# SPDX-FileCopyrightText: Copyright (c) 2021-2025 NVIDIA CORPORATION & AFFILIATES. All rights reserved. # SPDX-License-Identifier: LicenseRef-NVIDIA-SOFTWARE-LICENSE # This code was automatically generated with version 12.9.0. Do not modify it directly. from typing import Any, Optional from enum import IntEnum import cython import ctypes from libc.stdlib cimport calloc, malloc, free from libc cimport string from libc.stdint cimport int32_t, uint32_t, int64_t, uint64_t, uintptr_t from libc.stddef cimport wchar_t from libc.limits cimport CHAR_MIN from libcpp.vector cimport vector from cpython.buffer cimport PyObject_CheckBuffer, PyObject_GetBuffer, PyBuffer_Release, PyBUF_SIMPLE, PyBUF_ANY_CONTIGUOUS from cpython.bytes cimport PyBytes_FromStringAndSize import cuda.bindings.driver from libcpp.map cimport map import cuda.bindings.driver as _driver _driver = _driver.__dict__ include "_lib/utils.pxi" ctypedef unsigned long long signed_char_ptr ctypedef unsigned long long unsigned_char_ptr ctypedef unsigned long long char_ptr ctypedef unsigned long long short_ptr ctypedef unsigned long long unsigned_short_ptr ctypedef unsigned long long int_ptr ctypedef unsigned long long long_int_ptr ctypedef unsigned long long long_long_int_ptr ctypedef unsigned long long unsigned_int_ptr ctypedef unsigned long long unsigned_long_int_ptr ctypedef unsigned long long unsigned_long_long_int_ptr ctypedef unsigned long long uint32_t_ptr ctypedef unsigned long long uint64_t_ptr ctypedef unsigned long long int32_t_ptr ctypedef unsigned long long int64_t_ptr ctypedef unsigned long long unsigned_ptr ctypedef unsigned long long unsigned_long_long_ptr ctypedef unsigned long long long_long_ptr ctypedef unsigned long long size_t_ptr ctypedef unsigned long long long_ptr ctypedef unsigned long long float_ptr ctypedef unsigned long long double_ptr ctypedef unsigned long long void_ptr #: Default page-locked allocation flag cudaHostAllocDefault = cyruntime.cudaHostAllocDefault #: Pinned memory accessible by all CUDA contexts cudaHostAllocPortable = cyruntime.cudaHostAllocPortable #: Map allocation into device space cudaHostAllocMapped = cyruntime.cudaHostAllocMapped #: Write-combined memory cudaHostAllocWriteCombined = cyruntime.cudaHostAllocWriteCombined #: Default host memory registration flag cudaHostRegisterDefault = cyruntime.cudaHostRegisterDefault #: Pinned memory accessible by all CUDA contexts cudaHostRegisterPortable = cyruntime.cudaHostRegisterPortable #: Map registered memory into device space cudaHostRegisterMapped = cyruntime.cudaHostRegisterMapped #: Memory-mapped I/O space cudaHostRegisterIoMemory = cyruntime.cudaHostRegisterIoMemory #: Memory-mapped read-only cudaHostRegisterReadOnly = cyruntime.cudaHostRegisterReadOnly #: Default peer addressing enable flag cudaPeerAccessDefault = cyruntime.cudaPeerAccessDefault #: Default stream flag cudaStreamDefault = cyruntime.cudaStreamDefault #: Stream does not synchronize with stream 0 (the NULL stream) cudaStreamNonBlocking = cyruntime.cudaStreamNonBlocking #: Legacy stream handle #: #: Stream handle that can be passed as a cudaStream_t to use an implicit #: stream with legacy synchronization behavior. #: #: See details of the \link_sync_behavior cudaStreamLegacy = cyruntime.cudaStreamLegacy #: Per-thread stream handle #: #: Stream handle that can be passed as a cudaStream_t to use an implicit #: stream with per-thread synchronization behavior. #: #: See details of the \link_sync_behavior cudaStreamPerThread = cyruntime.cudaStreamPerThread #: Default event flag cudaEventDefault = cyruntime.cudaEventDefault #: Event uses blocking synchronization cudaEventBlockingSync = cyruntime.cudaEventBlockingSync #: Event will not record timing data cudaEventDisableTiming = cyruntime.cudaEventDisableTiming #: Event is suitable for interprocess use. cudaEventDisableTiming must be #: set cudaEventInterprocess = cyruntime.cudaEventInterprocess #: Default event record flag cudaEventRecordDefault = cyruntime.cudaEventRecordDefault #: Event is captured in the graph as an external event node when performing #: stream capture cudaEventRecordExternal = cyruntime.cudaEventRecordExternal #: Default event wait flag cudaEventWaitDefault = cyruntime.cudaEventWaitDefault #: Event is captured in the graph as an external event node when performing #: stream capture cudaEventWaitExternal = cyruntime.cudaEventWaitExternal #: Device flag - Automatic scheduling cudaDeviceScheduleAuto = cyruntime.cudaDeviceScheduleAuto #: Device flag - Spin default scheduling cudaDeviceScheduleSpin = cyruntime.cudaDeviceScheduleSpin #: Device flag - Yield default scheduling cudaDeviceScheduleYield = cyruntime.cudaDeviceScheduleYield #: Device flag - Use blocking synchronization cudaDeviceScheduleBlockingSync = cyruntime.cudaDeviceScheduleBlockingSync #: Device flag - Use blocking synchronization [Deprecated] cudaDeviceBlockingSync = cyruntime.cudaDeviceBlockingSync #: Device schedule flags mask cudaDeviceScheduleMask = cyruntime.cudaDeviceScheduleMask #: Device flag - Support mapped pinned allocations cudaDeviceMapHost = cyruntime.cudaDeviceMapHost #: Device flag - Keep local memory allocation after launch cudaDeviceLmemResizeToMax = cyruntime.cudaDeviceLmemResizeToMax #: Device flag - Ensure synchronous memory operations on this context will #: synchronize cudaDeviceSyncMemops = cyruntime.cudaDeviceSyncMemops #: Device flags mask cudaDeviceMask = cyruntime.cudaDeviceMask #: Default CUDA array allocation flag cudaArrayDefault = cyruntime.cudaArrayDefault #: Must be set in cudaMalloc3DArray to create a layered CUDA array cudaArrayLayered = cyruntime.cudaArrayLayered #: Must be set in cudaMallocArray or cudaMalloc3DArray in order to bind #: surfaces to the CUDA array cudaArraySurfaceLoadStore = cyruntime.cudaArraySurfaceLoadStore #: Must be set in cudaMalloc3DArray to create a cubemap CUDA array cudaArrayCubemap = cyruntime.cudaArrayCubemap #: Must be set in cudaMallocArray or cudaMalloc3DArray in order to perform #: texture gather operations on the CUDA array cudaArrayTextureGather = cyruntime.cudaArrayTextureGather #: Must be set in cudaExternalMemoryGetMappedMipmappedArray if the #: mipmapped array is used as a color target in a graphics API cudaArrayColorAttachment = cyruntime.cudaArrayColorAttachment #: Must be set in cudaMallocArray, cudaMalloc3DArray or #: cudaMallocMipmappedArray in order to create a sparse CUDA array or CUDA #: mipmapped array cudaArraySparse = cyruntime.cudaArraySparse #: Must be set in cudaMallocArray, cudaMalloc3DArray or #: cudaMallocMipmappedArray in order to create a deferred mapping CUDA #: array or CUDA mipmapped array cudaArrayDeferredMapping = cyruntime.cudaArrayDeferredMapping #: Automatically enable peer access between remote devices as needed cudaIpcMemLazyEnablePeerAccess = cyruntime.cudaIpcMemLazyEnablePeerAccess #: Memory can be accessed by any stream on any device cudaMemAttachGlobal = cyruntime.cudaMemAttachGlobal #: Memory cannot be accessed by any stream on any device cudaMemAttachHost = cyruntime.cudaMemAttachHost #: Memory can only be accessed by a single stream on the associated device cudaMemAttachSingle = cyruntime.cudaMemAttachSingle #: Default behavior cudaOccupancyDefault = cyruntime.cudaOccupancyDefault #: Assume global caching is enabled and cannot be automatically turned off cudaOccupancyDisableCachingOverride = cyruntime.cudaOccupancyDisableCachingOverride #: Device id that represents the CPU cudaCpuDeviceId = cyruntime.cudaCpuDeviceId #: Device id that represents an invalid device cudaInvalidDeviceId = cyruntime.cudaInvalidDeviceId #: Tell the CUDA runtime that DeviceFlags is being set in cudaInitDevice #: call cudaInitDeviceFlagsAreValid = cyruntime.cudaInitDeviceFlagsAreValid #: If set, each kernel launched as part of #: :py:obj:`~.cudaLaunchCooperativeKernelMultiDevice` only waits for prior #: work in the stream corresponding to that GPU to complete before the #: kernel begins execution. cudaCooperativeLaunchMultiDeviceNoPreSync = cyruntime.cudaCooperativeLaunchMultiDeviceNoPreSync #: If set, any subsequent work pushed in a stream that participated in a #: call to :py:obj:`~.cudaLaunchCooperativeKernelMultiDevice` will only #: wait for the kernel launched on the GPU corresponding to that stream to #: complete before it begins execution. cudaCooperativeLaunchMultiDeviceNoPostSync = cyruntime.cudaCooperativeLaunchMultiDeviceNoPostSync #: Indicates that the layered sparse CUDA array or CUDA mipmapped array has #: a single mip tail region for all layers cudaArraySparsePropertiesSingleMipTail = cyruntime.cudaArraySparsePropertiesSingleMipTail #: This flag, if set, indicates that the memory will be used as a buffer #: for hardware accelerated decompression. cudaMemPoolCreateUsageHwDecompress = cyruntime.cudaMemPoolCreateUsageHwDecompress #: CUDA IPC Handle Size CUDA_IPC_HANDLE_SIZE = cyruntime.CUDA_IPC_HANDLE_SIZE #: Indicates that the external memory object is a dedicated resource cudaExternalMemoryDedicated = cyruntime.cudaExternalMemoryDedicated #: When the /p flags parameter of #: :py:obj:`~.cudaExternalSemaphoreSignalParams` contains this flag, it #: indicates that signaling an external semaphore object should skip #: performing appropriate memory synchronization operations over all the #: external memory objects that are imported as #: :py:obj:`~.cudaExternalMemoryHandleTypeNvSciBuf`, which otherwise are #: performed by default to ensure data coherency with other importers of #: the same NvSciBuf memory objects. cudaExternalSemaphoreSignalSkipNvSciBufMemSync = cyruntime.cudaExternalSemaphoreSignalSkipNvSciBufMemSync #: When the /p flags parameter of #: :py:obj:`~.cudaExternalSemaphoreWaitParams` contains this flag, it #: indicates that waiting an external semaphore object should skip #: performing appropriate memory synchronization operations over all the #: external memory objects that are imported as #: :py:obj:`~.cudaExternalMemoryHandleTypeNvSciBuf`, which otherwise are #: performed by default to ensure data coherency with other importers of #: the same NvSciBuf memory objects. cudaExternalSemaphoreWaitSkipNvSciBufMemSync = cyruntime.cudaExternalSemaphoreWaitSkipNvSciBufMemSync #: When /p flags of :py:obj:`~.cudaDeviceGetNvSciSyncAttributes` is set to #: this, it indicates that application need signaler specific NvSciSyncAttr #: to be filled by :py:obj:`~.cudaDeviceGetNvSciSyncAttributes`. cudaNvSciSyncAttrSignal = cyruntime.cudaNvSciSyncAttrSignal #: When /p flags of :py:obj:`~.cudaDeviceGetNvSciSyncAttributes` is set to #: this, it indicates that application need waiter specific NvSciSyncAttr #: to be filled by :py:obj:`~.cudaDeviceGetNvSciSyncAttributes`. cudaNvSciSyncAttrWait = cyruntime.cudaNvSciSyncAttrWait #: This port activates when the kernel has finished executing. cudaGraphKernelNodePortDefault = cyruntime.cudaGraphKernelNodePortDefault #: This port activates when all blocks of the kernel have performed #: cudaTriggerProgrammaticLaunchCompletion() or have terminated. It must be #: used with edge type :py:obj:`~.cudaGraphDependencyTypeProgrammatic`. See #: also :py:obj:`~.cudaLaunchAttributeProgrammaticEvent`. cudaGraphKernelNodePortProgrammatic = cyruntime.cudaGraphKernelNodePortProgrammatic #: This port activates when all blocks of the kernel have begun execution. #: See also :py:obj:`~.cudaLaunchAttributeLaunchCompletionEvent`. cudaGraphKernelNodePortLaunchCompletion = cyruntime.cudaGraphKernelNodePortLaunchCompletion cudaStreamAttributeAccessPolicyWindow = cyruntime.cudaStreamAttributeAccessPolicyWindow cudaStreamAttributeSynchronizationPolicy = cyruntime.cudaStreamAttributeSynchronizationPolicy cudaStreamAttributeMemSyncDomainMap = cyruntime.cudaStreamAttributeMemSyncDomainMap cudaStreamAttributeMemSyncDomain = cyruntime.cudaStreamAttributeMemSyncDomain cudaStreamAttributePriority = cyruntime.cudaStreamAttributePriority cudaKernelNodeAttributeAccessPolicyWindow = cyruntime.cudaKernelNodeAttributeAccessPolicyWindow cudaKernelNodeAttributeCooperative = cyruntime.cudaKernelNodeAttributeCooperative cudaKernelNodeAttributePriority = cyruntime.cudaKernelNodeAttributePriority cudaKernelNodeAttributeClusterDimension = cyruntime.cudaKernelNodeAttributeClusterDimension cudaKernelNodeAttributeClusterSchedulingPolicyPreference = cyruntime.cudaKernelNodeAttributeClusterSchedulingPolicyPreference cudaKernelNodeAttributeMemSyncDomainMap = cyruntime.cudaKernelNodeAttributeMemSyncDomainMap cudaKernelNodeAttributeMemSyncDomain = cyruntime.cudaKernelNodeAttributeMemSyncDomain cudaKernelNodeAttributePreferredSharedMemoryCarveout = cyruntime.cudaKernelNodeAttributePreferredSharedMemoryCarveout cudaKernelNodeAttributeDeviceUpdatableKernelNode = cyruntime.cudaKernelNodeAttributeDeviceUpdatableKernelNode cudaSurfaceType1D = cyruntime.cudaSurfaceType1D cudaSurfaceType2D = cyruntime.cudaSurfaceType2D cudaSurfaceType3D = cyruntime.cudaSurfaceType3D cudaSurfaceTypeCubemap = cyruntime.cudaSurfaceTypeCubemap cudaSurfaceType1DLayered = cyruntime.cudaSurfaceType1DLayered cudaSurfaceType2DLayered = cyruntime.cudaSurfaceType2DLayered cudaSurfaceTypeCubemapLayered = cyruntime.cudaSurfaceTypeCubemapLayered cudaTextureType1D = cyruntime.cudaTextureType1D cudaTextureType2D = cyruntime.cudaTextureType2D cudaTextureType3D = cyruntime.cudaTextureType3D cudaTextureTypeCubemap = cyruntime.cudaTextureTypeCubemap cudaTextureType1DLayered = cyruntime.cudaTextureType1DLayered cudaTextureType2DLayered = cyruntime.cudaTextureType2DLayered cudaTextureTypeCubemapLayered = cyruntime.cudaTextureTypeCubemapLayered #: CUDA Runtime API Version CUDART_VERSION = cyruntime.CUDART_VERSION __CUDART_API_VERSION = cyruntime.__CUDART_API_VERSION #: Maximum number of planes per frame CUDA_EGL_MAX_PLANES = cyruntime.CUDA_EGL_MAX_PLANES class cudaError_t(IntEnum): """ impl_private CUDA error types """ #: The API call returned with no errors. In the case of query calls, #: this also means that the operation being queried is complete (see #: :py:obj:`~.cudaEventQuery()` and :py:obj:`~.cudaStreamQuery()`). cudaSuccess = cyruntime.cudaError.cudaSuccess #: This indicates that one or more of the parameters passed to the API #: call is not within an acceptable range of values. cudaErrorInvalidValue = cyruntime.cudaError.cudaErrorInvalidValue #: The API call failed because it was unable to allocate enough memory #: or other resources to perform the requested operation. cudaErrorMemoryAllocation = cyruntime.cudaError.cudaErrorMemoryAllocation #: The API call failed because the CUDA driver and runtime could not be #: initialized. cudaErrorInitializationError = cyruntime.cudaError.cudaErrorInitializationError #: This indicates that a CUDA Runtime API call cannot be executed #: because it is being called during process shut down, at a point in #: time after CUDA driver has been unloaded. cudaErrorCudartUnloading = cyruntime.cudaError.cudaErrorCudartUnloading #: This indicates profiler is not initialized for this run. This can #: happen when the application is running with external profiling tools #: like visual profiler. cudaErrorProfilerDisabled = cyruntime.cudaError.cudaErrorProfilerDisabled #: [Deprecated] cudaErrorProfilerNotInitialized = cyruntime.cudaError.cudaErrorProfilerNotInitialized #: [Deprecated] cudaErrorProfilerAlreadyStarted = cyruntime.cudaError.cudaErrorProfilerAlreadyStarted #: [Deprecated] cudaErrorProfilerAlreadyStopped = cyruntime.cudaError.cudaErrorProfilerAlreadyStopped #: This indicates that a kernel launch is requesting resources that can #: never be satisfied by the current device. Requesting more shared #: memory per block than the device supports will trigger this error, #: as will requesting too many threads or blocks. See #: :py:obj:`~.cudaDeviceProp` for more device limitations. cudaErrorInvalidConfiguration = cyruntime.cudaError.cudaErrorInvalidConfiguration #: This indicates that one or more of the pitch-related parameters #: passed to the API call is not within the acceptable range for pitch. cudaErrorInvalidPitchValue = cyruntime.cudaError.cudaErrorInvalidPitchValue #: This indicates that the symbol name/identifier passed to the API #: call is not a valid name or identifier. cudaErrorInvalidSymbol = cyruntime.cudaError.cudaErrorInvalidSymbol #: This indicates that at least one host pointer passed to the API call #: is not a valid host pointer. [Deprecated] cudaErrorInvalidHostPointer = cyruntime.cudaError.cudaErrorInvalidHostPointer #: This indicates that at least one device pointer passed to the API #: call is not a valid device pointer. [Deprecated] cudaErrorInvalidDevicePointer = cyruntime.cudaError.cudaErrorInvalidDevicePointer #: This indicates that the texture passed to the API call is not a #: valid texture. cudaErrorInvalidTexture = cyruntime.cudaError.cudaErrorInvalidTexture #: This indicates that the texture binding is not valid. This occurs if #: you call :py:obj:`~.cudaGetTextureAlignmentOffset()` with an unbound #: texture. cudaErrorInvalidTextureBinding = cyruntime.cudaError.cudaErrorInvalidTextureBinding #: This indicates that the channel descriptor passed to the API call is #: not valid. This occurs if the format is not one of the formats #: specified by :py:obj:`~.cudaChannelFormatKind`, or if one of the #: dimensions is invalid. cudaErrorInvalidChannelDescriptor = cyruntime.cudaError.cudaErrorInvalidChannelDescriptor #: This indicates that the direction of the memcpy passed to the API #: call is not one of the types specified by #: :py:obj:`~.cudaMemcpyKind`. cudaErrorInvalidMemcpyDirection = cyruntime.cudaError.cudaErrorInvalidMemcpyDirection #: This indicated that the user has taken the address of a constant #: variable, which was forbidden up until the CUDA 3.1 release. #: [Deprecated] cudaErrorAddressOfConstant = cyruntime.cudaError.cudaErrorAddressOfConstant #: This indicated that a texture fetch was not able to be performed. #: This was previously used for device emulation of texture operations. #: [Deprecated] cudaErrorTextureFetchFailed = cyruntime.cudaError.cudaErrorTextureFetchFailed #: This indicated that a texture was not bound for access. This was #: previously used for device emulation of texture operations. #: [Deprecated] cudaErrorTextureNotBound = cyruntime.cudaError.cudaErrorTextureNotBound #: This indicated that a synchronization operation had failed. This was #: previously used for some device emulation functions. [Deprecated] cudaErrorSynchronizationError = cyruntime.cudaError.cudaErrorSynchronizationError #: This indicates that a non-float texture was being accessed with #: linear filtering. This is not supported by CUDA. cudaErrorInvalidFilterSetting = cyruntime.cudaError.cudaErrorInvalidFilterSetting #: This indicates that an attempt was made to read an unsupported data #: type as a normalized float. This is not supported by CUDA. cudaErrorInvalidNormSetting = cyruntime.cudaError.cudaErrorInvalidNormSetting #: Mixing of device and device emulation code was not allowed. #: [Deprecated] cudaErrorMixedDeviceExecution = cyruntime.cudaError.cudaErrorMixedDeviceExecution #: This indicates that the API call is not yet implemented. Production #: releases of CUDA will never return this error. [Deprecated] cudaErrorNotYetImplemented = cyruntime.cudaError.cudaErrorNotYetImplemented #: This indicated that an emulated device pointer exceeded the 32-bit #: address range. [Deprecated] cudaErrorMemoryValueTooLarge = cyruntime.cudaError.cudaErrorMemoryValueTooLarge #: This indicates that the CUDA driver that the application has loaded #: is a stub library. Applications that run with the stub rather than a #: real driver loaded will result in CUDA API returning this error. cudaErrorStubLibrary = cyruntime.cudaError.cudaErrorStubLibrary #: This indicates that the installed NVIDIA CUDA driver is older than #: the CUDA runtime library. This is not a supported configuration. #: Users should install an updated NVIDIA display driver to allow the #: application to run. cudaErrorInsufficientDriver = cyruntime.cudaError.cudaErrorInsufficientDriver #: This indicates that the API call requires a newer CUDA driver than #: the one currently installed. Users should install an updated NVIDIA #: CUDA driver to allow the API call to succeed. cudaErrorCallRequiresNewerDriver = cyruntime.cudaError.cudaErrorCallRequiresNewerDriver #: This indicates that the surface passed to the API call is not a #: valid surface. cudaErrorInvalidSurface = cyruntime.cudaError.cudaErrorInvalidSurface #: This indicates that multiple global or constant variables (across #: separate CUDA source files in the application) share the same string #: name. cudaErrorDuplicateVariableName = cyruntime.cudaError.cudaErrorDuplicateVariableName #: This indicates that multiple textures (across separate CUDA source #: files in the application) share the same string name. cudaErrorDuplicateTextureName = cyruntime.cudaError.cudaErrorDuplicateTextureName #: This indicates that multiple surfaces (across separate CUDA source #: files in the application) share the same string name. cudaErrorDuplicateSurfaceName = cyruntime.cudaError.cudaErrorDuplicateSurfaceName #: This indicates that all CUDA devices are busy or unavailable at the #: current time. Devices are often busy/unavailable due to use of #: :py:obj:`~.cudaComputeModeProhibited`, #: :py:obj:`~.cudaComputeModeExclusiveProcess`, or when long running #: CUDA kernels have filled up the GPU and are blocking new work from #: starting. They can also be unavailable due to memory constraints on #: a device that already has active CUDA work being performed. cudaErrorDevicesUnavailable = cyruntime.cudaError.cudaErrorDevicesUnavailable #: This indicates that the current context is not compatible with this #: the CUDA Runtime. This can only occur if you are using CUDA #: Runtime/Driver interoperability and have created an existing Driver #: context using the driver API. The Driver context may be incompatible #: either because the Driver context was created using an older version #: of the API, because the Runtime API call expects a primary driver #: context and the Driver context is not primary, or because the Driver #: context has been destroyed. Please see :py:obj:`~.Interactions`with #: the CUDA Driver API" for more information. cudaErrorIncompatibleDriverContext = cyruntime.cudaError.cudaErrorIncompatibleDriverContext #: The device function being invoked (usually via #: :py:obj:`~.cudaLaunchKernel()`) was not previously configured via #: the :py:obj:`~.cudaConfigureCall()` function. cudaErrorMissingConfiguration = cyruntime.cudaError.cudaErrorMissingConfiguration #: This indicated that a previous kernel launch failed. This was #: previously used for device emulation of kernel launches. #: [Deprecated] cudaErrorPriorLaunchFailure = cyruntime.cudaError.cudaErrorPriorLaunchFailure #: This error indicates that a device runtime grid launch did not occur #: because the depth of the child grid would exceed the maximum #: supported number of nested grid launches. cudaErrorLaunchMaxDepthExceeded = cyruntime.cudaError.cudaErrorLaunchMaxDepthExceeded #: This error indicates that a grid launch did not occur because the #: kernel uses file-scoped textures which are unsupported by the device #: runtime. Kernels launched via the device runtime only support #: textures created with the Texture Object API's. cudaErrorLaunchFileScopedTex = cyruntime.cudaError.cudaErrorLaunchFileScopedTex #: This error indicates that a grid launch did not occur because the #: kernel uses file-scoped surfaces which are unsupported by the device #: runtime. Kernels launched via the device runtime only support #: surfaces created with the Surface Object API's. cudaErrorLaunchFileScopedSurf = cyruntime.cudaError.cudaErrorLaunchFileScopedSurf #: This error indicates that a call to #: :py:obj:`~.cudaDeviceSynchronize` made from the device runtime #: failed because the call was made at grid depth greater than than #: either the default (2 levels of grids) or user specified device #: limit :py:obj:`~.cudaLimitDevRuntimeSyncDepth`. To be able to #: synchronize on launched grids at a greater depth successfully, the #: maximum nested depth at which :py:obj:`~.cudaDeviceSynchronize` will #: be called must be specified with the #: :py:obj:`~.cudaLimitDevRuntimeSyncDepth` limit to the #: :py:obj:`~.cudaDeviceSetLimit` api before the host-side launch of a #: kernel using the device runtime. Keep in mind that additional levels #: of sync depth require the runtime to reserve large amounts of device #: memory that cannot be used for user allocations. Note that #: :py:obj:`~.cudaDeviceSynchronize` made from device runtime is only #: supported on devices of compute capability < 9.0. cudaErrorSyncDepthExceeded = cyruntime.cudaError.cudaErrorSyncDepthExceeded #: This error indicates that a device runtime grid launch failed #: because the launch would exceed the limit #: :py:obj:`~.cudaLimitDevRuntimePendingLaunchCount`. For this launch #: to proceed successfully, :py:obj:`~.cudaDeviceSetLimit` must be #: called to set the :py:obj:`~.cudaLimitDevRuntimePendingLaunchCount` #: to be higher than the upper bound of outstanding launches that can #: be issued to the device runtime. Keep in mind that raising the limit #: of pending device runtime launches will require the runtime to #: reserve device memory that cannot be used for user allocations. cudaErrorLaunchPendingCountExceeded = cyruntime.cudaError.cudaErrorLaunchPendingCountExceeded #: The requested device function does not exist or is not compiled for #: the proper device architecture. cudaErrorInvalidDeviceFunction = cyruntime.cudaError.cudaErrorInvalidDeviceFunction #: This indicates that no CUDA-capable devices were detected by the #: installed CUDA driver. cudaErrorNoDevice = cyruntime.cudaError.cudaErrorNoDevice #: This indicates that the device ordinal supplied by the user does not #: correspond to a valid CUDA device or that the action requested is #: invalid for the specified device. cudaErrorInvalidDevice = cyruntime.cudaError.cudaErrorInvalidDevice #: This indicates that the device doesn't have a valid Grid License. cudaErrorDeviceNotLicensed = cyruntime.cudaError.cudaErrorDeviceNotLicensed #: By default, the CUDA runtime may perform a minimal set of self- #: tests, as well as CUDA driver tests, to establish the validity of #: both. Introduced in CUDA 11.2, this error return indicates that at #: least one of these tests has failed and the validity of either the #: runtime or the driver could not be established. cudaErrorSoftwareValidityNotEstablished = cyruntime.cudaError.cudaErrorSoftwareValidityNotEstablished #: This indicates an internal startup failure in the CUDA runtime. cudaErrorStartupFailure = cyruntime.cudaError.cudaErrorStartupFailure #: This indicates that the device kernel image is invalid. cudaErrorInvalidKernelImage = cyruntime.cudaError.cudaErrorInvalidKernelImage #: This most frequently indicates that there is no context bound to the #: current thread. This can also be returned if the context passed to #: an API call is not a valid handle (such as a context that has had #: :py:obj:`~.cuCtxDestroy()` invoked on it). This can also be returned #: if a user mixes different API versions (i.e. 3010 context with 3020 #: API calls). See :py:obj:`~.cuCtxGetApiVersion()` for more details. cudaErrorDeviceUninitialized = cyruntime.cudaError.cudaErrorDeviceUninitialized #: This indicates that the buffer object could not be mapped. cudaErrorMapBufferObjectFailed = cyruntime.cudaError.cudaErrorMapBufferObjectFailed #: This indicates that the buffer object could not be unmapped. cudaErrorUnmapBufferObjectFailed = cyruntime.cudaError.cudaErrorUnmapBufferObjectFailed #: This indicates that the specified array is currently mapped and thus #: cannot be destroyed. cudaErrorArrayIsMapped = cyruntime.cudaError.cudaErrorArrayIsMapped #: This indicates that the resource is already mapped. cudaErrorAlreadyMapped = cyruntime.cudaError.cudaErrorAlreadyMapped #: This indicates that there is no kernel image available that is #: suitable for the device. This can occur when a user specifies code #: generation options for a particular CUDA source file that do not #: include the corresponding device configuration. cudaErrorNoKernelImageForDevice = cyruntime.cudaError.cudaErrorNoKernelImageForDevice #: This indicates that a resource has already been acquired. cudaErrorAlreadyAcquired = cyruntime.cudaError.cudaErrorAlreadyAcquired #: This indicates that a resource is not mapped. cudaErrorNotMapped = cyruntime.cudaError.cudaErrorNotMapped #: This indicates that a mapped resource is not available for access as #: an array. cudaErrorNotMappedAsArray = cyruntime.cudaError.cudaErrorNotMappedAsArray #: This indicates that a mapped resource is not available for access as #: a pointer. cudaErrorNotMappedAsPointer = cyruntime.cudaError.cudaErrorNotMappedAsPointer #: This indicates that an uncorrectable ECC error was detected during #: execution. cudaErrorECCUncorrectable = cyruntime.cudaError.cudaErrorECCUncorrectable #: This indicates that the :py:obj:`~.cudaLimit` passed to the API call #: is not supported by the active device. cudaErrorUnsupportedLimit = cyruntime.cudaError.cudaErrorUnsupportedLimit #: This indicates that a call tried to access an exclusive-thread #: device that is already in use by a different thread. cudaErrorDeviceAlreadyInUse = cyruntime.cudaError.cudaErrorDeviceAlreadyInUse #: This error indicates that P2P access is not supported across the #: given devices. cudaErrorPeerAccessUnsupported = cyruntime.cudaError.cudaErrorPeerAccessUnsupported #: A PTX compilation failed. The runtime may fall back to compiling PTX #: if an application does not contain a suitable binary for the current #: device. cudaErrorInvalidPtx = cyruntime.cudaError.cudaErrorInvalidPtx #: This indicates an error with the OpenGL or DirectX context. cudaErrorInvalidGraphicsContext = cyruntime.cudaError.cudaErrorInvalidGraphicsContext #: This indicates that an uncorrectable NVLink error was detected #: during the execution. cudaErrorNvlinkUncorrectable = cyruntime.cudaError.cudaErrorNvlinkUncorrectable #: This indicates that the PTX JIT compiler library was not found. The #: JIT Compiler library is used for PTX compilation. The runtime may #: fall back to compiling PTX if an application does not contain a #: suitable binary for the current device. cudaErrorJitCompilerNotFound = cyruntime.cudaError.cudaErrorJitCompilerNotFound #: This indicates that the provided PTX was compiled with an #: unsupported toolchain. The most common reason for this, is the PTX #: was generated by a compiler newer than what is supported by the CUDA #: driver and PTX JIT compiler. cudaErrorUnsupportedPtxVersion = cyruntime.cudaError.cudaErrorUnsupportedPtxVersion #: This indicates that the JIT compilation was disabled. The JIT #: compilation compiles PTX. The runtime may fall back to compiling PTX #: if an application does not contain a suitable binary for the current #: device. cudaErrorJitCompilationDisabled = cyruntime.cudaError.cudaErrorJitCompilationDisabled #: This indicates that the provided execution affinity is not supported #: by the device. cudaErrorUnsupportedExecAffinity = cyruntime.cudaError.cudaErrorUnsupportedExecAffinity #: This indicates that the code to be compiled by the PTX JIT contains #: unsupported call to cudaDeviceSynchronize. cudaErrorUnsupportedDevSideSync = cyruntime.cudaError.cudaErrorUnsupportedDevSideSync #: This indicates that an exception occurred on the device that is now #: contained by the GPU's error containment capability. Common causes #: are - a. Certain types of invalid accesses of peer GPU memory over #: nvlink b. Certain classes of hardware errors This leaves the process #: in an inconsistent state and any further CUDA work will return the #: same error. To continue using CUDA, the process must be terminated #: and relaunched. cudaErrorContained = cyruntime.cudaError.cudaErrorContained #: This indicates that the device kernel source is invalid. cudaErrorInvalidSource = cyruntime.cudaError.cudaErrorInvalidSource #: This indicates that the file specified was not found. cudaErrorFileNotFound = cyruntime.cudaError.cudaErrorFileNotFound #: This indicates that a link to a shared object failed to resolve. cudaErrorSharedObjectSymbolNotFound = cyruntime.cudaError.cudaErrorSharedObjectSymbolNotFound #: This indicates that initialization of a shared object failed. cudaErrorSharedObjectInitFailed = cyruntime.cudaError.cudaErrorSharedObjectInitFailed #: This error indicates that an OS call failed. cudaErrorOperatingSystem = cyruntime.cudaError.cudaErrorOperatingSystem #: This indicates that a resource handle passed to the API call was not #: valid. Resource handles are opaque types like #: :py:obj:`~.cudaStream_t` and :py:obj:`~.cudaEvent_t`. cudaErrorInvalidResourceHandle = cyruntime.cudaError.cudaErrorInvalidResourceHandle #: This indicates that a resource required by the API call is not in a #: valid state to perform the requested operation. cudaErrorIllegalState = cyruntime.cudaError.cudaErrorIllegalState #: This indicates an attempt was made to introspect an object in a way #: that would discard semantically important information. This is #: either due to the object using funtionality newer than the API #: version used to introspect it or omission of optional return #: arguments. cudaErrorLossyQuery = cyruntime.cudaError.cudaErrorLossyQuery #: This indicates that a named symbol was not found. Examples of #: symbols are global/constant variable names, driver function names, #: texture names, and surface names. cudaErrorSymbolNotFound = cyruntime.cudaError.cudaErrorSymbolNotFound #: This indicates that asynchronous operations issued previously have #: not completed yet. This result is not actually an error, but must be #: indicated differently than :py:obj:`~.cudaSuccess` (which indicates #: completion). Calls that may return this value include #: :py:obj:`~.cudaEventQuery()` and :py:obj:`~.cudaStreamQuery()`. cudaErrorNotReady = cyruntime.cudaError.cudaErrorNotReady #: The device encountered a load or store instruction on an invalid #: memory address. This leaves the process in an inconsistent state and #: any further CUDA work will return the same error. To continue using #: CUDA, the process must be terminated and relaunched. cudaErrorIllegalAddress = cyruntime.cudaError.cudaErrorIllegalAddress #: This indicates that a launch did not occur because it did not have #: appropriate resources. Although this error is similar to #: :py:obj:`~.cudaErrorInvalidConfiguration`, this error usually #: indicates that the user has attempted to pass too many arguments to #: the device kernel, or the kernel launch specifies too many threads #: for the kernel's register count. cudaErrorLaunchOutOfResources = cyruntime.cudaError.cudaErrorLaunchOutOfResources #: This indicates that the device kernel took too long to execute. This #: can only occur if timeouts are enabled - see the device property #: :py:obj:`~.kernelExecTimeoutEnabled` for more information. This #: leaves the process in an inconsistent state and any further CUDA #: work will return the same error. To continue using CUDA, the process #: must be terminated and relaunched. cudaErrorLaunchTimeout = cyruntime.cudaError.cudaErrorLaunchTimeout #: This error indicates a kernel launch that uses an incompatible #: texturing mode. cudaErrorLaunchIncompatibleTexturing = cyruntime.cudaError.cudaErrorLaunchIncompatibleTexturing #: This error indicates that a call to #: :py:obj:`~.cudaDeviceEnablePeerAccess()` is trying to re-enable peer #: addressing on from a context which has already had peer addressing #: enabled. cudaErrorPeerAccessAlreadyEnabled = cyruntime.cudaError.cudaErrorPeerAccessAlreadyEnabled #: This error indicates that :py:obj:`~.cudaDeviceDisablePeerAccess()` #: is trying to disable peer addressing which has not been enabled yet #: via :py:obj:`~.cudaDeviceEnablePeerAccess()`. cudaErrorPeerAccessNotEnabled = cyruntime.cudaError.cudaErrorPeerAccessNotEnabled #: This indicates that the user has called #: :py:obj:`~.cudaSetValidDevices()`, :py:obj:`~.cudaSetDeviceFlags()`, #: :py:obj:`~.cudaD3D9SetDirect3DDevice()`, #: :py:obj:`~.cudaD3D10SetDirect3DDevice`, #: :py:obj:`~.cudaD3D11SetDirect3DDevice()`, or #: :py:obj:`~.cudaVDPAUSetVDPAUDevice()` after initializing the CUDA #: runtime by calling non-device management operations (allocating #: memory and launching kernels are examples of non-device management #: operations). This error can also be returned if using runtime/driver #: interoperability and there is an existing :py:obj:`~.CUcontext` #: active on the host thread. cudaErrorSetOnActiveProcess = cyruntime.cudaError.cudaErrorSetOnActiveProcess #: This error indicates that the context current to the calling thread #: has been destroyed using :py:obj:`~.cuCtxDestroy`, or is a primary #: context which has not yet been initialized. cudaErrorContextIsDestroyed = cyruntime.cudaError.cudaErrorContextIsDestroyed #: An assert triggered in device code during kernel execution. The #: device cannot be used again. All existing allocations are invalid. #: To continue using CUDA, the process must be terminated and #: relaunched. cudaErrorAssert = cyruntime.cudaError.cudaErrorAssert #: This error indicates that the hardware resources required to enable #: peer access have been exhausted for one or more of the devices #: passed to :py:obj:`~.cudaEnablePeerAccess()`. cudaErrorTooManyPeers = cyruntime.cudaError.cudaErrorTooManyPeers #: This error indicates that the memory range passed to #: :py:obj:`~.cudaHostRegister()` has already been registered. cudaErrorHostMemoryAlreadyRegistered = cyruntime.cudaError.cudaErrorHostMemoryAlreadyRegistered #: This error indicates that the pointer passed to #: :py:obj:`~.cudaHostUnregister()` does not correspond to any #: currently registered memory region. cudaErrorHostMemoryNotRegistered = cyruntime.cudaError.cudaErrorHostMemoryNotRegistered #: Device encountered an error in the call stack during kernel #: execution, possibly due to stack corruption or exceeding the stack #: size limit. This leaves the process in an inconsistent state and any #: further CUDA work will return the same error. To continue using #: CUDA, the process must be terminated and relaunched. cudaErrorHardwareStackError = cyruntime.cudaError.cudaErrorHardwareStackError #: The device encountered an illegal instruction during kernel #: execution This leaves the process in an inconsistent state and any #: further CUDA work will return the same error. To continue using #: CUDA, the process must be terminated and relaunched. cudaErrorIllegalInstruction = cyruntime.cudaError.cudaErrorIllegalInstruction #: The device encountered a load or store instruction on a memory #: address which is not aligned. This leaves the process in an #: inconsistent state and any further CUDA work will return the same #: error. To continue using CUDA, the process must be terminated and #: relaunched. cudaErrorMisalignedAddress = cyruntime.cudaError.cudaErrorMisalignedAddress #: While executing a kernel, the device encountered an instruction #: which can only operate on memory locations in certain address spaces #: (global, shared, or local), but was supplied a memory address not #: belonging to an allowed address space. This leaves the process in an #: inconsistent state and any further CUDA work will return the same #: error. To continue using CUDA, the process must be terminated and #: relaunched. cudaErrorInvalidAddressSpace = cyruntime.cudaError.cudaErrorInvalidAddressSpace #: The device encountered an invalid program counter. This leaves the #: process in an inconsistent state and any further CUDA work will #: return the same error. To continue using CUDA, the process must be #: terminated and relaunched. cudaErrorInvalidPc = cyruntime.cudaError.cudaErrorInvalidPc #: An exception occurred on the device while executing a kernel. Common #: causes include dereferencing an invalid device pointer and accessing #: out of bounds shared memory. Less common cases can be system #: specific - more information about these cases can be found in the #: system specific user guide. This leaves the process in an #: inconsistent state and any further CUDA work will return the same #: error. To continue using CUDA, the process must be terminated and #: relaunched. cudaErrorLaunchFailure = cyruntime.cudaError.cudaErrorLaunchFailure #: This error indicates that the number of blocks launched per grid for #: a kernel that was launched via either #: :py:obj:`~.cudaLaunchCooperativeKernel` or #: :py:obj:`~.cudaLaunchCooperativeKernelMultiDevice` exceeds the #: maximum number of blocks as allowed by #: :py:obj:`~.cudaOccupancyMaxActiveBlocksPerMultiprocessor` or #: :py:obj:`~.cudaOccupancyMaxActiveBlocksPerMultiprocessorWithFlags` #: times the number of multiprocessors as specified by the device #: attribute :py:obj:`~.cudaDevAttrMultiProcessorCount`. cudaErrorCooperativeLaunchTooLarge = cyruntime.cudaError.cudaErrorCooperativeLaunchTooLarge #: An exception occurred on the device while exiting a kernel using #: tensor memory: the tensor memory was not completely deallocated. #: This leaves the process in an inconsistent state and any further #: CUDA work will return the same error. To continue using CUDA, the #: process must be terminated and relaunched. cudaErrorTensorMemoryLeak = cyruntime.cudaError.cudaErrorTensorMemoryLeak #: This error indicates the attempted operation is not permitted. cudaErrorNotPermitted = cyruntime.cudaError.cudaErrorNotPermitted #: This error indicates the attempted operation is not supported on the #: current system or device. cudaErrorNotSupported = cyruntime.cudaError.cudaErrorNotSupported #: This error indicates that the system is not yet ready to start any #: CUDA work. To continue using CUDA, verify the system configuration #: is in a valid state and all required driver daemons are actively #: running. More information about this error can be found in the #: system specific user guide. cudaErrorSystemNotReady = cyruntime.cudaError.cudaErrorSystemNotReady #: This error indicates that there is a mismatch between the versions #: of the display driver and the CUDA driver. Refer to the #: compatibility documentation for supported versions. cudaErrorSystemDriverMismatch = cyruntime.cudaError.cudaErrorSystemDriverMismatch #: This error indicates that the system was upgraded to run with #: forward compatibility but the visible hardware detected by CUDA does #: not support this configuration. Refer to the compatibility #: documentation for the supported hardware matrix or ensure that only #: supported hardware is visible during initialization via the #: CUDA_VISIBLE_DEVICES environment variable. cudaErrorCompatNotSupportedOnDevice = cyruntime.cudaError.cudaErrorCompatNotSupportedOnDevice #: This error indicates that the MPS client failed to connect to the #: MPS control daemon or the MPS server. cudaErrorMpsConnectionFailed = cyruntime.cudaError.cudaErrorMpsConnectionFailed #: This error indicates that the remote procedural call between the MPS #: server and the MPS client failed. cudaErrorMpsRpcFailure = cyruntime.cudaError.cudaErrorMpsRpcFailure #: This error indicates that the MPS server is not ready to accept new #: MPS client requests. This error can be returned when the MPS server #: is in the process of recovering from a fatal failure. cudaErrorMpsServerNotReady = cyruntime.cudaError.cudaErrorMpsServerNotReady #: This error indicates that the hardware resources required to create #: MPS client have been exhausted. cudaErrorMpsMaxClientsReached = cyruntime.cudaError.cudaErrorMpsMaxClientsReached #: This error indicates the the hardware resources required to device #: connections have been exhausted. cudaErrorMpsMaxConnectionsReached = cyruntime.cudaError.cudaErrorMpsMaxConnectionsReached #: This error indicates that the MPS client has been terminated by the #: server. To continue using CUDA, the process must be terminated and #: relaunched. cudaErrorMpsClientTerminated = cyruntime.cudaError.cudaErrorMpsClientTerminated #: This error indicates, that the program is using CUDA Dynamic #: Parallelism, but the current configuration, like MPS, does not #: support it. cudaErrorCdpNotSupported = cyruntime.cudaError.cudaErrorCdpNotSupported #: This error indicates, that the program contains an unsupported #: interaction between different versions of CUDA Dynamic Parallelism. cudaErrorCdpVersionMismatch = cyruntime.cudaError.cudaErrorCdpVersionMismatch #: The operation is not permitted when the stream is capturing. cudaErrorStreamCaptureUnsupported = cyruntime.cudaError.cudaErrorStreamCaptureUnsupported #: The current capture sequence on the stream has been invalidated due #: to a previous error. cudaErrorStreamCaptureInvalidated = cyruntime.cudaError.cudaErrorStreamCaptureInvalidated #: The operation would have resulted in a merge of two independent #: capture sequences. cudaErrorStreamCaptureMerge = cyruntime.cudaError.cudaErrorStreamCaptureMerge #: The capture was not initiated in this stream. cudaErrorStreamCaptureUnmatched = cyruntime.cudaError.cudaErrorStreamCaptureUnmatched #: The capture sequence contains a fork that was not joined to the #: primary stream. cudaErrorStreamCaptureUnjoined = cyruntime.cudaError.cudaErrorStreamCaptureUnjoined #: A dependency would have been created which crosses the capture #: sequence boundary. Only implicit in-stream ordering dependencies are #: allowed to cross the boundary. cudaErrorStreamCaptureIsolation = cyruntime.cudaError.cudaErrorStreamCaptureIsolation #: The operation would have resulted in a disallowed implicit #: dependency on a current capture sequence from cudaStreamLegacy. cudaErrorStreamCaptureImplicit = cyruntime.cudaError.cudaErrorStreamCaptureImplicit #: The operation is not permitted on an event which was last recorded #: in a capturing stream. cudaErrorCapturedEvent = cyruntime.cudaError.cudaErrorCapturedEvent #: A stream capture sequence not initiated with the #: :py:obj:`~.cudaStreamCaptureModeRelaxed` argument to #: :py:obj:`~.cudaStreamBeginCapture` was passed to #: :py:obj:`~.cudaStreamEndCapture` in a different thread. cudaErrorStreamCaptureWrongThread = cyruntime.cudaError.cudaErrorStreamCaptureWrongThread #: This indicates that the wait operation has timed out. cudaErrorTimeout = cyruntime.cudaError.cudaErrorTimeout #: This error indicates that the graph update was not performed because #: it included changes which violated constraints specific to #: instantiated graph update. cudaErrorGraphExecUpdateFailure = cyruntime.cudaError.cudaErrorGraphExecUpdateFailure #: This indicates that an async error has occurred in a device outside #: of CUDA. If CUDA was waiting for an external device's signal before #: consuming shared data, the external device signaled an error #: indicating that the data is not valid for consumption. This leaves #: the process in an inconsistent state and any further CUDA work will #: return the same error. To continue using CUDA, the process must be #: terminated and relaunched. cudaErrorExternalDevice = cyruntime.cudaError.cudaErrorExternalDevice #: This indicates that a kernel launch error has occurred due to #: cluster misconfiguration. cudaErrorInvalidClusterSize = cyruntime.cudaError.cudaErrorInvalidClusterSize #: Indiciates a function handle is not loaded when calling an API that #: requires a loaded function. cudaErrorFunctionNotLoaded = cyruntime.cudaError.cudaErrorFunctionNotLoaded #: This error indicates one or more resources passed in are not valid #: resource types for the operation. cudaErrorInvalidResourceType = cyruntime.cudaError.cudaErrorInvalidResourceType #: This error indicates one or more resources are insufficient or non- #: applicable for the operation. cudaErrorInvalidResourceConfiguration = cyruntime.cudaError.cudaErrorInvalidResourceConfiguration #: This indicates that an unknown internal error has occurred. cudaErrorUnknown = cyruntime.cudaError.cudaErrorUnknown cudaErrorApiFailureBase = cyruntime.cudaError.cudaErrorApiFailureBase _dict_cudaError_t = dict(((int(v), v) for k, v in cudaError_t.__members__.items())) class cudaGraphDependencyType(IntEnum): """ Type annotations that can be applied to graph edges as part of :py:obj:`~.cudaGraphEdgeData`. """ #: This is an ordinary dependency. cudaGraphDependencyTypeDefault = cyruntime.cudaGraphDependencyType_enum.cudaGraphDependencyTypeDefault #: This dependency type allows the downstream node to use #: `cudaGridDependencySynchronize()`. It may only be used between #: kernel nodes, and must be used with either the #: :py:obj:`~.cudaGraphKernelNodePortProgrammatic` or #: :py:obj:`~.cudaGraphKernelNodePortLaunchCompletion` outgoing port. cudaGraphDependencyTypeProgrammatic = cyruntime.cudaGraphDependencyType_enum.cudaGraphDependencyTypeProgrammatic _dict_cudaGraphDependencyType = dict(((int(v), v) for k, v in cudaGraphDependencyType.__members__.items())) class cudaGraphInstantiateResult(IntEnum): """ Graph instantiation results """ #: Instantiation succeeded cudaGraphInstantiateSuccess = cyruntime.cudaGraphInstantiateResult.cudaGraphInstantiateSuccess #: Instantiation failed for an unexpected reason which is described in #: the return value of the function cudaGraphInstantiateError = cyruntime.cudaGraphInstantiateResult.cudaGraphInstantiateError #: Instantiation failed due to invalid structure, such as cycles cudaGraphInstantiateInvalidStructure = cyruntime.cudaGraphInstantiateResult.cudaGraphInstantiateInvalidStructure #: Instantiation for device launch failed because the graph contained #: an unsupported operation cudaGraphInstantiateNodeOperationNotSupported = cyruntime.cudaGraphInstantiateResult.cudaGraphInstantiateNodeOperationNotSupported #: Instantiation for device launch failed due to the nodes belonging to #: different contexts cudaGraphInstantiateMultipleDevicesNotSupported = cyruntime.cudaGraphInstantiateResult.cudaGraphInstantiateMultipleDevicesNotSupported #: One or more conditional handles are not associated with conditional #: nodes cudaGraphInstantiateConditionalHandleUnused = cyruntime.cudaGraphInstantiateResult.cudaGraphInstantiateConditionalHandleUnused _dict_cudaGraphInstantiateResult = dict(((int(v), v) for k, v in cudaGraphInstantiateResult.__members__.items())) class cudaLaunchMemSyncDomain(IntEnum): """ Memory Synchronization Domain A kernel can be launched in a specified memory synchronization domain that affects all memory operations issued by that kernel. A memory barrier issued in one domain will only order memory operations in that domain, thus eliminating latency increase from memory barriers ordering unrelated traffic. By default, kernels are launched in domain 0. Kernel launched with :py:obj:`~.cudaLaunchMemSyncDomainRemote` will have a different domain ID. User may also alter the domain ID with :py:obj:`~.cudaLaunchMemSyncDomainMap` for a specific stream / graph node / kernel launch. See :py:obj:`~.cudaLaunchAttributeMemSyncDomain`, :py:obj:`~.cudaStreamSetAttribute`, :py:obj:`~.cudaLaunchKernelEx`, :py:obj:`~.cudaGraphKernelNodeSetAttribute`. Memory operations done in kernels launched in different domains are considered system-scope distanced. In other words, a GPU scoped memory synchronization is not sufficient for memory order to be observed by kernels in another memory synchronization domain even if they are on the same GPU. """ #: Launch kernels in the default domain cudaLaunchMemSyncDomainDefault = cyruntime.cudaLaunchMemSyncDomain.cudaLaunchMemSyncDomainDefault #: Launch kernels in the remote domain cudaLaunchMemSyncDomainRemote = cyruntime.cudaLaunchMemSyncDomain.cudaLaunchMemSyncDomainRemote _dict_cudaLaunchMemSyncDomain = dict(((int(v), v) for k, v in cudaLaunchMemSyncDomain.__members__.items())) class cudaLaunchAttributeID(IntEnum): """ Launch attributes enum; used as id field of :py:obj:`~.cudaLaunchAttribute` """ #: Ignored entry, for convenient composition cudaLaunchAttributeIgnore = cyruntime.cudaLaunchAttributeID.cudaLaunchAttributeIgnore #: Valid for streams, graph nodes, launches. See #: :py:obj:`~.cudaLaunchAttributeValue.accessPolicyWindow`. cudaLaunchAttributeAccessPolicyWindow = cyruntime.cudaLaunchAttributeID.cudaLaunchAttributeAccessPolicyWindow #: Valid for graph nodes, launches. See #: :py:obj:`~.cudaLaunchAttributeValue.cooperative`. cudaLaunchAttributeCooperative = cyruntime.cudaLaunchAttributeID.cudaLaunchAttributeCooperative #: Valid for streams. See #: :py:obj:`~.cudaLaunchAttributeValue.syncPolicy`. cudaLaunchAttributeSynchronizationPolicy = cyruntime.cudaLaunchAttributeID.cudaLaunchAttributeSynchronizationPolicy #: Valid for graph nodes, launches. See #: :py:obj:`~.cudaLaunchAttributeValue.clusterDim`. cudaLaunchAttributeClusterDimension = cyruntime.cudaLaunchAttributeID.cudaLaunchAttributeClusterDimension #: Valid for graph nodes, launches. See #: :py:obj:`~.cudaLaunchAttributeValue.clusterSchedulingPolicyPreference`. cudaLaunchAttributeClusterSchedulingPolicyPreference = cyruntime.cudaLaunchAttributeID.cudaLaunchAttributeClusterSchedulingPolicyPreference #: Valid for launches. Setting #: :py:obj:`~.cudaLaunchAttributeValue.programmaticStreamSerializationAllowed` #: to non-0 signals that the kernel will use programmatic means to #: resolve its stream dependency, so that the CUDA runtime should #: opportunistically allow the grid's execution to overlap with the #: previous kernel in the stream, if that kernel requests the overlap. #: The dependent launches can choose to wait on the dependency using #: the programmatic sync (cudaGridDependencySynchronize() or equivalent #: PTX instructions). cudaLaunchAttributeProgrammaticStreamSerialization = cyruntime.cudaLaunchAttributeID.cudaLaunchAttributeProgrammaticStreamSerialization #: Valid for launches. Set #: :py:obj:`~.cudaLaunchAttributeValue.programmaticEvent` to record the #: event. Event recorded through this launch attribute is guaranteed to #: only trigger after all block in the associated kernel trigger the #: event. A block can trigger the event programmatically in a future #: CUDA release. A trigger can also be inserted at the beginning of #: each block's execution if triggerAtBlockStart is set to non-0. The #: dependent launches can choose to wait on the dependency using the #: programmatic sync (cudaGridDependencySynchronize() or equivalent PTX #: instructions). Note that dependents (including the CPU thread #: calling :py:obj:`~.cudaEventSynchronize()`) are not guaranteed to #: observe the release precisely when it is released. For example, #: :py:obj:`~.cudaEventSynchronize()` may only observe the event #: trigger long after the associated kernel has completed. This #: recording type is primarily meant for establishing programmatic #: dependency between device tasks. Note also this type of dependency #: allows, but does not guarantee, concurrent execution of tasks. #: The event supplied must not be an interprocess or interop event. #: The event must disable timing (i.e. must be created with the #: :py:obj:`~.cudaEventDisableTiming` flag set). cudaLaunchAttributeProgrammaticEvent = cyruntime.cudaLaunchAttributeID.cudaLaunchAttributeProgrammaticEvent #: Valid for streams, graph nodes, launches. See #: :py:obj:`~.cudaLaunchAttributeValue.priority`. cudaLaunchAttributePriority = cyruntime.cudaLaunchAttributeID.cudaLaunchAttributePriority #: Valid for streams, graph nodes, launches. See #: :py:obj:`~.cudaLaunchAttributeValue.memSyncDomainMap`. cudaLaunchAttributeMemSyncDomainMap = cyruntime.cudaLaunchAttributeID.cudaLaunchAttributeMemSyncDomainMap #: Valid for streams, graph nodes, launches. See #: :py:obj:`~.cudaLaunchAttributeValue.memSyncDomain`. cudaLaunchAttributeMemSyncDomain = cyruntime.cudaLaunchAttributeID.cudaLaunchAttributeMemSyncDomain #: Valid for graph nodes and launches. Set #: :py:obj:`~.cudaLaunchAttributeValue.preferredClusterDim` to allow #: the kernel launch to specify a preferred substitute cluster #: dimension. Blocks may be grouped according to either the dimensions #: specified with this attribute (grouped into a "preferred substitute #: cluster"), or the one specified with #: :py:obj:`~.cudaLaunchAttributeClusterDimension` attribute (grouped #: into a "regular cluster"). The cluster dimensions of a "preferred #: substitute cluster" shall be an integer multiple greater than zero #: of the regular cluster dimensions. The device will attempt - on a #: best-effort basis - to group thread blocks into preferred clusters #: over grouping them into regular clusters. When it deems necessary #: (primarily when the device temporarily runs out of physical #: resources to launch the larger preferred clusters), the device may #: switch to launch the regular clusters instead to attempt to utilize #: as much of the physical device resources as possible. #: Each type of cluster will have its enumeration / coordinate setup #: as if the grid consists solely of its type of cluster. For example, #: if the preferred substitute cluster dimensions double the regular #: cluster dimensions, there might be simultaneously a regular cluster #: indexed at (1,0,0), and a preferred cluster indexed at (1,0,0). In #: this example, the preferred substitute cluster (1,0,0) replaces #: regular clusters (2,0,0) and (3,0,0) and groups their blocks. #: This attribute will only take effect when a regular cluster #: dimension has been specified. The preferred substitute cluster #: dimension must be an integer multiple greater than zero of the #: regular cluster dimension and must divide the grid. It must also be #: no more than `maxBlocksPerCluster`, if it is set in the kernel's #: `__launch_bounds__`. Otherwise it must be less than the maximum #: value the driver can support. Otherwise, setting this attribute to a #: value physically unable to fit on any particular device is #: permitted. cudaLaunchAttributePreferredClusterDimension = cyruntime.cudaLaunchAttributeID.cudaLaunchAttributePreferredClusterDimension #: Valid for launches. Set #: :py:obj:`~.cudaLaunchAttributeValue.launchCompletionEvent` to record #: the event. #: Nominally, the event is triggered once all blocks of the kernel #: have begun execution. Currently this is a best effort. If a kernel B #: has a launch completion dependency on a kernel A, B may wait until A #: is complete. Alternatively, blocks of B may begin before all blocks #: of A have begun, for example if B can claim execution resources #: unavailable to A (e.g. they run on different GPUs) or if B is a #: higher priority than A. Exercise caution if such an ordering #: inversion could lead to deadlock. #: A launch completion event is nominally similar to a programmatic #: event with `triggerAtBlockStart` set except that it is not visible #: to `cudaGridDependencySynchronize()` and can be used with compute #: capability less than 9.0. #: The event supplied must not be an interprocess or interop event. #: The event must disable timing (i.e. must be created with the #: :py:obj:`~.cudaEventDisableTiming` flag set). cudaLaunchAttributeLaunchCompletionEvent = cyruntime.cudaLaunchAttributeID.cudaLaunchAttributeLaunchCompletionEvent #: Valid for graph nodes, launches. This attribute is graphs-only, and #: passing it to a launch in a non-capturing stream will result in an #: error. #: :cudaLaunchAttributeValue::deviceUpdatableKernelNode::deviceUpdatable #: can only be set to 0 or 1. Setting the field to 1 indicates that the #: corresponding kernel node should be device-updatable. On success, a #: handle will be returned via #: :py:obj:`~.cudaLaunchAttributeValue`::deviceUpdatableKernelNode::devNode #: which can be passed to the various device-side update functions to #: update the node's kernel parameters from within another kernel. For #: more information on the types of device updates that can be made, as #: well as the relevant limitations thereof, see #: :py:obj:`~.cudaGraphKernelNodeUpdatesApply`. #: Nodes which are device-updatable have additional restrictions #: compared to regular kernel nodes. Firstly, device-updatable nodes #: cannot be removed from their graph via #: :py:obj:`~.cudaGraphDestroyNode`. Additionally, once opted-in to #: this functionality, a node cannot opt out, and any attempt to set #: the deviceUpdatable attribute to 0 will result in an error. Device- #: updatable kernel nodes also cannot have their attributes copied #: to/from another kernel node via #: :py:obj:`~.cudaGraphKernelNodeCopyAttributes`. Graphs containing one #: or more device-updatable nodes also do not allow multiple #: instantiation, and neither the graph nor its instantiated version #: can be passed to :py:obj:`~.cudaGraphExecUpdate`. #: If a graph contains device-updatable nodes and updates those nodes #: from the device from within the graph, the graph must be uploaded #: with :py:obj:`~.cuGraphUpload` before it is launched. For such a #: graph, if host-side executable graph updates are made to the device- #: updatable nodes, the graph must be uploaded before it is launched #: again. cudaLaunchAttributeDeviceUpdatableKernelNode = cyruntime.cudaLaunchAttributeID.cudaLaunchAttributeDeviceUpdatableKernelNode #: Valid for launches. On devices where the L1 cache and shared memory #: use the same hardware resources, setting #: :py:obj:`~.cudaLaunchAttributeValue.sharedMemCarveout` to a #: percentage between 0-100 signals sets the shared memory carveout #: preference in percent of the total shared memory for that kernel #: launch. This attribute takes precedence over #: :py:obj:`~.cudaFuncAttributePreferredSharedMemoryCarveout`. This is #: only a hint, and the driver can choose a different configuration if #: required for the launch. cudaLaunchAttributePreferredSharedMemoryCarveout = cyruntime.cudaLaunchAttributeID.cudaLaunchAttributePreferredSharedMemoryCarveout _dict_cudaLaunchAttributeID = dict(((int(v), v) for k, v in cudaLaunchAttributeID.__members__.items())) class cudaAsyncNotificationType(IntEnum): """ Types of async notification that can occur """ #: Sent when the process has exceeded its device memory budget cudaAsyncNotificationTypeOverBudget = cyruntime.cudaAsyncNotificationType_enum.cudaAsyncNotificationTypeOverBudget _dict_cudaAsyncNotificationType = dict(((int(v), v) for k, v in cudaAsyncNotificationType.__members__.items())) class cudaDataType(IntEnum): """""" CUDA_R_32F = cyruntime.cudaDataType_t.CUDA_R_32F CUDA_R_64F = cyruntime.cudaDataType_t.CUDA_R_64F CUDA_R_16F = cyruntime.cudaDataType_t.CUDA_R_16F CUDA_R_8I = cyruntime.cudaDataType_t.CUDA_R_8I CUDA_C_32F = cyruntime.cudaDataType_t.CUDA_C_32F CUDA_C_64F = cyruntime.cudaDataType_t.CUDA_C_64F CUDA_C_16F = cyruntime.cudaDataType_t.CUDA_C_16F CUDA_C_8I = cyruntime.cudaDataType_t.CUDA_C_8I CUDA_R_8U = cyruntime.cudaDataType_t.CUDA_R_8U CUDA_C_8U = cyruntime.cudaDataType_t.CUDA_C_8U CUDA_R_32I = cyruntime.cudaDataType_t.CUDA_R_32I CUDA_C_32I = cyruntime.cudaDataType_t.CUDA_C_32I CUDA_R_32U = cyruntime.cudaDataType_t.CUDA_R_32U CUDA_C_32U = cyruntime.cudaDataType_t.CUDA_C_32U CUDA_R_16BF = cyruntime.cudaDataType_t.CUDA_R_16BF CUDA_C_16BF = cyruntime.cudaDataType_t.CUDA_C_16BF CUDA_R_4I = cyruntime.cudaDataType_t.CUDA_R_4I CUDA_C_4I = cyruntime.cudaDataType_t.CUDA_C_4I CUDA_R_4U = cyruntime.cudaDataType_t.CUDA_R_4U CUDA_C_4U = cyruntime.cudaDataType_t.CUDA_C_4U CUDA_R_16I = cyruntime.cudaDataType_t.CUDA_R_16I CUDA_C_16I = cyruntime.cudaDataType_t.CUDA_C_16I CUDA_R_16U = cyruntime.cudaDataType_t.CUDA_R_16U CUDA_C_16U = cyruntime.cudaDataType_t.CUDA_C_16U CUDA_R_64I = cyruntime.cudaDataType_t.CUDA_R_64I CUDA_C_64I = cyruntime.cudaDataType_t.CUDA_C_64I CUDA_R_64U = cyruntime.cudaDataType_t.CUDA_R_64U CUDA_C_64U = cyruntime.cudaDataType_t.CUDA_C_64U CUDA_R_8F_E4M3 = cyruntime.cudaDataType_t.CUDA_R_8F_E4M3 CUDA_R_8F_UE4M3 = cyruntime.cudaDataType_t.CUDA_R_8F_UE4M3 CUDA_R_8F_E5M2 = cyruntime.cudaDataType_t.CUDA_R_8F_E5M2 CUDA_R_8F_UE8M0 = cyruntime.cudaDataType_t.CUDA_R_8F_UE8M0 CUDA_R_6F_E2M3 = cyruntime.cudaDataType_t.CUDA_R_6F_E2M3 CUDA_R_6F_E3M2 = cyruntime.cudaDataType_t.CUDA_R_6F_E3M2 CUDA_R_4F_E2M1 = cyruntime.cudaDataType_t.CUDA_R_4F_E2M1 _dict_cudaDataType = dict(((int(v), v) for k, v in cudaDataType.__members__.items())) class libraryPropertyType(IntEnum): """""" MAJOR_VERSION = cyruntime.libraryPropertyType_t.MAJOR_VERSION MINOR_VERSION = cyruntime.libraryPropertyType_t.MINOR_VERSION PATCH_LEVEL = cyruntime.libraryPropertyType_t.PATCH_LEVEL _dict_libraryPropertyType = dict(((int(v), v) for k, v in libraryPropertyType.__members__.items())) class cudaEglFrameType(IntEnum): """ CUDA EglFrame type - array or pointer """ #: Frame type CUDA array cudaEglFrameTypeArray = cyruntime.cudaEglFrameType_enum.cudaEglFrameTypeArray #: Frame type CUDA pointer cudaEglFrameTypePitch = cyruntime.cudaEglFrameType_enum.cudaEglFrameTypePitch _dict_cudaEglFrameType = dict(((int(v), v) for k, v in cudaEglFrameType.__members__.items())) class cudaEglResourceLocationFlags(IntEnum): """ Resource location flags- sysmem or vidmem For CUDA context on iGPU, since video and system memory are equivalent - these flags will not have an effect on the execution. For CUDA context on dGPU, applications can use the flag :py:obj:`~.cudaEglResourceLocationFlags` to give a hint about the desired location. :py:obj:`~.cudaEglResourceLocationSysmem` - the frame data is made resident on the system memory to be accessed by CUDA. :py:obj:`~.cudaEglResourceLocationVidmem` - the frame data is made resident on the dedicated video memory to be accessed by CUDA. There may be an additional latency due to new allocation and data migration, if the frame is produced on a different memory. """ #: Resource location sysmem cudaEglResourceLocationSysmem = cyruntime.cudaEglResourceLocationFlags_enum.cudaEglResourceLocationSysmem #: Resource location vidmem cudaEglResourceLocationVidmem = cyruntime.cudaEglResourceLocationFlags_enum.cudaEglResourceLocationVidmem _dict_cudaEglResourceLocationFlags = dict(((int(v), v) for k, v in cudaEglResourceLocationFlags.__members__.items())) class cudaEglColorFormat(IntEnum): """ CUDA EGL Color Format - The different planar and multiplanar formats currently supported for CUDA_EGL interops. """ #: Y, U, V in three surfaces, each in a separate surface, U/V width = #: 1/2 Y width, U/V height = 1/2 Y height. cudaEglColorFormatYUV420Planar = cyruntime.cudaEglColorFormat_enum.cudaEglColorFormatYUV420Planar #: Y, UV in two surfaces (UV as one surface) with VU byte ordering, #: width, height ratio same as YUV420Planar. cudaEglColorFormatYUV420SemiPlanar = cyruntime.cudaEglColorFormat_enum.cudaEglColorFormatYUV420SemiPlanar #: Y, U, V each in a separate surface, U/V width = 1/2 Y width, U/V #: height = Y height. cudaEglColorFormatYUV422Planar = cyruntime.cudaEglColorFormat_enum.cudaEglColorFormatYUV422Planar #: Y, UV in two surfaces with VU byte ordering, width, height ratio #: same as YUV422Planar. cudaEglColorFormatYUV422SemiPlanar = cyruntime.cudaEglColorFormat_enum.cudaEglColorFormatYUV422SemiPlanar #: R/G/B/A four channels in one surface with BGRA byte ordering. cudaEglColorFormatARGB = cyruntime.cudaEglColorFormat_enum.cudaEglColorFormatARGB #: R/G/B/A four channels in one surface with ABGR byte ordering. cudaEglColorFormatRGBA = cyruntime.cudaEglColorFormat_enum.cudaEglColorFormatRGBA #: single luminance channel in one surface. cudaEglColorFormatL = cyruntime.cudaEglColorFormat_enum.cudaEglColorFormatL #: single color channel in one surface. cudaEglColorFormatR = cyruntime.cudaEglColorFormat_enum.cudaEglColorFormatR #: Y, U, V in three surfaces, each in a separate surface, U/V width = Y #: width, U/V height = Y height. cudaEglColorFormatYUV444Planar = cyruntime.cudaEglColorFormat_enum.cudaEglColorFormatYUV444Planar #: Y, UV in two surfaces (UV as one surface) with VU byte ordering, #: width, height ratio same as YUV444Planar. cudaEglColorFormatYUV444SemiPlanar = cyruntime.cudaEglColorFormat_enum.cudaEglColorFormatYUV444SemiPlanar #: Y, U, V in one surface, interleaved as UYVY in one channel. cudaEglColorFormatYUYV422 = cyruntime.cudaEglColorFormat_enum.cudaEglColorFormatYUYV422 #: Y, U, V in one surface, interleaved as YUYV in one channel. cudaEglColorFormatUYVY422 = cyruntime.cudaEglColorFormat_enum.cudaEglColorFormatUYVY422 #: R/G/B/A four channels in one surface with RGBA byte ordering. cudaEglColorFormatABGR = cyruntime.cudaEglColorFormat_enum.cudaEglColorFormatABGR #: R/G/B/A four channels in one surface with ARGB byte ordering. cudaEglColorFormatBGRA = cyruntime.cudaEglColorFormat_enum.cudaEglColorFormatBGRA #: Alpha color format - one channel in one surface. cudaEglColorFormatA = cyruntime.cudaEglColorFormat_enum.cudaEglColorFormatA #: R/G color format - two channels in one surface with GR byte ordering cudaEglColorFormatRG = cyruntime.cudaEglColorFormat_enum.cudaEglColorFormatRG #: Y, U, V, A four channels in one surface, interleaved as VUYA. cudaEglColorFormatAYUV = cyruntime.cudaEglColorFormat_enum.cudaEglColorFormatAYUV #: Y, VU in two surfaces (VU as one surface) with UV byte ordering, U/V #: width = Y width, U/V height = Y height. cudaEglColorFormatYVU444SemiPlanar = cyruntime.cudaEglColorFormat_enum.cudaEglColorFormatYVU444SemiPlanar #: Y, VU in two surfaces (VU as one surface) with UV byte ordering, U/V #: width = 1/2 Y width, U/V height = Y height. cudaEglColorFormatYVU422SemiPlanar = cyruntime.cudaEglColorFormat_enum.cudaEglColorFormatYVU422SemiPlanar #: Y, VU in two surfaces (VU as one surface) with UV byte ordering, U/V #: width = 1/2 Y width, U/V height = 1/2 Y height. cudaEglColorFormatYVU420SemiPlanar = cyruntime.cudaEglColorFormat_enum.cudaEglColorFormatYVU420SemiPlanar #: Y10, V10U10 in two surfaces (VU as one surface) with UV byte #: ordering, U/V width = Y width, U/V height = Y height. cudaEglColorFormatY10V10U10_444SemiPlanar = cyruntime.cudaEglColorFormat_enum.cudaEglColorFormatY10V10U10_444SemiPlanar #: Y10, V10U10 in two surfaces (VU as one surface) with UV byte #: ordering, U/V width = 1/2 Y width, U/V height = 1/2 Y height. cudaEglColorFormatY10V10U10_420SemiPlanar = cyruntime.cudaEglColorFormat_enum.cudaEglColorFormatY10V10U10_420SemiPlanar #: Y12, V12U12 in two surfaces (VU as one surface) with UV byte #: ordering, U/V width = Y width, U/V height = Y height. cudaEglColorFormatY12V12U12_444SemiPlanar = cyruntime.cudaEglColorFormat_enum.cudaEglColorFormatY12V12U12_444SemiPlanar #: Y12, V12U12 in two surfaces (VU as one surface) with UV byte #: ordering, U/V width = 1/2 Y width, U/V height = 1/2 Y height. cudaEglColorFormatY12V12U12_420SemiPlanar = cyruntime.cudaEglColorFormat_enum.cudaEglColorFormatY12V12U12_420SemiPlanar #: Extended Range Y, U, V in one surface, interleaved as YVYU in one #: channel. cudaEglColorFormatVYUY_ER = cyruntime.cudaEglColorFormat_enum.cudaEglColorFormatVYUY_ER #: Extended Range Y, U, V in one surface, interleaved as YUYV in one #: channel. cudaEglColorFormatUYVY_ER = cyruntime.cudaEglColorFormat_enum.cudaEglColorFormatUYVY_ER #: Extended Range Y, U, V in one surface, interleaved as UYVY in one #: channel. cudaEglColorFormatYUYV_ER = cyruntime.cudaEglColorFormat_enum.cudaEglColorFormatYUYV_ER #: Extended Range Y, U, V in one surface, interleaved as VYUY in one #: channel. cudaEglColorFormatYVYU_ER = cyruntime.cudaEglColorFormat_enum.cudaEglColorFormatYVYU_ER #: Extended Range Y, U, V, A four channels in one surface, interleaved #: as AVUY. cudaEglColorFormatYUVA_ER = cyruntime.cudaEglColorFormat_enum.cudaEglColorFormatYUVA_ER #: Extended Range Y, U, V, A four channels in one surface, interleaved #: as VUYA. cudaEglColorFormatAYUV_ER = cyruntime.cudaEglColorFormat_enum.cudaEglColorFormatAYUV_ER #: Extended Range Y, U, V in three surfaces, U/V width = Y width, U/V #: height = Y height. cudaEglColorFormatYUV444Planar_ER = cyruntime.cudaEglColorFormat_enum.cudaEglColorFormatYUV444Planar_ER #: Extended Range Y, U, V in three surfaces, U/V width = 1/2 Y width, #: U/V height = Y height. cudaEglColorFormatYUV422Planar_ER = cyruntime.cudaEglColorFormat_enum.cudaEglColorFormatYUV422Planar_ER #: Extended Range Y, U, V in three surfaces, U/V width = 1/2 Y width, #: U/V height = 1/2 Y height. cudaEglColorFormatYUV420Planar_ER = cyruntime.cudaEglColorFormat_enum.cudaEglColorFormatYUV420Planar_ER #: Extended Range Y, UV in two surfaces (UV as one surface) with VU #: byte ordering, U/V width = Y width, U/V height = Y height. cudaEglColorFormatYUV444SemiPlanar_ER = cyruntime.cudaEglColorFormat_enum.cudaEglColorFormatYUV444SemiPlanar_ER #: Extended Range Y, UV in two surfaces (UV as one surface) with VU #: byte ordering, U/V width = 1/2 Y width, U/V height = Y height. cudaEglColorFormatYUV422SemiPlanar_ER = cyruntime.cudaEglColorFormat_enum.cudaEglColorFormatYUV422SemiPlanar_ER #: Extended Range Y, UV in two surfaces (UV as one surface) with VU #: byte ordering, U/V width = 1/2 Y width, U/V height = 1/2 Y height. cudaEglColorFormatYUV420SemiPlanar_ER = cyruntime.cudaEglColorFormat_enum.cudaEglColorFormatYUV420SemiPlanar_ER #: Extended Range Y, V, U in three surfaces, U/V width = Y width, U/V #: height = Y height. cudaEglColorFormatYVU444Planar_ER = cyruntime.cudaEglColorFormat_enum.cudaEglColorFormatYVU444Planar_ER #: Extended Range Y, V, U in three surfaces, U/V width = 1/2 Y width, #: U/V height = Y height. cudaEglColorFormatYVU422Planar_ER = cyruntime.cudaEglColorFormat_enum.cudaEglColorFormatYVU422Planar_ER #: Extended Range Y, V, U in three surfaces, U/V width = 1/2 Y width, #: U/V height = 1/2 Y height. cudaEglColorFormatYVU420Planar_ER = cyruntime.cudaEglColorFormat_enum.cudaEglColorFormatYVU420Planar_ER #: Extended Range Y, VU in two surfaces (VU as one surface) with UV #: byte ordering, U/V width = Y width, U/V height = Y height. cudaEglColorFormatYVU444SemiPlanar_ER = cyruntime.cudaEglColorFormat_enum.cudaEglColorFormatYVU444SemiPlanar_ER #: Extended Range Y, VU in two surfaces (VU as one surface) with UV #: byte ordering, U/V width = 1/2 Y width, U/V height = Y height. cudaEglColorFormatYVU422SemiPlanar_ER = cyruntime.cudaEglColorFormat_enum.cudaEglColorFormatYVU422SemiPlanar_ER #: Extended Range Y, VU in two surfaces (VU as one surface) with UV #: byte ordering, U/V width = 1/2 Y width, U/V height = 1/2 Y height. cudaEglColorFormatYVU420SemiPlanar_ER = cyruntime.cudaEglColorFormat_enum.cudaEglColorFormatYVU420SemiPlanar_ER #: Bayer format - one channel in one surface with interleaved RGGB #: ordering. cudaEglColorFormatBayerRGGB = cyruntime.cudaEglColorFormat_enum.cudaEglColorFormatBayerRGGB #: Bayer format - one channel in one surface with interleaved BGGR #: ordering. cudaEglColorFormatBayerBGGR = cyruntime.cudaEglColorFormat_enum.cudaEglColorFormatBayerBGGR #: Bayer format - one channel in one surface with interleaved GRBG #: ordering. cudaEglColorFormatBayerGRBG = cyruntime.cudaEglColorFormat_enum.cudaEglColorFormatBayerGRBG #: Bayer format - one channel in one surface with interleaved GBRG #: ordering. cudaEglColorFormatBayerGBRG = cyruntime.cudaEglColorFormat_enum.cudaEglColorFormatBayerGBRG #: Bayer10 format - one channel in one surface with interleaved RGGB #: ordering. Out of 16 bits, 10 bits used 6 bits No-op. cudaEglColorFormatBayer10RGGB = cyruntime.cudaEglColorFormat_enum.cudaEglColorFormatBayer10RGGB #: Bayer10 format - one channel in one surface with interleaved BGGR #: ordering. Out of 16 bits, 10 bits used 6 bits No-op. cudaEglColorFormatBayer10BGGR = cyruntime.cudaEglColorFormat_enum.cudaEglColorFormatBayer10BGGR #: Bayer10 format - one channel in one surface with interleaved GRBG #: ordering. Out of 16 bits, 10 bits used 6 bits No-op. cudaEglColorFormatBayer10GRBG = cyruntime.cudaEglColorFormat_enum.cudaEglColorFormatBayer10GRBG #: Bayer10 format - one channel in one surface with interleaved GBRG #: ordering. Out of 16 bits, 10 bits used 6 bits No-op. cudaEglColorFormatBayer10GBRG = cyruntime.cudaEglColorFormat_enum.cudaEglColorFormatBayer10GBRG #: Bayer12 format - one channel in one surface with interleaved RGGB #: ordering. Out of 16 bits, 12 bits used 4 bits No-op. cudaEglColorFormatBayer12RGGB = cyruntime.cudaEglColorFormat_enum.cudaEglColorFormatBayer12RGGB #: Bayer12 format - one channel in one surface with interleaved BGGR #: ordering. Out of 16 bits, 12 bits used 4 bits No-op. cudaEglColorFormatBayer12BGGR = cyruntime.cudaEglColorFormat_enum.cudaEglColorFormatBayer12BGGR #: Bayer12 format - one channel in one surface with interleaved GRBG #: ordering. Out of 16 bits, 12 bits used 4 bits No-op. cudaEglColorFormatBayer12GRBG = cyruntime.cudaEglColorFormat_enum.cudaEglColorFormatBayer12GRBG #: Bayer12 format - one channel in one surface with interleaved GBRG #: ordering. Out of 16 bits, 12 bits used 4 bits No-op. cudaEglColorFormatBayer12GBRG = cyruntime.cudaEglColorFormat_enum.cudaEglColorFormatBayer12GBRG #: Bayer14 format - one channel in one surface with interleaved RGGB #: ordering. Out of 16 bits, 14 bits used 2 bits No-op. cudaEglColorFormatBayer14RGGB = cyruntime.cudaEglColorFormat_enum.cudaEglColorFormatBayer14RGGB #: Bayer14 format - one channel in one surface with interleaved BGGR #: ordering. Out of 16 bits, 14 bits used 2 bits No-op. cudaEglColorFormatBayer14BGGR = cyruntime.cudaEglColorFormat_enum.cudaEglColorFormatBayer14BGGR #: Bayer14 format - one channel in one surface with interleaved GRBG #: ordering. Out of 16 bits, 14 bits used 2 bits No-op. cudaEglColorFormatBayer14GRBG = cyruntime.cudaEglColorFormat_enum.cudaEglColorFormatBayer14GRBG #: Bayer14 format - one channel in one surface with interleaved GBRG #: ordering. Out of 16 bits, 14 bits used 2 bits No-op. cudaEglColorFormatBayer14GBRG = cyruntime.cudaEglColorFormat_enum.cudaEglColorFormatBayer14GBRG #: Bayer20 format - one channel in one surface with interleaved RGGB #: ordering. Out of 32 bits, 20 bits used 12 bits No-op. cudaEglColorFormatBayer20RGGB = cyruntime.cudaEglColorFormat_enum.cudaEglColorFormatBayer20RGGB #: Bayer20 format - one channel in one surface with interleaved BGGR #: ordering. Out of 32 bits, 20 bits used 12 bits No-op. cudaEglColorFormatBayer20BGGR = cyruntime.cudaEglColorFormat_enum.cudaEglColorFormatBayer20BGGR #: Bayer20 format - one channel in one surface with interleaved GRBG #: ordering. Out of 32 bits, 20 bits used 12 bits No-op. cudaEglColorFormatBayer20GRBG = cyruntime.cudaEglColorFormat_enum.cudaEglColorFormatBayer20GRBG #: Bayer20 format - one channel in one surface with interleaved GBRG #: ordering. Out of 32 bits, 20 bits used 12 bits No-op. cudaEglColorFormatBayer20GBRG = cyruntime.cudaEglColorFormat_enum.cudaEglColorFormatBayer20GBRG #: Y, V, U in three surfaces, each in a separate surface, U/V width = Y #: width, U/V height = Y height. cudaEglColorFormatYVU444Planar = cyruntime.cudaEglColorFormat_enum.cudaEglColorFormatYVU444Planar #: Y, V, U in three surfaces, each in a separate surface, U/V width = #: 1/2 Y width, U/V height = Y height. cudaEglColorFormatYVU422Planar = cyruntime.cudaEglColorFormat_enum.cudaEglColorFormatYVU422Planar #: Y, V, U in three surfaces, each in a separate surface, U/V width = #: 1/2 Y width, U/V height = 1/2 Y height. cudaEglColorFormatYVU420Planar = cyruntime.cudaEglColorFormat_enum.cudaEglColorFormatYVU420Planar #: Nvidia proprietary Bayer ISP format - one channel in one surface #: with interleaved RGGB ordering and mapped to opaque integer #: datatype. cudaEglColorFormatBayerIspRGGB = cyruntime.cudaEglColorFormat_enum.cudaEglColorFormatBayerIspRGGB #: Nvidia proprietary Bayer ISP format - one channel in one surface #: with interleaved BGGR ordering and mapped to opaque integer #: datatype. cudaEglColorFormatBayerIspBGGR = cyruntime.cudaEglColorFormat_enum.cudaEglColorFormatBayerIspBGGR #: Nvidia proprietary Bayer ISP format - one channel in one surface #: with interleaved GRBG ordering and mapped to opaque integer #: datatype. cudaEglColorFormatBayerIspGRBG = cyruntime.cudaEglColorFormat_enum.cudaEglColorFormatBayerIspGRBG #: Nvidia proprietary Bayer ISP format - one channel in one surface #: with interleaved GBRG ordering and mapped to opaque integer #: datatype. cudaEglColorFormatBayerIspGBRG = cyruntime.cudaEglColorFormat_enum.cudaEglColorFormatBayerIspGBRG #: Bayer format - one channel in one surface with interleaved BCCR #: ordering. cudaEglColorFormatBayerBCCR = cyruntime.cudaEglColorFormat_enum.cudaEglColorFormatBayerBCCR #: Bayer format - one channel in one surface with interleaved RCCB #: ordering. cudaEglColorFormatBayerRCCB = cyruntime.cudaEglColorFormat_enum.cudaEglColorFormatBayerRCCB #: Bayer format - one channel in one surface with interleaved CRBC #: ordering. cudaEglColorFormatBayerCRBC = cyruntime.cudaEglColorFormat_enum.cudaEglColorFormatBayerCRBC #: Bayer format - one channel in one surface with interleaved CBRC #: ordering. cudaEglColorFormatBayerCBRC = cyruntime.cudaEglColorFormat_enum.cudaEglColorFormatBayerCBRC #: Bayer10 format - one channel in one surface with interleaved CCCC #: ordering. Out of 16 bits, 10 bits used 6 bits No-op. cudaEglColorFormatBayer10CCCC = cyruntime.cudaEglColorFormat_enum.cudaEglColorFormatBayer10CCCC #: Bayer12 format - one channel in one surface with interleaved BCCR #: ordering. Out of 16 bits, 12 bits used 4 bits No-op. cudaEglColorFormatBayer12BCCR = cyruntime.cudaEglColorFormat_enum.cudaEglColorFormatBayer12BCCR #: Bayer12 format - one channel in one surface with interleaved RCCB #: ordering. Out of 16 bits, 12 bits used 4 bits No-op. cudaEglColorFormatBayer12RCCB = cyruntime.cudaEglColorFormat_enum.cudaEglColorFormatBayer12RCCB #: Bayer12 format - one channel in one surface with interleaved CRBC #: ordering. Out of 16 bits, 12 bits used 4 bits No-op. cudaEglColorFormatBayer12CRBC = cyruntime.cudaEglColorFormat_enum.cudaEglColorFormatBayer12CRBC #: Bayer12 format - one channel in one surface with interleaved CBRC #: ordering. Out of 16 bits, 12 bits used 4 bits No-op. cudaEglColorFormatBayer12CBRC = cyruntime.cudaEglColorFormat_enum.cudaEglColorFormatBayer12CBRC #: Bayer12 format - one channel in one surface with interleaved CCCC #: ordering. Out of 16 bits, 12 bits used 4 bits No-op. cudaEglColorFormatBayer12CCCC = cyruntime.cudaEglColorFormat_enum.cudaEglColorFormatBayer12CCCC #: Color format for single Y plane. cudaEglColorFormatY = cyruntime.cudaEglColorFormat_enum.cudaEglColorFormatY #: Y, UV in two surfaces (UV as one surface) U/V width = 1/2 Y width, #: U/V height = 1/2 Y height. cudaEglColorFormatYUV420SemiPlanar_2020 = cyruntime.cudaEglColorFormat_enum.cudaEglColorFormatYUV420SemiPlanar_2020 #: Y, VU in two surfaces (VU as one surface) U/V width = 1/2 Y width, #: U/V height = 1/2 Y height. cudaEglColorFormatYVU420SemiPlanar_2020 = cyruntime.cudaEglColorFormat_enum.cudaEglColorFormatYVU420SemiPlanar_2020 #: Y, U, V in three surfaces, each in a separate surface, U/V width = #: 1/2 Y width, U/V height = 1/2 Y height. cudaEglColorFormatYUV420Planar_2020 = cyruntime.cudaEglColorFormat_enum.cudaEglColorFormatYUV420Planar_2020 #: Y, V, U in three surfaces, each in a separate surface, U/V width = #: 1/2 Y width, U/V height = 1/2 Y height. cudaEglColorFormatYVU420Planar_2020 = cyruntime.cudaEglColorFormat_enum.cudaEglColorFormatYVU420Planar_2020 #: Y, UV in two surfaces (UV as one surface) U/V width = 1/2 Y width, #: U/V height = 1/2 Y height. cudaEglColorFormatYUV420SemiPlanar_709 = cyruntime.cudaEglColorFormat_enum.cudaEglColorFormatYUV420SemiPlanar_709 #: Y, VU in two surfaces (VU as one surface) U/V width = 1/2 Y width, #: U/V height = 1/2 Y height. cudaEglColorFormatYVU420SemiPlanar_709 = cyruntime.cudaEglColorFormat_enum.cudaEglColorFormatYVU420SemiPlanar_709 #: Y, U, V in three surfaces, each in a separate surface, U/V width = #: 1/2 Y width, U/V height = 1/2 Y height. cudaEglColorFormatYUV420Planar_709 = cyruntime.cudaEglColorFormat_enum.cudaEglColorFormatYUV420Planar_709 #: Y, V, U in three surfaces, each in a separate surface, U/V width = #: 1/2 Y width, U/V height = 1/2 Y height. cudaEglColorFormatYVU420Planar_709 = cyruntime.cudaEglColorFormat_enum.cudaEglColorFormatYVU420Planar_709 #: Y10, V10U10 in two surfaces (VU as one surface) U/V width = 1/2 Y #: width, U/V height = 1/2 Y height. cudaEglColorFormatY10V10U10_420SemiPlanar_709 = cyruntime.cudaEglColorFormat_enum.cudaEglColorFormatY10V10U10_420SemiPlanar_709 #: Y10, V10U10 in two surfaces (VU as one surface) U/V width = 1/2 Y #: width, U/V height = 1/2 Y height. cudaEglColorFormatY10V10U10_420SemiPlanar_2020 = cyruntime.cudaEglColorFormat_enum.cudaEglColorFormatY10V10U10_420SemiPlanar_2020 #: Y10, V10U10 in two surfaces (VU as one surface) U/V width = 1/2 Y #: width, U/V height = Y height. cudaEglColorFormatY10V10U10_422SemiPlanar_2020 = cyruntime.cudaEglColorFormat_enum.cudaEglColorFormatY10V10U10_422SemiPlanar_2020 #: Y10, V10U10 in two surfaces (VU as one surface) U/V width = 1/2 Y #: width, U/V height = Y height. cudaEglColorFormatY10V10U10_422SemiPlanar = cyruntime.cudaEglColorFormat_enum.cudaEglColorFormatY10V10U10_422SemiPlanar #: Y10, V10U10 in two surfaces (VU as one surface) U/V width = 1/2 Y #: width, U/V height = Y height. cudaEglColorFormatY10V10U10_422SemiPlanar_709 = cyruntime.cudaEglColorFormat_enum.cudaEglColorFormatY10V10U10_422SemiPlanar_709 #: Extended Range Color format for single Y plane. cudaEglColorFormatY_ER = cyruntime.cudaEglColorFormat_enum.cudaEglColorFormatY_ER #: Extended Range Color format for single Y plane. cudaEglColorFormatY_709_ER = cyruntime.cudaEglColorFormat_enum.cudaEglColorFormatY_709_ER #: Extended Range Color format for single Y10 plane. cudaEglColorFormatY10_ER = cyruntime.cudaEglColorFormat_enum.cudaEglColorFormatY10_ER #: Extended Range Color format for single Y10 plane. cudaEglColorFormatY10_709_ER = cyruntime.cudaEglColorFormat_enum.cudaEglColorFormatY10_709_ER #: Extended Range Color format for single Y12 plane. cudaEglColorFormatY12_ER = cyruntime.cudaEglColorFormat_enum.cudaEglColorFormatY12_ER #: Extended Range Color format for single Y12 plane. cudaEglColorFormatY12_709_ER = cyruntime.cudaEglColorFormat_enum.cudaEglColorFormatY12_709_ER #: Y, U, V, A four channels in one surface, interleaved as AVUY. cudaEglColorFormatYUVA = cyruntime.cudaEglColorFormat_enum.cudaEglColorFormatYUVA #: Y, U, V in one surface, interleaved as YVYU in one channel. cudaEglColorFormatYVYU = cyruntime.cudaEglColorFormat_enum.cudaEglColorFormatYVYU #: Y, U, V in one surface, interleaved as VYUY in one channel. cudaEglColorFormatVYUY = cyruntime.cudaEglColorFormat_enum.cudaEglColorFormatVYUY #: Extended Range Y10, V10U10 in two surfaces (VU as one surface) U/V #: width = 1/2 Y width, U/V height = 1/2 Y height. cudaEglColorFormatY10V10U10_420SemiPlanar_ER = cyruntime.cudaEglColorFormat_enum.cudaEglColorFormatY10V10U10_420SemiPlanar_ER #: Extended Range Y10, V10U10 in two surfaces (VU as one surface) U/V #: width = 1/2 Y width, U/V height = 1/2 Y height. cudaEglColorFormatY10V10U10_420SemiPlanar_709_ER = cyruntime.cudaEglColorFormat_enum.cudaEglColorFormatY10V10U10_420SemiPlanar_709_ER #: Extended Range Y10, V10U10 in two surfaces (VU as one surface) U/V #: width = Y width, U/V height = Y height. cudaEglColorFormatY10V10U10_444SemiPlanar_ER = cyruntime.cudaEglColorFormat_enum.cudaEglColorFormatY10V10U10_444SemiPlanar_ER #: Extended Range Y10, V10U10 in two surfaces (VU as one surface) U/V #: width = Y width, U/V height = Y height. cudaEglColorFormatY10V10U10_444SemiPlanar_709_ER = cyruntime.cudaEglColorFormat_enum.cudaEglColorFormatY10V10U10_444SemiPlanar_709_ER #: Extended Range Y12, V12U12 in two surfaces (VU as one surface) U/V #: width = 1/2 Y width, U/V height = 1/2 Y height. cudaEglColorFormatY12V12U12_420SemiPlanar_ER = cyruntime.cudaEglColorFormat_enum.cudaEglColorFormatY12V12U12_420SemiPlanar_ER #: Extended Range Y12, V12U12 in two surfaces (VU as one surface) U/V #: width = 1/2 Y width, U/V height = 1/2 Y height. cudaEglColorFormatY12V12U12_420SemiPlanar_709_ER = cyruntime.cudaEglColorFormat_enum.cudaEglColorFormatY12V12U12_420SemiPlanar_709_ER #: Extended Range Y12, V12U12 in two surfaces (VU as one surface) U/V #: width = Y width, U/V height = Y height. cudaEglColorFormatY12V12U12_444SemiPlanar_ER = cyruntime.cudaEglColorFormat_enum.cudaEglColorFormatY12V12U12_444SemiPlanar_ER #: Extended Range Y12, V12U12 in two surfaces (VU as one surface) U/V #: width = Y width, U/V height = Y height. cudaEglColorFormatY12V12U12_444SemiPlanar_709_ER = cyruntime.cudaEglColorFormat_enum.cudaEglColorFormatY12V12U12_444SemiPlanar_709_ER #: Y, U, V in one surface, interleaved as UYVY in one channel. cudaEglColorFormatUYVY709 = cyruntime.cudaEglColorFormat_enum.cudaEglColorFormatUYVY709 #: Extended Range Y, U, V in one surface, interleaved as UYVY in one #: channel. cudaEglColorFormatUYVY709_ER = cyruntime.cudaEglColorFormat_enum.cudaEglColorFormatUYVY709_ER #: Y, U, V in one surface, interleaved as UYVY in one channel. cudaEglColorFormatUYVY2020 = cyruntime.cudaEglColorFormat_enum.cudaEglColorFormatUYVY2020 _dict_cudaEglColorFormat = dict(((int(v), v) for k, v in cudaEglColorFormat.__members__.items())) class cudaChannelFormatKind(IntEnum): """ Channel format kind """ #: Signed channel format cudaChannelFormatKindSigned = cyruntime.cudaChannelFormatKind.cudaChannelFormatKindSigned #: Unsigned channel format cudaChannelFormatKindUnsigned = cyruntime.cudaChannelFormatKind.cudaChannelFormatKindUnsigned #: Float channel format cudaChannelFormatKindFloat = cyruntime.cudaChannelFormatKind.cudaChannelFormatKindFloat #: No channel format cudaChannelFormatKindNone = cyruntime.cudaChannelFormatKind.cudaChannelFormatKindNone #: Unsigned 8-bit integers, planar 4:2:0 YUV format cudaChannelFormatKindNV12 = cyruntime.cudaChannelFormatKind.cudaChannelFormatKindNV12 #: 1 channel unsigned 8-bit normalized integer cudaChannelFormatKindUnsignedNormalized8X1 = cyruntime.cudaChannelFormatKind.cudaChannelFormatKindUnsignedNormalized8X1 #: 2 channel unsigned 8-bit normalized integer cudaChannelFormatKindUnsignedNormalized8X2 = cyruntime.cudaChannelFormatKind.cudaChannelFormatKindUnsignedNormalized8X2 #: 4 channel unsigned 8-bit normalized integer cudaChannelFormatKindUnsignedNormalized8X4 = cyruntime.cudaChannelFormatKind.cudaChannelFormatKindUnsignedNormalized8X4 #: 1 channel unsigned 16-bit normalized integer cudaChannelFormatKindUnsignedNormalized16X1 = cyruntime.cudaChannelFormatKind.cudaChannelFormatKindUnsignedNormalized16X1 #: 2 channel unsigned 16-bit normalized integer cudaChannelFormatKindUnsignedNormalized16X2 = cyruntime.cudaChannelFormatKind.cudaChannelFormatKindUnsignedNormalized16X2 #: 4 channel unsigned 16-bit normalized integer cudaChannelFormatKindUnsignedNormalized16X4 = cyruntime.cudaChannelFormatKind.cudaChannelFormatKindUnsignedNormalized16X4 #: 1 channel signed 8-bit normalized integer cudaChannelFormatKindSignedNormalized8X1 = cyruntime.cudaChannelFormatKind.cudaChannelFormatKindSignedNormalized8X1 #: 2 channel signed 8-bit normalized integer cudaChannelFormatKindSignedNormalized8X2 = cyruntime.cudaChannelFormatKind.cudaChannelFormatKindSignedNormalized8X2 #: 4 channel signed 8-bit normalized integer cudaChannelFormatKindSignedNormalized8X4 = cyruntime.cudaChannelFormatKind.cudaChannelFormatKindSignedNormalized8X4 #: 1 channel signed 16-bit normalized integer cudaChannelFormatKindSignedNormalized16X1 = cyruntime.cudaChannelFormatKind.cudaChannelFormatKindSignedNormalized16X1 #: 2 channel signed 16-bit normalized integer cudaChannelFormatKindSignedNormalized16X2 = cyruntime.cudaChannelFormatKind.cudaChannelFormatKindSignedNormalized16X2 #: 4 channel signed 16-bit normalized integer cudaChannelFormatKindSignedNormalized16X4 = cyruntime.cudaChannelFormatKind.cudaChannelFormatKindSignedNormalized16X4 #: 4 channel unsigned normalized block-compressed (BC1 compression) #: format cudaChannelFormatKindUnsignedBlockCompressed1 = cyruntime.cudaChannelFormatKind.cudaChannelFormatKindUnsignedBlockCompressed1 #: 4 channel unsigned normalized block-compressed (BC1 compression) #: format with sRGB encoding cudaChannelFormatKindUnsignedBlockCompressed1SRGB = cyruntime.cudaChannelFormatKind.cudaChannelFormatKindUnsignedBlockCompressed1SRGB #: 4 channel unsigned normalized block-compressed (BC2 compression) #: format cudaChannelFormatKindUnsignedBlockCompressed2 = cyruntime.cudaChannelFormatKind.cudaChannelFormatKindUnsignedBlockCompressed2 #: 4 channel unsigned normalized block-compressed (BC2 compression) #: format with sRGB encoding cudaChannelFormatKindUnsignedBlockCompressed2SRGB = cyruntime.cudaChannelFormatKind.cudaChannelFormatKindUnsignedBlockCompressed2SRGB #: 4 channel unsigned normalized block-compressed (BC3 compression) #: format cudaChannelFormatKindUnsignedBlockCompressed3 = cyruntime.cudaChannelFormatKind.cudaChannelFormatKindUnsignedBlockCompressed3 #: 4 channel unsigned normalized block-compressed (BC3 compression) #: format with sRGB encoding cudaChannelFormatKindUnsignedBlockCompressed3SRGB = cyruntime.cudaChannelFormatKind.cudaChannelFormatKindUnsignedBlockCompressed3SRGB #: 1 channel unsigned normalized block-compressed (BC4 compression) #: format cudaChannelFormatKindUnsignedBlockCompressed4 = cyruntime.cudaChannelFormatKind.cudaChannelFormatKindUnsignedBlockCompressed4 #: 1 channel signed normalized block-compressed (BC4 compression) #: format cudaChannelFormatKindSignedBlockCompressed4 = cyruntime.cudaChannelFormatKind.cudaChannelFormatKindSignedBlockCompressed4 #: 2 channel unsigned normalized block-compressed (BC5 compression) #: format cudaChannelFormatKindUnsignedBlockCompressed5 = cyruntime.cudaChannelFormatKind.cudaChannelFormatKindUnsignedBlockCompressed5 #: 2 channel signed normalized block-compressed (BC5 compression) #: format cudaChannelFormatKindSignedBlockCompressed5 = cyruntime.cudaChannelFormatKind.cudaChannelFormatKindSignedBlockCompressed5 #: 3 channel unsigned half-float block-compressed (BC6H compression) #: format cudaChannelFormatKindUnsignedBlockCompressed6H = cyruntime.cudaChannelFormatKind.cudaChannelFormatKindUnsignedBlockCompressed6H #: 3 channel signed half-float block-compressed (BC6H compression) #: format cudaChannelFormatKindSignedBlockCompressed6H = cyruntime.cudaChannelFormatKind.cudaChannelFormatKindSignedBlockCompressed6H #: 4 channel unsigned normalized block-compressed (BC7 compression) #: format cudaChannelFormatKindUnsignedBlockCompressed7 = cyruntime.cudaChannelFormatKind.cudaChannelFormatKindUnsignedBlockCompressed7 #: 4 channel unsigned normalized block-compressed (BC7 compression) #: format with sRGB encoding cudaChannelFormatKindUnsignedBlockCompressed7SRGB = cyruntime.cudaChannelFormatKind.cudaChannelFormatKindUnsignedBlockCompressed7SRGB #: 4 channel unsigned normalized (10-bit, 10-bit, 10-bit, 2-bit) format cudaChannelFormatKindUnsignedNormalized1010102 = cyruntime.cudaChannelFormatKind.cudaChannelFormatKindUnsignedNormalized1010102 _dict_cudaChannelFormatKind = dict(((int(v), v) for k, v in cudaChannelFormatKind.__members__.items())) class cudaMemoryType(IntEnum): """ CUDA memory types """ #: Unregistered memory cudaMemoryTypeUnregistered = cyruntime.cudaMemoryType.cudaMemoryTypeUnregistered #: Host memory cudaMemoryTypeHost = cyruntime.cudaMemoryType.cudaMemoryTypeHost #: Device memory cudaMemoryTypeDevice = cyruntime.cudaMemoryType.cudaMemoryTypeDevice #: Managed memory cudaMemoryTypeManaged = cyruntime.cudaMemoryType.cudaMemoryTypeManaged _dict_cudaMemoryType = dict(((int(v), v) for k, v in cudaMemoryType.__members__.items())) class cudaMemcpyKind(IntEnum): """ CUDA memory copy types """ #: Host -> Host cudaMemcpyHostToHost = cyruntime.cudaMemcpyKind.cudaMemcpyHostToHost #: Host -> Device cudaMemcpyHostToDevice = cyruntime.cudaMemcpyKind.cudaMemcpyHostToDevice #: Device -> Host cudaMemcpyDeviceToHost = cyruntime.cudaMemcpyKind.cudaMemcpyDeviceToHost #: Device -> Device cudaMemcpyDeviceToDevice = cyruntime.cudaMemcpyKind.cudaMemcpyDeviceToDevice #: Direction of the transfer is inferred from the pointer values. #: Requires unified virtual addressing cudaMemcpyDefault = cyruntime.cudaMemcpyKind.cudaMemcpyDefault _dict_cudaMemcpyKind = dict(((int(v), v) for k, v in cudaMemcpyKind.__members__.items())) class cudaAccessProperty(IntEnum): """ Specifies performance hint with :py:obj:`~.cudaAccessPolicyWindow` for hitProp and missProp members. """ #: Normal cache persistence. cudaAccessPropertyNormal = cyruntime.cudaAccessProperty.cudaAccessPropertyNormal #: Streaming access is less likely to persit from cache. cudaAccessPropertyStreaming = cyruntime.cudaAccessProperty.cudaAccessPropertyStreaming #: Persisting access is more likely to persist in cache. cudaAccessPropertyPersisting = cyruntime.cudaAccessProperty.cudaAccessPropertyPersisting _dict_cudaAccessProperty = dict(((int(v), v) for k, v in cudaAccessProperty.__members__.items())) class cudaStreamCaptureStatus(IntEnum): """ Possible stream capture statuses returned by :py:obj:`~.cudaStreamIsCapturing` """ #: Stream is not capturing cudaStreamCaptureStatusNone = cyruntime.cudaStreamCaptureStatus.cudaStreamCaptureStatusNone #: Stream is actively capturing cudaStreamCaptureStatusActive = cyruntime.cudaStreamCaptureStatus.cudaStreamCaptureStatusActive #: Stream is part of a capture sequence that has been invalidated, but #: not terminated cudaStreamCaptureStatusInvalidated = cyruntime.cudaStreamCaptureStatus.cudaStreamCaptureStatusInvalidated _dict_cudaStreamCaptureStatus = dict(((int(v), v) for k, v in cudaStreamCaptureStatus.__members__.items())) class cudaStreamCaptureMode(IntEnum): """ Possible modes for stream capture thread interactions. For more details see :py:obj:`~.cudaStreamBeginCapture` and :py:obj:`~.cudaThreadExchangeStreamCaptureMode` """ cudaStreamCaptureModeGlobal = cyruntime.cudaStreamCaptureMode.cudaStreamCaptureModeGlobal cudaStreamCaptureModeThreadLocal = cyruntime.cudaStreamCaptureMode.cudaStreamCaptureModeThreadLocal cudaStreamCaptureModeRelaxed = cyruntime.cudaStreamCaptureMode.cudaStreamCaptureModeRelaxed _dict_cudaStreamCaptureMode = dict(((int(v), v) for k, v in cudaStreamCaptureMode.__members__.items())) class cudaSynchronizationPolicy(IntEnum): """ """ cudaSyncPolicyAuto = cyruntime.cudaSynchronizationPolicy.cudaSyncPolicyAuto cudaSyncPolicySpin = cyruntime.cudaSynchronizationPolicy.cudaSyncPolicySpin cudaSyncPolicyYield = cyruntime.cudaSynchronizationPolicy.cudaSyncPolicyYield cudaSyncPolicyBlockingSync = cyruntime.cudaSynchronizationPolicy.cudaSyncPolicyBlockingSync _dict_cudaSynchronizationPolicy = dict(((int(v), v) for k, v in cudaSynchronizationPolicy.__members__.items())) class cudaClusterSchedulingPolicy(IntEnum): """ Cluster scheduling policies. These may be passed to :py:obj:`~.cudaFuncSetAttribute` """ #: the default policy cudaClusterSchedulingPolicyDefault = cyruntime.cudaClusterSchedulingPolicy.cudaClusterSchedulingPolicyDefault #: spread the blocks within a cluster to the SMs cudaClusterSchedulingPolicySpread = cyruntime.cudaClusterSchedulingPolicy.cudaClusterSchedulingPolicySpread #: allow the hardware to load-balance the blocks in a cluster to the #: SMs cudaClusterSchedulingPolicyLoadBalancing = cyruntime.cudaClusterSchedulingPolicy.cudaClusterSchedulingPolicyLoadBalancing _dict_cudaClusterSchedulingPolicy = dict(((int(v), v) for k, v in cudaClusterSchedulingPolicy.__members__.items())) class cudaStreamUpdateCaptureDependenciesFlags(IntEnum): """ Flags for :py:obj:`~.cudaStreamUpdateCaptureDependencies` """ #: Add new nodes to the dependency set cudaStreamAddCaptureDependencies = cyruntime.cudaStreamUpdateCaptureDependenciesFlags.cudaStreamAddCaptureDependencies #: Replace the dependency set with the new nodes cudaStreamSetCaptureDependencies = cyruntime.cudaStreamUpdateCaptureDependenciesFlags.cudaStreamSetCaptureDependencies _dict_cudaStreamUpdateCaptureDependenciesFlags = dict(((int(v), v) for k, v in cudaStreamUpdateCaptureDependenciesFlags.__members__.items())) class cudaUserObjectFlags(IntEnum): """ Flags for user objects for graphs """ #: Indicates the destructor execution is not synchronized by any CUDA #: handle. cudaUserObjectNoDestructorSync = cyruntime.cudaUserObjectFlags.cudaUserObjectNoDestructorSync _dict_cudaUserObjectFlags = dict(((int(v), v) for k, v in cudaUserObjectFlags.__members__.items())) class cudaUserObjectRetainFlags(IntEnum): """ Flags for retaining user object references for graphs """ #: Transfer references from the caller rather than creating new #: references. cudaGraphUserObjectMove = cyruntime.cudaUserObjectRetainFlags.cudaGraphUserObjectMove _dict_cudaUserObjectRetainFlags = dict(((int(v), v) for k, v in cudaUserObjectRetainFlags.__members__.items())) class cudaGraphicsRegisterFlags(IntEnum): """ CUDA graphics interop register flags """ #: Default cudaGraphicsRegisterFlagsNone = cyruntime.cudaGraphicsRegisterFlags.cudaGraphicsRegisterFlagsNone #: CUDA will not write to this resource cudaGraphicsRegisterFlagsReadOnly = cyruntime.cudaGraphicsRegisterFlags.cudaGraphicsRegisterFlagsReadOnly #: CUDA will only write to and will not read from this resource cudaGraphicsRegisterFlagsWriteDiscard = cyruntime.cudaGraphicsRegisterFlags.cudaGraphicsRegisterFlagsWriteDiscard #: CUDA will bind this resource to a surface reference cudaGraphicsRegisterFlagsSurfaceLoadStore = cyruntime.cudaGraphicsRegisterFlags.cudaGraphicsRegisterFlagsSurfaceLoadStore #: CUDA will perform texture gather operations on this resource cudaGraphicsRegisterFlagsTextureGather = cyruntime.cudaGraphicsRegisterFlags.cudaGraphicsRegisterFlagsTextureGather _dict_cudaGraphicsRegisterFlags = dict(((int(v), v) for k, v in cudaGraphicsRegisterFlags.__members__.items())) class cudaGraphicsMapFlags(IntEnum): """ CUDA graphics interop map flags """ #: Default; Assume resource can be read/written cudaGraphicsMapFlagsNone = cyruntime.cudaGraphicsMapFlags.cudaGraphicsMapFlagsNone #: CUDA will not write to this resource cudaGraphicsMapFlagsReadOnly = cyruntime.cudaGraphicsMapFlags.cudaGraphicsMapFlagsReadOnly #: CUDA will only write to and will not read from this resource cudaGraphicsMapFlagsWriteDiscard = cyruntime.cudaGraphicsMapFlags.cudaGraphicsMapFlagsWriteDiscard _dict_cudaGraphicsMapFlags = dict(((int(v), v) for k, v in cudaGraphicsMapFlags.__members__.items())) class cudaGraphicsCubeFace(IntEnum): """ CUDA graphics interop array indices for cube maps """ #: Positive X face of cubemap cudaGraphicsCubeFacePositiveX = cyruntime.cudaGraphicsCubeFace.cudaGraphicsCubeFacePositiveX #: Negative X face of cubemap cudaGraphicsCubeFaceNegativeX = cyruntime.cudaGraphicsCubeFace.cudaGraphicsCubeFaceNegativeX #: Positive Y face of cubemap cudaGraphicsCubeFacePositiveY = cyruntime.cudaGraphicsCubeFace.cudaGraphicsCubeFacePositiveY #: Negative Y face of cubemap cudaGraphicsCubeFaceNegativeY = cyruntime.cudaGraphicsCubeFace.cudaGraphicsCubeFaceNegativeY #: Positive Z face of cubemap cudaGraphicsCubeFacePositiveZ = cyruntime.cudaGraphicsCubeFace.cudaGraphicsCubeFacePositiveZ #: Negative Z face of cubemap cudaGraphicsCubeFaceNegativeZ = cyruntime.cudaGraphicsCubeFace.cudaGraphicsCubeFaceNegativeZ _dict_cudaGraphicsCubeFace = dict(((int(v), v) for k, v in cudaGraphicsCubeFace.__members__.items())) class cudaResourceType(IntEnum): """ CUDA resource types """ #: Array resource cudaResourceTypeArray = cyruntime.cudaResourceType.cudaResourceTypeArray #: Mipmapped array resource cudaResourceTypeMipmappedArray = cyruntime.cudaResourceType.cudaResourceTypeMipmappedArray #: Linear resource cudaResourceTypeLinear = cyruntime.cudaResourceType.cudaResourceTypeLinear #: Pitch 2D resource cudaResourceTypePitch2D = cyruntime.cudaResourceType.cudaResourceTypePitch2D _dict_cudaResourceType = dict(((int(v), v) for k, v in cudaResourceType.__members__.items())) class cudaResourceViewFormat(IntEnum): """ CUDA texture resource view formats """ #: No resource view format (use underlying resource format) cudaResViewFormatNone = cyruntime.cudaResourceViewFormat.cudaResViewFormatNone #: 1 channel unsigned 8-bit integers cudaResViewFormatUnsignedChar1 = cyruntime.cudaResourceViewFormat.cudaResViewFormatUnsignedChar1 #: 2 channel unsigned 8-bit integers cudaResViewFormatUnsignedChar2 = cyruntime.cudaResourceViewFormat.cudaResViewFormatUnsignedChar2 #: 4 channel unsigned 8-bit integers cudaResViewFormatUnsignedChar4 = cyruntime.cudaResourceViewFormat.cudaResViewFormatUnsignedChar4 #: 1 channel signed 8-bit integers cudaResViewFormatSignedChar1 = cyruntime.cudaResourceViewFormat.cudaResViewFormatSignedChar1 #: 2 channel signed 8-bit integers cudaResViewFormatSignedChar2 = cyruntime.cudaResourceViewFormat.cudaResViewFormatSignedChar2 #: 4 channel signed 8-bit integers cudaResViewFormatSignedChar4 = cyruntime.cudaResourceViewFormat.cudaResViewFormatSignedChar4 #: 1 channel unsigned 16-bit integers cudaResViewFormatUnsignedShort1 = cyruntime.cudaResourceViewFormat.cudaResViewFormatUnsignedShort1 #: 2 channel unsigned 16-bit integers cudaResViewFormatUnsignedShort2 = cyruntime.cudaResourceViewFormat.cudaResViewFormatUnsignedShort2 #: 4 channel unsigned 16-bit integers cudaResViewFormatUnsignedShort4 = cyruntime.cudaResourceViewFormat.cudaResViewFormatUnsignedShort4 #: 1 channel signed 16-bit integers cudaResViewFormatSignedShort1 = cyruntime.cudaResourceViewFormat.cudaResViewFormatSignedShort1 #: 2 channel signed 16-bit integers cudaResViewFormatSignedShort2 = cyruntime.cudaResourceViewFormat.cudaResViewFormatSignedShort2 #: 4 channel signed 16-bit integers cudaResViewFormatSignedShort4 = cyruntime.cudaResourceViewFormat.cudaResViewFormatSignedShort4 #: 1 channel unsigned 32-bit integers cudaResViewFormatUnsignedInt1 = cyruntime.cudaResourceViewFormat.cudaResViewFormatUnsignedInt1 #: 2 channel unsigned 32-bit integers cudaResViewFormatUnsignedInt2 = cyruntime.cudaResourceViewFormat.cudaResViewFormatUnsignedInt2 #: 4 channel unsigned 32-bit integers cudaResViewFormatUnsignedInt4 = cyruntime.cudaResourceViewFormat.cudaResViewFormatUnsignedInt4 #: 1 channel signed 32-bit integers cudaResViewFormatSignedInt1 = cyruntime.cudaResourceViewFormat.cudaResViewFormatSignedInt1 #: 2 channel signed 32-bit integers cudaResViewFormatSignedInt2 = cyruntime.cudaResourceViewFormat.cudaResViewFormatSignedInt2 #: 4 channel signed 32-bit integers cudaResViewFormatSignedInt4 = cyruntime.cudaResourceViewFormat.cudaResViewFormatSignedInt4 #: 1 channel 16-bit floating point cudaResViewFormatHalf1 = cyruntime.cudaResourceViewFormat.cudaResViewFormatHalf1 #: 2 channel 16-bit floating point cudaResViewFormatHalf2 = cyruntime.cudaResourceViewFormat.cudaResViewFormatHalf2 #: 4 channel 16-bit floating point cudaResViewFormatHalf4 = cyruntime.cudaResourceViewFormat.cudaResViewFormatHalf4 #: 1 channel 32-bit floating point cudaResViewFormatFloat1 = cyruntime.cudaResourceViewFormat.cudaResViewFormatFloat1 #: 2 channel 32-bit floating point cudaResViewFormatFloat2 = cyruntime.cudaResourceViewFormat.cudaResViewFormatFloat2 #: 4 channel 32-bit floating point cudaResViewFormatFloat4 = cyruntime.cudaResourceViewFormat.cudaResViewFormatFloat4 #: Block compressed 1 cudaResViewFormatUnsignedBlockCompressed1 = cyruntime.cudaResourceViewFormat.cudaResViewFormatUnsignedBlockCompressed1 #: Block compressed 2 cudaResViewFormatUnsignedBlockCompressed2 = cyruntime.cudaResourceViewFormat.cudaResViewFormatUnsignedBlockCompressed2 #: Block compressed 3 cudaResViewFormatUnsignedBlockCompressed3 = cyruntime.cudaResourceViewFormat.cudaResViewFormatUnsignedBlockCompressed3 #: Block compressed 4 unsigned cudaResViewFormatUnsignedBlockCompressed4 = cyruntime.cudaResourceViewFormat.cudaResViewFormatUnsignedBlockCompressed4 #: Block compressed 4 signed cudaResViewFormatSignedBlockCompressed4 = cyruntime.cudaResourceViewFormat.cudaResViewFormatSignedBlockCompressed4 #: Block compressed 5 unsigned cudaResViewFormatUnsignedBlockCompressed5 = cyruntime.cudaResourceViewFormat.cudaResViewFormatUnsignedBlockCompressed5 #: Block compressed 5 signed cudaResViewFormatSignedBlockCompressed5 = cyruntime.cudaResourceViewFormat.cudaResViewFormatSignedBlockCompressed5 #: Block compressed 6 unsigned half-float cudaResViewFormatUnsignedBlockCompressed6H = cyruntime.cudaResourceViewFormat.cudaResViewFormatUnsignedBlockCompressed6H #: Block compressed 6 signed half-float cudaResViewFormatSignedBlockCompressed6H = cyruntime.cudaResourceViewFormat.cudaResViewFormatSignedBlockCompressed6H #: Block compressed 7 cudaResViewFormatUnsignedBlockCompressed7 = cyruntime.cudaResourceViewFormat.cudaResViewFormatUnsignedBlockCompressed7 _dict_cudaResourceViewFormat = dict(((int(v), v) for k, v in cudaResourceViewFormat.__members__.items())) class cudaFuncAttribute(IntEnum): """ CUDA function attributes that can be set using :py:obj:`~.cudaFuncSetAttribute` """ #: Maximum dynamic shared memory size cudaFuncAttributeMaxDynamicSharedMemorySize = cyruntime.cudaFuncAttribute.cudaFuncAttributeMaxDynamicSharedMemorySize #: Preferred shared memory-L1 cache split cudaFuncAttributePreferredSharedMemoryCarveout = cyruntime.cudaFuncAttribute.cudaFuncAttributePreferredSharedMemoryCarveout #: Indicator to enforce valid cluster dimension specification on kernel #: launch cudaFuncAttributeClusterDimMustBeSet = cyruntime.cudaFuncAttribute.cudaFuncAttributeClusterDimMustBeSet #: Required cluster width cudaFuncAttributeRequiredClusterWidth = cyruntime.cudaFuncAttribute.cudaFuncAttributeRequiredClusterWidth #: Required cluster height cudaFuncAttributeRequiredClusterHeight = cyruntime.cudaFuncAttribute.cudaFuncAttributeRequiredClusterHeight #: Required cluster depth cudaFuncAttributeRequiredClusterDepth = cyruntime.cudaFuncAttribute.cudaFuncAttributeRequiredClusterDepth #: Whether non-portable cluster scheduling policy is supported cudaFuncAttributeNonPortableClusterSizeAllowed = cyruntime.cudaFuncAttribute.cudaFuncAttributeNonPortableClusterSizeAllowed #: Required cluster scheduling policy preference cudaFuncAttributeClusterSchedulingPolicyPreference = cyruntime.cudaFuncAttribute.cudaFuncAttributeClusterSchedulingPolicyPreference cudaFuncAttributeMax = cyruntime.cudaFuncAttribute.cudaFuncAttributeMax _dict_cudaFuncAttribute = dict(((int(v), v) for k, v in cudaFuncAttribute.__members__.items())) class cudaFuncCache(IntEnum): """ CUDA function cache configurations """ #: Default function cache configuration, no preference cudaFuncCachePreferNone = cyruntime.cudaFuncCache.cudaFuncCachePreferNone #: Prefer larger shared memory and smaller L1 cache cudaFuncCachePreferShared = cyruntime.cudaFuncCache.cudaFuncCachePreferShared #: Prefer larger L1 cache and smaller shared memory cudaFuncCachePreferL1 = cyruntime.cudaFuncCache.cudaFuncCachePreferL1 #: Prefer equal size L1 cache and shared memory cudaFuncCachePreferEqual = cyruntime.cudaFuncCache.cudaFuncCachePreferEqual _dict_cudaFuncCache = dict(((int(v), v) for k, v in cudaFuncCache.__members__.items())) class cudaSharedMemConfig(IntEnum): """ CUDA shared memory configuration [Deprecated] """ cudaSharedMemBankSizeDefault = cyruntime.cudaSharedMemConfig.cudaSharedMemBankSizeDefault cudaSharedMemBankSizeFourByte = cyruntime.cudaSharedMemConfig.cudaSharedMemBankSizeFourByte cudaSharedMemBankSizeEightByte = cyruntime.cudaSharedMemConfig.cudaSharedMemBankSizeEightByte _dict_cudaSharedMemConfig = dict(((int(v), v) for k, v in cudaSharedMemConfig.__members__.items())) class cudaSharedCarveout(IntEnum): """ Shared memory carveout configurations. These may be passed to cudaFuncSetAttribute """ #: No preference for shared memory or L1 (default) cudaSharedmemCarveoutDefault = cyruntime.cudaSharedCarveout.cudaSharedmemCarveoutDefault #: Prefer maximum available L1 cache, minimum shared memory cudaSharedmemCarveoutMaxL1 = cyruntime.cudaSharedCarveout.cudaSharedmemCarveoutMaxL1 #: Prefer maximum available shared memory, minimum L1 cache cudaSharedmemCarveoutMaxShared = cyruntime.cudaSharedCarveout.cudaSharedmemCarveoutMaxShared _dict_cudaSharedCarveout = dict(((int(v), v) for k, v in cudaSharedCarveout.__members__.items())) class cudaComputeMode(IntEnum): """ CUDA device compute modes """ #: Default compute mode (Multiple threads can use #: :py:obj:`~.cudaSetDevice()` with this device) cudaComputeModeDefault = cyruntime.cudaComputeMode.cudaComputeModeDefault #: Compute-exclusive-thread mode (Only one thread in one process will #: be able to use :py:obj:`~.cudaSetDevice()` with this device) cudaComputeModeExclusive = cyruntime.cudaComputeMode.cudaComputeModeExclusive #: Compute-prohibited mode (No threads can use #: :py:obj:`~.cudaSetDevice()` with this device) cudaComputeModeProhibited = cyruntime.cudaComputeMode.cudaComputeModeProhibited #: Compute-exclusive-process mode (Many threads in one process will be #: able to use :py:obj:`~.cudaSetDevice()` with this device) cudaComputeModeExclusiveProcess = cyruntime.cudaComputeMode.cudaComputeModeExclusiveProcess _dict_cudaComputeMode = dict(((int(v), v) for k, v in cudaComputeMode.__members__.items())) class cudaLimit(IntEnum): """ CUDA Limits """ #: GPU thread stack size cudaLimitStackSize = cyruntime.cudaLimit.cudaLimitStackSize #: GPU printf FIFO size cudaLimitPrintfFifoSize = cyruntime.cudaLimit.cudaLimitPrintfFifoSize #: GPU malloc heap size cudaLimitMallocHeapSize = cyruntime.cudaLimit.cudaLimitMallocHeapSize #: GPU device runtime synchronize depth cudaLimitDevRuntimeSyncDepth = cyruntime.cudaLimit.cudaLimitDevRuntimeSyncDepth #: GPU device runtime pending launch count cudaLimitDevRuntimePendingLaunchCount = cyruntime.cudaLimit.cudaLimitDevRuntimePendingLaunchCount #: A value between 0 and 128 that indicates the maximum fetch #: granularity of L2 (in Bytes). This is a hint cudaLimitMaxL2FetchGranularity = cyruntime.cudaLimit.cudaLimitMaxL2FetchGranularity #: A size in bytes for L2 persisting lines cache size cudaLimitPersistingL2CacheSize = cyruntime.cudaLimit.cudaLimitPersistingL2CacheSize _dict_cudaLimit = dict(((int(v), v) for k, v in cudaLimit.__members__.items())) class cudaMemoryAdvise(IntEnum): """ CUDA Memory Advise values """ #: Data will mostly be read and only occassionally be written to cudaMemAdviseSetReadMostly = cyruntime.cudaMemoryAdvise.cudaMemAdviseSetReadMostly #: Undo the effect of :py:obj:`~.cudaMemAdviseSetReadMostly` cudaMemAdviseUnsetReadMostly = cyruntime.cudaMemoryAdvise.cudaMemAdviseUnsetReadMostly #: Set the preferred location for the data as the specified device cudaMemAdviseSetPreferredLocation = cyruntime.cudaMemoryAdvise.cudaMemAdviseSetPreferredLocation #: Clear the preferred location for the data cudaMemAdviseUnsetPreferredLocation = cyruntime.cudaMemoryAdvise.cudaMemAdviseUnsetPreferredLocation #: Data will be accessed by the specified device, so prevent page #: faults as much as possible cudaMemAdviseSetAccessedBy = cyruntime.cudaMemoryAdvise.cudaMemAdviseSetAccessedBy #: Let the Unified Memory subsystem decide on the page faulting policy #: for the specified device cudaMemAdviseUnsetAccessedBy = cyruntime.cudaMemoryAdvise.cudaMemAdviseUnsetAccessedBy _dict_cudaMemoryAdvise = dict(((int(v), v) for k, v in cudaMemoryAdvise.__members__.items())) class cudaMemRangeAttribute(IntEnum): """ CUDA range attributes """ #: Whether the range will mostly be read and only occassionally be #: written to cudaMemRangeAttributeReadMostly = cyruntime.cudaMemRangeAttribute.cudaMemRangeAttributeReadMostly #: The preferred location of the range cudaMemRangeAttributePreferredLocation = cyruntime.cudaMemRangeAttribute.cudaMemRangeAttributePreferredLocation #: Memory range has :py:obj:`~.cudaMemAdviseSetAccessedBy` set for #: specified device cudaMemRangeAttributeAccessedBy = cyruntime.cudaMemRangeAttribute.cudaMemRangeAttributeAccessedBy #: The last location to which the range was prefetched cudaMemRangeAttributeLastPrefetchLocation = cyruntime.cudaMemRangeAttribute.cudaMemRangeAttributeLastPrefetchLocation #: The preferred location type of the range cudaMemRangeAttributePreferredLocationType = cyruntime.cudaMemRangeAttribute.cudaMemRangeAttributePreferredLocationType #: The preferred location id of the range cudaMemRangeAttributePreferredLocationId = cyruntime.cudaMemRangeAttribute.cudaMemRangeAttributePreferredLocationId #: The last location type to which the range was prefetched cudaMemRangeAttributeLastPrefetchLocationType = cyruntime.cudaMemRangeAttribute.cudaMemRangeAttributeLastPrefetchLocationType #: The last location id to which the range was prefetched cudaMemRangeAttributeLastPrefetchLocationId = cyruntime.cudaMemRangeAttribute.cudaMemRangeAttributeLastPrefetchLocationId _dict_cudaMemRangeAttribute = dict(((int(v), v) for k, v in cudaMemRangeAttribute.__members__.items())) class cudaFlushGPUDirectRDMAWritesOptions(IntEnum): """ CUDA GPUDirect RDMA flush writes APIs supported on the device """ #: :py:obj:`~.cudaDeviceFlushGPUDirectRDMAWrites()` and its CUDA Driver #: API counterpart are supported on the device. cudaFlushGPUDirectRDMAWritesOptionHost = cyruntime.cudaFlushGPUDirectRDMAWritesOptions.cudaFlushGPUDirectRDMAWritesOptionHost #: The :py:obj:`~.CU_STREAM_WAIT_VALUE_FLUSH` flag and the #: :py:obj:`~.CU_STREAM_MEM_OP_FLUSH_REMOTE_WRITES` MemOp are supported #: on the CUDA device. cudaFlushGPUDirectRDMAWritesOptionMemOps = cyruntime.cudaFlushGPUDirectRDMAWritesOptions.cudaFlushGPUDirectRDMAWritesOptionMemOps _dict_cudaFlushGPUDirectRDMAWritesOptions = dict(((int(v), v) for k, v in cudaFlushGPUDirectRDMAWritesOptions.__members__.items())) class cudaGPUDirectRDMAWritesOrdering(IntEnum): """ CUDA GPUDirect RDMA flush writes ordering features of the device """ #: The device does not natively support ordering of GPUDirect RDMA #: writes. :py:obj:`~.cudaFlushGPUDirectRDMAWrites()` can be leveraged #: if supported. cudaGPUDirectRDMAWritesOrderingNone = cyruntime.cudaGPUDirectRDMAWritesOrdering.cudaGPUDirectRDMAWritesOrderingNone #: Natively, the device can consistently consume GPUDirect RDMA writes, #: although other CUDA devices may not. cudaGPUDirectRDMAWritesOrderingOwner = cyruntime.cudaGPUDirectRDMAWritesOrdering.cudaGPUDirectRDMAWritesOrderingOwner #: Any CUDA device in the system can consistently consume GPUDirect #: RDMA writes to this device. cudaGPUDirectRDMAWritesOrderingAllDevices = cyruntime.cudaGPUDirectRDMAWritesOrdering.cudaGPUDirectRDMAWritesOrderingAllDevices _dict_cudaGPUDirectRDMAWritesOrdering = dict(((int(v), v) for k, v in cudaGPUDirectRDMAWritesOrdering.__members__.items())) class cudaFlushGPUDirectRDMAWritesScope(IntEnum): """ CUDA GPUDirect RDMA flush writes scopes """ #: Blocks until remote writes are visible to the CUDA device context #: owning the data. cudaFlushGPUDirectRDMAWritesToOwner = cyruntime.cudaFlushGPUDirectRDMAWritesScope.cudaFlushGPUDirectRDMAWritesToOwner #: Blocks until remote writes are visible to all CUDA device contexts. cudaFlushGPUDirectRDMAWritesToAllDevices = cyruntime.cudaFlushGPUDirectRDMAWritesScope.cudaFlushGPUDirectRDMAWritesToAllDevices _dict_cudaFlushGPUDirectRDMAWritesScope = dict(((int(v), v) for k, v in cudaFlushGPUDirectRDMAWritesScope.__members__.items())) class cudaFlushGPUDirectRDMAWritesTarget(IntEnum): """ CUDA GPUDirect RDMA flush writes targets """ #: Sets the target for :py:obj:`~.cudaDeviceFlushGPUDirectRDMAWrites()` #: to the currently active CUDA device context. cudaFlushGPUDirectRDMAWritesTargetCurrentDevice = cyruntime.cudaFlushGPUDirectRDMAWritesTarget.cudaFlushGPUDirectRDMAWritesTargetCurrentDevice _dict_cudaFlushGPUDirectRDMAWritesTarget = dict(((int(v), v) for k, v in cudaFlushGPUDirectRDMAWritesTarget.__members__.items())) class cudaDeviceAttr(IntEnum): """ CUDA device attributes """ #: Maximum number of threads per block cudaDevAttrMaxThreadsPerBlock = cyruntime.cudaDeviceAttr.cudaDevAttrMaxThreadsPerBlock #: Maximum block dimension X cudaDevAttrMaxBlockDimX = cyruntime.cudaDeviceAttr.cudaDevAttrMaxBlockDimX #: Maximum block dimension Y cudaDevAttrMaxBlockDimY = cyruntime.cudaDeviceAttr.cudaDevAttrMaxBlockDimY #: Maximum block dimension Z cudaDevAttrMaxBlockDimZ = cyruntime.cudaDeviceAttr.cudaDevAttrMaxBlockDimZ #: Maximum grid dimension X cudaDevAttrMaxGridDimX = cyruntime.cudaDeviceAttr.cudaDevAttrMaxGridDimX #: Maximum grid dimension Y cudaDevAttrMaxGridDimY = cyruntime.cudaDeviceAttr.cudaDevAttrMaxGridDimY #: Maximum grid dimension Z cudaDevAttrMaxGridDimZ = cyruntime.cudaDeviceAttr.cudaDevAttrMaxGridDimZ #: Maximum shared memory available per block in bytes cudaDevAttrMaxSharedMemoryPerBlock = cyruntime.cudaDeviceAttr.cudaDevAttrMaxSharedMemoryPerBlock #: Memory available on device for constant variables in a CUDA C kernel #: in bytes cudaDevAttrTotalConstantMemory = cyruntime.cudaDeviceAttr.cudaDevAttrTotalConstantMemory #: Warp size in threads cudaDevAttrWarpSize = cyruntime.cudaDeviceAttr.cudaDevAttrWarpSize #: Maximum pitch in bytes allowed by memory copies cudaDevAttrMaxPitch = cyruntime.cudaDeviceAttr.cudaDevAttrMaxPitch #: Maximum number of 32-bit registers available per block cudaDevAttrMaxRegistersPerBlock = cyruntime.cudaDeviceAttr.cudaDevAttrMaxRegistersPerBlock #: Peak clock frequency in kilohertz cudaDevAttrClockRate = cyruntime.cudaDeviceAttr.cudaDevAttrClockRate #: Alignment requirement for textures cudaDevAttrTextureAlignment = cyruntime.cudaDeviceAttr.cudaDevAttrTextureAlignment #: Device can possibly copy memory and execute a kernel concurrently cudaDevAttrGpuOverlap = cyruntime.cudaDeviceAttr.cudaDevAttrGpuOverlap #: Number of multiprocessors on device cudaDevAttrMultiProcessorCount = cyruntime.cudaDeviceAttr.cudaDevAttrMultiProcessorCount #: Specifies whether there is a run time limit on kernels cudaDevAttrKernelExecTimeout = cyruntime.cudaDeviceAttr.cudaDevAttrKernelExecTimeout #: Device is integrated with host memory cudaDevAttrIntegrated = cyruntime.cudaDeviceAttr.cudaDevAttrIntegrated #: Device can map host memory into CUDA address space cudaDevAttrCanMapHostMemory = cyruntime.cudaDeviceAttr.cudaDevAttrCanMapHostMemory #: Compute mode (See :py:obj:`~.cudaComputeMode` for details) cudaDevAttrComputeMode = cyruntime.cudaDeviceAttr.cudaDevAttrComputeMode #: Maximum 1D texture width cudaDevAttrMaxTexture1DWidth = cyruntime.cudaDeviceAttr.cudaDevAttrMaxTexture1DWidth #: Maximum 2D texture width cudaDevAttrMaxTexture2DWidth = cyruntime.cudaDeviceAttr.cudaDevAttrMaxTexture2DWidth #: Maximum 2D texture height cudaDevAttrMaxTexture2DHeight = cyruntime.cudaDeviceAttr.cudaDevAttrMaxTexture2DHeight #: Maximum 3D texture width cudaDevAttrMaxTexture3DWidth = cyruntime.cudaDeviceAttr.cudaDevAttrMaxTexture3DWidth #: Maximum 3D texture height cudaDevAttrMaxTexture3DHeight = cyruntime.cudaDeviceAttr.cudaDevAttrMaxTexture3DHeight #: Maximum 3D texture depth cudaDevAttrMaxTexture3DDepth = cyruntime.cudaDeviceAttr.cudaDevAttrMaxTexture3DDepth #: Maximum 2D layered texture width cudaDevAttrMaxTexture2DLayeredWidth = cyruntime.cudaDeviceAttr.cudaDevAttrMaxTexture2DLayeredWidth #: Maximum 2D layered texture height cudaDevAttrMaxTexture2DLayeredHeight = cyruntime.cudaDeviceAttr.cudaDevAttrMaxTexture2DLayeredHeight #: Maximum layers in a 2D layered texture cudaDevAttrMaxTexture2DLayeredLayers = cyruntime.cudaDeviceAttr.cudaDevAttrMaxTexture2DLayeredLayers #: Alignment requirement for surfaces cudaDevAttrSurfaceAlignment = cyruntime.cudaDeviceAttr.cudaDevAttrSurfaceAlignment #: Device can possibly execute multiple kernels concurrently cudaDevAttrConcurrentKernels = cyruntime.cudaDeviceAttr.cudaDevAttrConcurrentKernels #: Device has ECC support enabled cudaDevAttrEccEnabled = cyruntime.cudaDeviceAttr.cudaDevAttrEccEnabled #: PCI bus ID of the device cudaDevAttrPciBusId = cyruntime.cudaDeviceAttr.cudaDevAttrPciBusId #: PCI device ID of the device cudaDevAttrPciDeviceId = cyruntime.cudaDeviceAttr.cudaDevAttrPciDeviceId #: Device is using TCC driver model cudaDevAttrTccDriver = cyruntime.cudaDeviceAttr.cudaDevAttrTccDriver #: Peak memory clock frequency in kilohertz cudaDevAttrMemoryClockRate = cyruntime.cudaDeviceAttr.cudaDevAttrMemoryClockRate #: Global memory bus width in bits cudaDevAttrGlobalMemoryBusWidth = cyruntime.cudaDeviceAttr.cudaDevAttrGlobalMemoryBusWidth #: Size of L2 cache in bytes cudaDevAttrL2CacheSize = cyruntime.cudaDeviceAttr.cudaDevAttrL2CacheSize #: Maximum resident threads per multiprocessor cudaDevAttrMaxThreadsPerMultiProcessor = cyruntime.cudaDeviceAttr.cudaDevAttrMaxThreadsPerMultiProcessor #: Number of asynchronous engines cudaDevAttrAsyncEngineCount = cyruntime.cudaDeviceAttr.cudaDevAttrAsyncEngineCount #: Device shares a unified address space with the host cudaDevAttrUnifiedAddressing = cyruntime.cudaDeviceAttr.cudaDevAttrUnifiedAddressing #: Maximum 1D layered texture width cudaDevAttrMaxTexture1DLayeredWidth = cyruntime.cudaDeviceAttr.cudaDevAttrMaxTexture1DLayeredWidth #: Maximum layers in a 1D layered texture cudaDevAttrMaxTexture1DLayeredLayers = cyruntime.cudaDeviceAttr.cudaDevAttrMaxTexture1DLayeredLayers #: Maximum 2D texture width if cudaArrayTextureGather is set cudaDevAttrMaxTexture2DGatherWidth = cyruntime.cudaDeviceAttr.cudaDevAttrMaxTexture2DGatherWidth #: Maximum 2D texture height if cudaArrayTextureGather is set cudaDevAttrMaxTexture2DGatherHeight = cyruntime.cudaDeviceAttr.cudaDevAttrMaxTexture2DGatherHeight #: Alternate maximum 3D texture width cudaDevAttrMaxTexture3DWidthAlt = cyruntime.cudaDeviceAttr.cudaDevAttrMaxTexture3DWidthAlt #: Alternate maximum 3D texture height cudaDevAttrMaxTexture3DHeightAlt = cyruntime.cudaDeviceAttr.cudaDevAttrMaxTexture3DHeightAlt #: Alternate maximum 3D texture depth cudaDevAttrMaxTexture3DDepthAlt = cyruntime.cudaDeviceAttr.cudaDevAttrMaxTexture3DDepthAlt #: PCI domain ID of the device cudaDevAttrPciDomainId = cyruntime.cudaDeviceAttr.cudaDevAttrPciDomainId #: Pitch alignment requirement for textures cudaDevAttrTexturePitchAlignment = cyruntime.cudaDeviceAttr.cudaDevAttrTexturePitchAlignment #: Maximum cubemap texture width/height cudaDevAttrMaxTextureCubemapWidth = cyruntime.cudaDeviceAttr.cudaDevAttrMaxTextureCubemapWidth #: Maximum cubemap layered texture width/height cudaDevAttrMaxTextureCubemapLayeredWidth = cyruntime.cudaDeviceAttr.cudaDevAttrMaxTextureCubemapLayeredWidth #: Maximum layers in a cubemap layered texture cudaDevAttrMaxTextureCubemapLayeredLayers = cyruntime.cudaDeviceAttr.cudaDevAttrMaxTextureCubemapLayeredLayers #: Maximum 1D surface width cudaDevAttrMaxSurface1DWidth = cyruntime.cudaDeviceAttr.cudaDevAttrMaxSurface1DWidth #: Maximum 2D surface width cudaDevAttrMaxSurface2DWidth = cyruntime.cudaDeviceAttr.cudaDevAttrMaxSurface2DWidth #: Maximum 2D surface height cudaDevAttrMaxSurface2DHeight = cyruntime.cudaDeviceAttr.cudaDevAttrMaxSurface2DHeight #: Maximum 3D surface width cudaDevAttrMaxSurface3DWidth = cyruntime.cudaDeviceAttr.cudaDevAttrMaxSurface3DWidth #: Maximum 3D surface height cudaDevAttrMaxSurface3DHeight = cyruntime.cudaDeviceAttr.cudaDevAttrMaxSurface3DHeight #: Maximum 3D surface depth cudaDevAttrMaxSurface3DDepth = cyruntime.cudaDeviceAttr.cudaDevAttrMaxSurface3DDepth #: Maximum 1D layered surface width cudaDevAttrMaxSurface1DLayeredWidth = cyruntime.cudaDeviceAttr.cudaDevAttrMaxSurface1DLayeredWidth #: Maximum layers in a 1D layered surface cudaDevAttrMaxSurface1DLayeredLayers = cyruntime.cudaDeviceAttr.cudaDevAttrMaxSurface1DLayeredLayers #: Maximum 2D layered surface width cudaDevAttrMaxSurface2DLayeredWidth = cyruntime.cudaDeviceAttr.cudaDevAttrMaxSurface2DLayeredWidth #: Maximum 2D layered surface height cudaDevAttrMaxSurface2DLayeredHeight = cyruntime.cudaDeviceAttr.cudaDevAttrMaxSurface2DLayeredHeight #: Maximum layers in a 2D layered surface cudaDevAttrMaxSurface2DLayeredLayers = cyruntime.cudaDeviceAttr.cudaDevAttrMaxSurface2DLayeredLayers #: Maximum cubemap surface width cudaDevAttrMaxSurfaceCubemapWidth = cyruntime.cudaDeviceAttr.cudaDevAttrMaxSurfaceCubemapWidth #: Maximum cubemap layered surface width cudaDevAttrMaxSurfaceCubemapLayeredWidth = cyruntime.cudaDeviceAttr.cudaDevAttrMaxSurfaceCubemapLayeredWidth #: Maximum layers in a cubemap layered surface cudaDevAttrMaxSurfaceCubemapLayeredLayers = cyruntime.cudaDeviceAttr.cudaDevAttrMaxSurfaceCubemapLayeredLayers #: Maximum 1D linear texture width cudaDevAttrMaxTexture1DLinearWidth = cyruntime.cudaDeviceAttr.cudaDevAttrMaxTexture1DLinearWidth #: Maximum 2D linear texture width cudaDevAttrMaxTexture2DLinearWidth = cyruntime.cudaDeviceAttr.cudaDevAttrMaxTexture2DLinearWidth #: Maximum 2D linear texture height cudaDevAttrMaxTexture2DLinearHeight = cyruntime.cudaDeviceAttr.cudaDevAttrMaxTexture2DLinearHeight #: Maximum 2D linear texture pitch in bytes cudaDevAttrMaxTexture2DLinearPitch = cyruntime.cudaDeviceAttr.cudaDevAttrMaxTexture2DLinearPitch #: Maximum mipmapped 2D texture width cudaDevAttrMaxTexture2DMipmappedWidth = cyruntime.cudaDeviceAttr.cudaDevAttrMaxTexture2DMipmappedWidth #: Maximum mipmapped 2D texture height cudaDevAttrMaxTexture2DMipmappedHeight = cyruntime.cudaDeviceAttr.cudaDevAttrMaxTexture2DMipmappedHeight #: Major compute capability version number cudaDevAttrComputeCapabilityMajor = cyruntime.cudaDeviceAttr.cudaDevAttrComputeCapabilityMajor #: Minor compute capability version number cudaDevAttrComputeCapabilityMinor = cyruntime.cudaDeviceAttr.cudaDevAttrComputeCapabilityMinor #: Maximum mipmapped 1D texture width cudaDevAttrMaxTexture1DMipmappedWidth = cyruntime.cudaDeviceAttr.cudaDevAttrMaxTexture1DMipmappedWidth #: Device supports stream priorities cudaDevAttrStreamPrioritiesSupported = cyruntime.cudaDeviceAttr.cudaDevAttrStreamPrioritiesSupported #: Device supports caching globals in L1 cudaDevAttrGlobalL1CacheSupported = cyruntime.cudaDeviceAttr.cudaDevAttrGlobalL1CacheSupported #: Device supports caching locals in L1 cudaDevAttrLocalL1CacheSupported = cyruntime.cudaDeviceAttr.cudaDevAttrLocalL1CacheSupported #: Maximum shared memory available per multiprocessor in bytes cudaDevAttrMaxSharedMemoryPerMultiprocessor = cyruntime.cudaDeviceAttr.cudaDevAttrMaxSharedMemoryPerMultiprocessor #: Maximum number of 32-bit registers available per multiprocessor cudaDevAttrMaxRegistersPerMultiprocessor = cyruntime.cudaDeviceAttr.cudaDevAttrMaxRegistersPerMultiprocessor #: Device can allocate managed memory on this system cudaDevAttrManagedMemory = cyruntime.cudaDeviceAttr.cudaDevAttrManagedMemory #: Device is on a multi-GPU board cudaDevAttrIsMultiGpuBoard = cyruntime.cudaDeviceAttr.cudaDevAttrIsMultiGpuBoard #: Unique identifier for a group of devices on the same multi-GPU board cudaDevAttrMultiGpuBoardGroupID = cyruntime.cudaDeviceAttr.cudaDevAttrMultiGpuBoardGroupID #: Link between the device and the host supports native atomic #: operations cudaDevAttrHostNativeAtomicSupported = cyruntime.cudaDeviceAttr.cudaDevAttrHostNativeAtomicSupported #: Ratio of single precision performance (in floating-point operations #: per second) to double precision performance cudaDevAttrSingleToDoublePrecisionPerfRatio = cyruntime.cudaDeviceAttr.cudaDevAttrSingleToDoublePrecisionPerfRatio #: Device supports coherently accessing pageable memory without calling #: cudaHostRegister on it cudaDevAttrPageableMemoryAccess = cyruntime.cudaDeviceAttr.cudaDevAttrPageableMemoryAccess #: Device can coherently access managed memory concurrently with the #: CPU cudaDevAttrConcurrentManagedAccess = cyruntime.cudaDeviceAttr.cudaDevAttrConcurrentManagedAccess #: Device supports Compute Preemption cudaDevAttrComputePreemptionSupported = cyruntime.cudaDeviceAttr.cudaDevAttrComputePreemptionSupported #: Device can access host registered memory at the same virtual address #: as the CPU cudaDevAttrCanUseHostPointerForRegisteredMem = cyruntime.cudaDeviceAttr.cudaDevAttrCanUseHostPointerForRegisteredMem cudaDevAttrReserved92 = cyruntime.cudaDeviceAttr.cudaDevAttrReserved92 cudaDevAttrReserved93 = cyruntime.cudaDeviceAttr.cudaDevAttrReserved93 cudaDevAttrReserved94 = cyruntime.cudaDeviceAttr.cudaDevAttrReserved94 #: Device supports launching cooperative kernels via #: :py:obj:`~.cudaLaunchCooperativeKernel` cudaDevAttrCooperativeLaunch = cyruntime.cudaDeviceAttr.cudaDevAttrCooperativeLaunch #: Deprecated, cudaLaunchCooperativeKernelMultiDevice is deprecated. cudaDevAttrCooperativeMultiDeviceLaunch = cyruntime.cudaDeviceAttr.cudaDevAttrCooperativeMultiDeviceLaunch #: The maximum optin shared memory per block. This value may vary by #: chip. See :py:obj:`~.cudaFuncSetAttribute` cudaDevAttrMaxSharedMemoryPerBlockOptin = cyruntime.cudaDeviceAttr.cudaDevAttrMaxSharedMemoryPerBlockOptin #: Device supports flushing of outstanding remote writes. cudaDevAttrCanFlushRemoteWrites = cyruntime.cudaDeviceAttr.cudaDevAttrCanFlushRemoteWrites #: Device supports host memory registration via #: :py:obj:`~.cudaHostRegister`. cudaDevAttrHostRegisterSupported = cyruntime.cudaDeviceAttr.cudaDevAttrHostRegisterSupported #: Device accesses pageable memory via the host's page tables. cudaDevAttrPageableMemoryAccessUsesHostPageTables = cyruntime.cudaDeviceAttr.cudaDevAttrPageableMemoryAccessUsesHostPageTables #: Host can directly access managed memory on the device without #: migration. cudaDevAttrDirectManagedMemAccessFromHost = cyruntime.cudaDeviceAttr.cudaDevAttrDirectManagedMemAccessFromHost #: Maximum number of blocks per multiprocessor cudaDevAttrMaxBlocksPerMultiprocessor = cyruntime.cudaDeviceAttr.cudaDevAttrMaxBlocksPerMultiprocessor #: Maximum L2 persisting lines capacity setting in bytes. cudaDevAttrMaxPersistingL2CacheSize = cyruntime.cudaDeviceAttr.cudaDevAttrMaxPersistingL2CacheSize #: Maximum value of :py:obj:`~.cudaAccessPolicyWindow.num_bytes`. cudaDevAttrMaxAccessPolicyWindowSize = cyruntime.cudaDeviceAttr.cudaDevAttrMaxAccessPolicyWindowSize #: Shared memory reserved by CUDA driver per block in bytes cudaDevAttrReservedSharedMemoryPerBlock = cyruntime.cudaDeviceAttr.cudaDevAttrReservedSharedMemoryPerBlock #: Device supports sparse CUDA arrays and sparse CUDA mipmapped arrays cudaDevAttrSparseCudaArraySupported = cyruntime.cudaDeviceAttr.cudaDevAttrSparseCudaArraySupported #: Device supports using the :py:obj:`~.cudaHostRegister` flag #: cudaHostRegisterReadOnly to register memory that must be mapped as #: read-only to the GPU cudaDevAttrHostRegisterReadOnlySupported = cyruntime.cudaDeviceAttr.cudaDevAttrHostRegisterReadOnlySupported #: External timeline semaphore interop is supported on the device cudaDevAttrTimelineSemaphoreInteropSupported = cyruntime.cudaDeviceAttr.cudaDevAttrTimelineSemaphoreInteropSupported #: Deprecated, External timeline semaphore interop is supported on the #: device cudaDevAttrMaxTimelineSemaphoreInteropSupported = cyruntime.cudaDeviceAttr.cudaDevAttrMaxTimelineSemaphoreInteropSupported #: Device supports using the :py:obj:`~.cudaMallocAsync` and #: :py:obj:`~.cudaMemPool` family of APIs cudaDevAttrMemoryPoolsSupported = cyruntime.cudaDeviceAttr.cudaDevAttrMemoryPoolsSupported #: Device supports GPUDirect RDMA APIs, like nvidia_p2p_get_pages (see #: https://docs.nvidia.com/cuda/gpudirect-rdma for more information) cudaDevAttrGPUDirectRDMASupported = cyruntime.cudaDeviceAttr.cudaDevAttrGPUDirectRDMASupported #: The returned attribute shall be interpreted as a bitmask, where the #: individual bits are listed in the #: :py:obj:`~.cudaFlushGPUDirectRDMAWritesOptions` enum cudaDevAttrGPUDirectRDMAFlushWritesOptions = cyruntime.cudaDeviceAttr.cudaDevAttrGPUDirectRDMAFlushWritesOptions #: GPUDirect RDMA writes to the device do not need to be flushed for #: consumers within the scope indicated by the returned attribute. See #: :py:obj:`~.cudaGPUDirectRDMAWritesOrdering` for the numerical values #: returned here. cudaDevAttrGPUDirectRDMAWritesOrdering = cyruntime.cudaDeviceAttr.cudaDevAttrGPUDirectRDMAWritesOrdering #: Handle types supported with mempool based IPC cudaDevAttrMemoryPoolSupportedHandleTypes = cyruntime.cudaDeviceAttr.cudaDevAttrMemoryPoolSupportedHandleTypes #: Indicates device supports cluster launch cudaDevAttrClusterLaunch = cyruntime.cudaDeviceAttr.cudaDevAttrClusterLaunch #: Device supports deferred mapping CUDA arrays and CUDA mipmapped #: arrays cudaDevAttrDeferredMappingCudaArraySupported = cyruntime.cudaDeviceAttr.cudaDevAttrDeferredMappingCudaArraySupported cudaDevAttrReserved122 = cyruntime.cudaDeviceAttr.cudaDevAttrReserved122 cudaDevAttrReserved123 = cyruntime.cudaDeviceAttr.cudaDevAttrReserved123 cudaDevAttrReserved124 = cyruntime.cudaDeviceAttr.cudaDevAttrReserved124 #: Device supports IPC Events. cudaDevAttrIpcEventSupport = cyruntime.cudaDeviceAttr.cudaDevAttrIpcEventSupport #: Number of memory synchronization domains the device supports. cudaDevAttrMemSyncDomainCount = cyruntime.cudaDeviceAttr.cudaDevAttrMemSyncDomainCount cudaDevAttrReserved127 = cyruntime.cudaDeviceAttr.cudaDevAttrReserved127 cudaDevAttrReserved128 = cyruntime.cudaDeviceAttr.cudaDevAttrReserved128 cudaDevAttrReserved129 = cyruntime.cudaDeviceAttr.cudaDevAttrReserved129 #: NUMA configuration of a device: value is of type #: :py:obj:`~.cudaDeviceNumaConfig` enum cudaDevAttrNumaConfig = cyruntime.cudaDeviceAttr.cudaDevAttrNumaConfig #: NUMA node ID of the GPU memory cudaDevAttrNumaId = cyruntime.cudaDeviceAttr.cudaDevAttrNumaId cudaDevAttrReserved132 = cyruntime.cudaDeviceAttr.cudaDevAttrReserved132 #: Contexts created on this device will be shared via MPS cudaDevAttrMpsEnabled = cyruntime.cudaDeviceAttr.cudaDevAttrMpsEnabled #: NUMA ID of the host node closest to the device or -1 when system #: does not support NUMA cudaDevAttrHostNumaId = cyruntime.cudaDeviceAttr.cudaDevAttrHostNumaId #: Device supports CIG with D3D12. cudaDevAttrD3D12CigSupported = cyruntime.cudaDeviceAttr.cudaDevAttrD3D12CigSupported #: Device supports CIG with Vulkan. cudaDevAttrVulkanCigSupported = cyruntime.cudaDeviceAttr.cudaDevAttrVulkanCigSupported #: The combined 16-bit PCI device ID and 16-bit PCI vendor ID. cudaDevAttrGpuPciDeviceId = cyruntime.cudaDeviceAttr.cudaDevAttrGpuPciDeviceId #: The combined 16-bit PCI subsystem ID and 16-bit PCI subsystem vendor #: ID. cudaDevAttrGpuPciSubsystemId = cyruntime.cudaDeviceAttr.cudaDevAttrGpuPciSubsystemId cudaDevAttrReserved141 = cyruntime.cudaDeviceAttr.cudaDevAttrReserved141 #: Device supports HOST_NUMA location with the #: :py:obj:`~.cudaMallocAsync` and :py:obj:`~.cudaMemPool` family of #: APIs cudaDevAttrHostNumaMemoryPoolsSupported = cyruntime.cudaDeviceAttr.cudaDevAttrHostNumaMemoryPoolsSupported #: Device supports HostNuma location IPC between nodes in a multi-node #: system. cudaDevAttrHostNumaMultinodeIpcSupported = cyruntime.cudaDeviceAttr.cudaDevAttrHostNumaMultinodeIpcSupported cudaDevAttrMax = cyruntime.cudaDeviceAttr.cudaDevAttrMax _dict_cudaDeviceAttr = dict(((int(v), v) for k, v in cudaDeviceAttr.__members__.items())) class cudaMemPoolAttr(IntEnum): """ CUDA memory pool attributes """ #: (value type = int) Allow cuMemAllocAsync to use memory #: asynchronously freed in another streams as long as a stream ordering #: dependency of the allocating stream on the free action exists. Cuda #: events and null stream interactions can create the required stream #: ordered dependencies. (default enabled) cudaMemPoolReuseFollowEventDependencies = cyruntime.cudaMemPoolAttr.cudaMemPoolReuseFollowEventDependencies #: (value type = int) Allow reuse of already completed frees when there #: is no dependency between the free and allocation. (default enabled) cudaMemPoolReuseAllowOpportunistic = cyruntime.cudaMemPoolAttr.cudaMemPoolReuseAllowOpportunistic #: (value type = int) Allow cuMemAllocAsync to insert new stream #: dependencies in order to establish the stream ordering required to #: reuse a piece of memory released by cuFreeAsync (default enabled). cudaMemPoolReuseAllowInternalDependencies = cyruntime.cudaMemPoolAttr.cudaMemPoolReuseAllowInternalDependencies #: (value type = cuuint64_t) Amount of reserved memory in bytes to hold #: onto before trying to release memory back to the OS. When more than #: the release threshold bytes of memory are held by the memory pool, #: the allocator will try to release memory back to the OS on the next #: call to stream, event or context synchronize. (default 0) cudaMemPoolAttrReleaseThreshold = cyruntime.cudaMemPoolAttr.cudaMemPoolAttrReleaseThreshold #: (value type = cuuint64_t) Amount of backing memory currently #: allocated for the mempool. cudaMemPoolAttrReservedMemCurrent = cyruntime.cudaMemPoolAttr.cudaMemPoolAttrReservedMemCurrent #: (value type = cuuint64_t) High watermark of backing memory allocated #: for the mempool since the last time it was reset. High watermark can #: only be reset to zero. cudaMemPoolAttrReservedMemHigh = cyruntime.cudaMemPoolAttr.cudaMemPoolAttrReservedMemHigh #: (value type = cuuint64_t) Amount of memory from the pool that is #: currently in use by the application. cudaMemPoolAttrUsedMemCurrent = cyruntime.cudaMemPoolAttr.cudaMemPoolAttrUsedMemCurrent #: (value type = cuuint64_t) High watermark of the amount of memory #: from the pool that was in use by the application since the last time #: it was reset. High watermark can only be reset to zero. cudaMemPoolAttrUsedMemHigh = cyruntime.cudaMemPoolAttr.cudaMemPoolAttrUsedMemHigh _dict_cudaMemPoolAttr = dict(((int(v), v) for k, v in cudaMemPoolAttr.__members__.items())) class cudaMemLocationType(IntEnum): """ Specifies the type of location """ cudaMemLocationTypeInvalid = cyruntime.cudaMemLocationType.cudaMemLocationTypeInvalid #: Location is a device location, thus id is a device ordinal cudaMemLocationTypeDevice = cyruntime.cudaMemLocationType.cudaMemLocationTypeDevice #: Location is host, id is ignored cudaMemLocationTypeHost = cyruntime.cudaMemLocationType.cudaMemLocationTypeHost #: Location is a host NUMA node, thus id is a host NUMA node id cudaMemLocationTypeHostNuma = cyruntime.cudaMemLocationType.cudaMemLocationTypeHostNuma #: Location is the host NUMA node closest to the current thread's CPU, #: id is ignored cudaMemLocationTypeHostNumaCurrent = cyruntime.cudaMemLocationType.cudaMemLocationTypeHostNumaCurrent _dict_cudaMemLocationType = dict(((int(v), v) for k, v in cudaMemLocationType.__members__.items())) class cudaMemAccessFlags(IntEnum): """ Specifies the memory protection flags for mapping. """ #: Default, make the address range not accessible cudaMemAccessFlagsProtNone = cyruntime.cudaMemAccessFlags.cudaMemAccessFlagsProtNone #: Make the address range read accessible cudaMemAccessFlagsProtRead = cyruntime.cudaMemAccessFlags.cudaMemAccessFlagsProtRead #: Make the address range read-write accessible cudaMemAccessFlagsProtReadWrite = cyruntime.cudaMemAccessFlags.cudaMemAccessFlagsProtReadWrite _dict_cudaMemAccessFlags = dict(((int(v), v) for k, v in cudaMemAccessFlags.__members__.items())) class cudaMemAllocationType(IntEnum): """ Defines the allocation types available """ cudaMemAllocationTypeInvalid = cyruntime.cudaMemAllocationType.cudaMemAllocationTypeInvalid #: This allocation type is 'pinned', i.e. cannot migrate from its #: current location while the application is actively using it cudaMemAllocationTypePinned = cyruntime.cudaMemAllocationType.cudaMemAllocationTypePinned cudaMemAllocationTypeMax = cyruntime.cudaMemAllocationType.cudaMemAllocationTypeMax _dict_cudaMemAllocationType = dict(((int(v), v) for k, v in cudaMemAllocationType.__members__.items())) class cudaMemAllocationHandleType(IntEnum): """ Flags for specifying particular handle types """ #: Does not allow any export mechanism. > cudaMemHandleTypeNone = cyruntime.cudaMemAllocationHandleType.cudaMemHandleTypeNone #: Allows a file descriptor to be used for exporting. Permitted only on #: POSIX systems. (int) cudaMemHandleTypePosixFileDescriptor = cyruntime.cudaMemAllocationHandleType.cudaMemHandleTypePosixFileDescriptor #: Allows a Win32 NT handle to be used for exporting. (HANDLE) cudaMemHandleTypeWin32 = cyruntime.cudaMemAllocationHandleType.cudaMemHandleTypeWin32 #: Allows a Win32 KMT handle to be used for exporting. (D3DKMT_HANDLE) cudaMemHandleTypeWin32Kmt = cyruntime.cudaMemAllocationHandleType.cudaMemHandleTypeWin32Kmt #: Allows a fabric handle to be used for exporting. #: (cudaMemFabricHandle_t) cudaMemHandleTypeFabric = cyruntime.cudaMemAllocationHandleType.cudaMemHandleTypeFabric _dict_cudaMemAllocationHandleType = dict(((int(v), v) for k, v in cudaMemAllocationHandleType.__members__.items())) class cudaGraphMemAttributeType(IntEnum): """ Graph memory attributes """ #: (value type = cuuint64_t) Amount of memory, in bytes, currently #: associated with graphs. cudaGraphMemAttrUsedMemCurrent = cyruntime.cudaGraphMemAttributeType.cudaGraphMemAttrUsedMemCurrent #: (value type = cuuint64_t) High watermark of memory, in bytes, #: associated with graphs since the last time it was reset. High #: watermark can only be reset to zero. cudaGraphMemAttrUsedMemHigh = cyruntime.cudaGraphMemAttributeType.cudaGraphMemAttrUsedMemHigh #: (value type = cuuint64_t) Amount of memory, in bytes, currently #: allocated for use by the CUDA graphs asynchronous allocator. cudaGraphMemAttrReservedMemCurrent = cyruntime.cudaGraphMemAttributeType.cudaGraphMemAttrReservedMemCurrent #: (value type = cuuint64_t) High watermark of memory, in bytes, #: currently allocated for use by the CUDA graphs asynchronous #: allocator. cudaGraphMemAttrReservedMemHigh = cyruntime.cudaGraphMemAttributeType.cudaGraphMemAttrReservedMemHigh _dict_cudaGraphMemAttributeType = dict(((int(v), v) for k, v in cudaGraphMemAttributeType.__members__.items())) class cudaMemcpyFlags(IntEnum): """ Flags to specify for copies within a batch. For more details see :py:obj:`~.cudaMemcpyBatchAsync`. """ cudaMemcpyFlagDefault = cyruntime.cudaMemcpyFlags.cudaMemcpyFlagDefault #: Hint to the driver to try and overlap the copy with compute work on #: the SMs. cudaMemcpyFlagPreferOverlapWithCompute = cyruntime.cudaMemcpyFlags.cudaMemcpyFlagPreferOverlapWithCompute _dict_cudaMemcpyFlags = dict(((int(v), v) for k, v in cudaMemcpyFlags.__members__.items())) class cudaMemcpySrcAccessOrder(IntEnum): """ """ #: Default invalid. cudaMemcpySrcAccessOrderInvalid = cyruntime.cudaMemcpySrcAccessOrder.cudaMemcpySrcAccessOrderInvalid #: Indicates that access to the source pointer must be in stream order. cudaMemcpySrcAccessOrderStream = cyruntime.cudaMemcpySrcAccessOrder.cudaMemcpySrcAccessOrderStream #: Indicates that access to the source pointer can be out of stream #: order and all accesses must be complete before the API call returns. #: This flag is suited for ephemeral sources (ex., stack variables) #: when it's known that no prior operations in the stream can be #: accessing the memory and also that the lifetime of the memory is #: limited to the scope that the source variable was declared in. #: Specifying this flag allows the driver to optimize the copy and #: removes the need for the user to synchronize the stream after the #: API call. cudaMemcpySrcAccessOrderDuringApiCall = cyruntime.cudaMemcpySrcAccessOrder.cudaMemcpySrcAccessOrderDuringApiCall #: Indicates that access to the source pointer can be out of stream #: order and the accesses can happen even after the API call returns. #: This flag is suited for host pointers allocated outside CUDA (ex., #: via malloc) when it's known that no prior operations in the stream #: can be accessing the memory. Specifying this flag allows the driver #: to optimize the copy on certain platforms. cudaMemcpySrcAccessOrderAny = cyruntime.cudaMemcpySrcAccessOrder.cudaMemcpySrcAccessOrderAny cudaMemcpySrcAccessOrderMax = cyruntime.cudaMemcpySrcAccessOrder.cudaMemcpySrcAccessOrderMax _dict_cudaMemcpySrcAccessOrder = dict(((int(v), v) for k, v in cudaMemcpySrcAccessOrder.__members__.items())) class cudaMemcpy3DOperandType(IntEnum): """ These flags allow applications to convey the operand type for individual copies specified in :py:obj:`~.cudaMemcpy3DBatchAsync`. """ #: Memcpy operand is a valid pointer. cudaMemcpyOperandTypePointer = cyruntime.cudaMemcpy3DOperandType.cudaMemcpyOperandTypePointer #: Memcpy operand is a CUarray. cudaMemcpyOperandTypeArray = cyruntime.cudaMemcpy3DOperandType.cudaMemcpyOperandTypeArray cudaMemcpyOperandTypeMax = cyruntime.cudaMemcpy3DOperandType.cudaMemcpyOperandTypeMax _dict_cudaMemcpy3DOperandType = dict(((int(v), v) for k, v in cudaMemcpy3DOperandType.__members__.items())) class cudaDeviceP2PAttr(IntEnum): """ CUDA device P2P attributes """ #: A relative value indicating the performance of the link between two #: devices cudaDevP2PAttrPerformanceRank = cyruntime.cudaDeviceP2PAttr.cudaDevP2PAttrPerformanceRank #: Peer access is enabled cudaDevP2PAttrAccessSupported = cyruntime.cudaDeviceP2PAttr.cudaDevP2PAttrAccessSupported #: Native atomic operation over the link supported cudaDevP2PAttrNativeAtomicSupported = cyruntime.cudaDeviceP2PAttr.cudaDevP2PAttrNativeAtomicSupported #: Accessing CUDA arrays over the link supported cudaDevP2PAttrCudaArrayAccessSupported = cyruntime.cudaDeviceP2PAttr.cudaDevP2PAttrCudaArrayAccessSupported _dict_cudaDeviceP2PAttr = dict(((int(v), v) for k, v in cudaDeviceP2PAttr.__members__.items())) class cudaExternalMemoryHandleType(IntEnum): """ External memory handle types """ #: Handle is an opaque file descriptor cudaExternalMemoryHandleTypeOpaqueFd = cyruntime.cudaExternalMemoryHandleType.cudaExternalMemoryHandleTypeOpaqueFd #: Handle is an opaque shared NT handle cudaExternalMemoryHandleTypeOpaqueWin32 = cyruntime.cudaExternalMemoryHandleType.cudaExternalMemoryHandleTypeOpaqueWin32 #: Handle is an opaque, globally shared handle cudaExternalMemoryHandleTypeOpaqueWin32Kmt = cyruntime.cudaExternalMemoryHandleType.cudaExternalMemoryHandleTypeOpaqueWin32Kmt #: Handle is a D3D12 heap object cudaExternalMemoryHandleTypeD3D12Heap = cyruntime.cudaExternalMemoryHandleType.cudaExternalMemoryHandleTypeD3D12Heap #: Handle is a D3D12 committed resource cudaExternalMemoryHandleTypeD3D12Resource = cyruntime.cudaExternalMemoryHandleType.cudaExternalMemoryHandleTypeD3D12Resource #: Handle is a shared NT handle to a D3D11 resource cudaExternalMemoryHandleTypeD3D11Resource = cyruntime.cudaExternalMemoryHandleType.cudaExternalMemoryHandleTypeD3D11Resource #: Handle is a globally shared handle to a D3D11 resource cudaExternalMemoryHandleTypeD3D11ResourceKmt = cyruntime.cudaExternalMemoryHandleType.cudaExternalMemoryHandleTypeD3D11ResourceKmt #: Handle is an NvSciBuf object cudaExternalMemoryHandleTypeNvSciBuf = cyruntime.cudaExternalMemoryHandleType.cudaExternalMemoryHandleTypeNvSciBuf _dict_cudaExternalMemoryHandleType = dict(((int(v), v) for k, v in cudaExternalMemoryHandleType.__members__.items())) class cudaExternalSemaphoreHandleType(IntEnum): """ External semaphore handle types """ #: Handle is an opaque file descriptor cudaExternalSemaphoreHandleTypeOpaqueFd = cyruntime.cudaExternalSemaphoreHandleType.cudaExternalSemaphoreHandleTypeOpaqueFd #: Handle is an opaque shared NT handle cudaExternalSemaphoreHandleTypeOpaqueWin32 = cyruntime.cudaExternalSemaphoreHandleType.cudaExternalSemaphoreHandleTypeOpaqueWin32 #: Handle is an opaque, globally shared handle cudaExternalSemaphoreHandleTypeOpaqueWin32Kmt = cyruntime.cudaExternalSemaphoreHandleType.cudaExternalSemaphoreHandleTypeOpaqueWin32Kmt #: Handle is a shared NT handle referencing a D3D12 fence object cudaExternalSemaphoreHandleTypeD3D12Fence = cyruntime.cudaExternalSemaphoreHandleType.cudaExternalSemaphoreHandleTypeD3D12Fence #: Handle is a shared NT handle referencing a D3D11 fence object cudaExternalSemaphoreHandleTypeD3D11Fence = cyruntime.cudaExternalSemaphoreHandleType.cudaExternalSemaphoreHandleTypeD3D11Fence #: Opaque handle to NvSciSync Object cudaExternalSemaphoreHandleTypeNvSciSync = cyruntime.cudaExternalSemaphoreHandleType.cudaExternalSemaphoreHandleTypeNvSciSync #: Handle is a shared NT handle referencing a D3D11 keyed mutex object cudaExternalSemaphoreHandleTypeKeyedMutex = cyruntime.cudaExternalSemaphoreHandleType.cudaExternalSemaphoreHandleTypeKeyedMutex #: Handle is a shared KMT handle referencing a D3D11 keyed mutex object cudaExternalSemaphoreHandleTypeKeyedMutexKmt = cyruntime.cudaExternalSemaphoreHandleType.cudaExternalSemaphoreHandleTypeKeyedMutexKmt #: Handle is an opaque handle file descriptor referencing a timeline #: semaphore cudaExternalSemaphoreHandleTypeTimelineSemaphoreFd = cyruntime.cudaExternalSemaphoreHandleType.cudaExternalSemaphoreHandleTypeTimelineSemaphoreFd #: Handle is an opaque handle file descriptor referencing a timeline #: semaphore cudaExternalSemaphoreHandleTypeTimelineSemaphoreWin32 = cyruntime.cudaExternalSemaphoreHandleType.cudaExternalSemaphoreHandleTypeTimelineSemaphoreWin32 _dict_cudaExternalSemaphoreHandleType = dict(((int(v), v) for k, v in cudaExternalSemaphoreHandleType.__members__.items())) class cudaJitOption(IntEnum): """ Online compiler and linker options """ #: Max number of registers that a thread may use. #: Option type: unsigned int #: Applies to: compiler only cudaJitMaxRegisters = cyruntime.cudaJitOption.cudaJitMaxRegisters #: IN: Specifies minimum number of threads per block to target #: compilation for #: OUT: Returns the number of threads the compiler actually targeted. #: This restricts the resource utilization of the compiler (e.g. max #: registers) such that a block with the given number of threads should #: be able to launch based on register limitations. Note, this option #: does not currently take into account any other resource limitations, #: such as shared memory utilization. #: Option type: unsigned int #: Applies to: compiler only cudaJitThreadsPerBlock = cyruntime.cudaJitOption.cudaJitThreadsPerBlock #: Overwrites the option value with the total wall clock time, in #: milliseconds, spent in the compiler and linker #: Option type: float #: Applies to: compiler and linker cudaJitWallTime = cyruntime.cudaJitOption.cudaJitWallTime #: Pointer to a buffer in which to print any log messages that are #: informational in nature (the buffer size is specified via option #: :py:obj:`~.cudaJitInfoLogBufferSizeBytes`) #: Option type: char * #: Applies to: compiler and linker cudaJitInfoLogBuffer = cyruntime.cudaJitOption.cudaJitInfoLogBuffer #: IN: Log buffer size in bytes. Log messages will be capped at this #: size (including null terminator) #: OUT: Amount of log buffer filled with messages #: Option type: unsigned int #: Applies to: compiler and linker cudaJitInfoLogBufferSizeBytes = cyruntime.cudaJitOption.cudaJitInfoLogBufferSizeBytes #: Pointer to a buffer in which to print any log messages that reflect #: errors (the buffer size is specified via option #: :py:obj:`~.cudaJitErrorLogBufferSizeBytes`) #: Option type: char * #: Applies to: compiler and linker cudaJitErrorLogBuffer = cyruntime.cudaJitOption.cudaJitErrorLogBuffer #: IN: Log buffer size in bytes. Log messages will be capped at this #: size (including null terminator) #: OUT: Amount of log buffer filled with messages #: Option type: unsigned int #: Applies to: compiler and linker cudaJitErrorLogBufferSizeBytes = cyruntime.cudaJitOption.cudaJitErrorLogBufferSizeBytes #: Level of optimizations to apply to generated code (0 - 4), with 4 #: being the default and highest level of optimizations. #: Option type: unsigned int #: Applies to: compiler only cudaJitOptimizationLevel = cyruntime.cudaJitOption.cudaJitOptimizationLevel #: Specifies choice of fallback strategy if matching cubin is not #: found. Choice is based on supplied :py:obj:`~.cudaJit_Fallback`. #: Option type: unsigned int for enumerated type #: :py:obj:`~.cudaJit_Fallback` #: Applies to: compiler only cudaJitFallbackStrategy = cyruntime.cudaJitOption.cudaJitFallbackStrategy #: Specifies whether to create debug information in output (-g) (0: #: false, default) #: Option type: int #: Applies to: compiler and linker cudaJitGenerateDebugInfo = cyruntime.cudaJitOption.cudaJitGenerateDebugInfo #: Generate verbose log messages (0: false, default) #: Option type: int #: Applies to: compiler and linker cudaJitLogVerbose = cyruntime.cudaJitOption.cudaJitLogVerbose #: Generate line number information (-lineinfo) (0: false, default) #: Option type: int #: Applies to: compiler only cudaJitGenerateLineInfo = cyruntime.cudaJitOption.cudaJitGenerateLineInfo #: Specifies whether to enable caching explicitly (-dlcm) #: Choice is based on supplied :py:obj:`~.cudaJit_CacheMode`. #: Option type: unsigned int for enumerated type #: :py:obj:`~.cudaJit_CacheMode` #: Applies to: compiler only cudaJitCacheMode = cyruntime.cudaJitOption.cudaJitCacheMode #: Generate position independent code (0: false) #: Option type: int #: Applies to: compiler only cudaJitPositionIndependentCode = cyruntime.cudaJitOption.cudaJitPositionIndependentCode #: This option hints to the JIT compiler the minimum number of CTAs #: from the kernel’s grid to be mapped to a SM. This option is ignored #: when used together with :py:obj:`~.cudaJitMaxRegisters` or #: :py:obj:`~.cudaJitThreadsPerBlock`. Optimizations based on this #: option need :py:obj:`~.cudaJitMaxThreadsPerBlock` to be specified as #: well. For kernels already using PTX directive .minnctapersm, this #: option will be ignored by default. Use #: :py:obj:`~.cudaJitOverrideDirectiveValues` to let this option take #: precedence over the PTX directive. Option type: unsigned int #: Applies to: compiler only cudaJitMinCtaPerSm = cyruntime.cudaJitOption.cudaJitMinCtaPerSm #: Maximum number threads in a thread block, computed as the product of #: the maximum extent specifed for each dimension of the block. This #: limit is guaranteed not to be exeeded in any invocation of the #: kernel. Exceeding the the maximum number of threads results in #: runtime error or kernel launch failure. For kernels already using #: PTX directive .maxntid, this option will be ignored by default. Use #: :py:obj:`~.cudaJitOverrideDirectiveValues` to let this option take #: precedence over the PTX directive. Option type: int #: Applies to: compiler only cudaJitMaxThreadsPerBlock = cyruntime.cudaJitOption.cudaJitMaxThreadsPerBlock #: This option lets the values specified using #: :py:obj:`~.cudaJitMaxRegisters`, :py:obj:`~.cudaJitThreadsPerBlock`, #: :py:obj:`~.cudaJitMaxThreadsPerBlock` and #: :py:obj:`~.cudaJitMinCtaPerSm` take precedence over any PTX #: directives. (0: Disable, default; 1: Enable) Option type: int #: Applies to: compiler only cudaJitOverrideDirectiveValues = cyruntime.cudaJitOption.cudaJitOverrideDirectiveValues _dict_cudaJitOption = dict(((int(v), v) for k, v in cudaJitOption.__members__.items())) class cudaLibraryOption(IntEnum): """ Library options to be specified with :py:obj:`~.cudaLibraryLoadData()` or :py:obj:`~.cudaLibraryLoadFromFile()` """ cudaLibraryHostUniversalFunctionAndDataTable = cyruntime.cudaLibraryOption.cudaLibraryHostUniversalFunctionAndDataTable #: Specifes that the argument `code` passed to #: :py:obj:`~.cudaLibraryLoadData()` will be preserved. Specifying this #: option will let the driver know that `code` can be accessed at any #: point until :py:obj:`~.cudaLibraryUnload()`. The default behavior is #: for the driver to allocate and maintain its own copy of `code`. Note #: that this is only a memory usage optimization hint and the driver #: can choose to ignore it if required. Specifying this option with #: :py:obj:`~.cudaLibraryLoadFromFile()` is invalid and will return #: :py:obj:`~.cudaErrorInvalidValue`. cudaLibraryBinaryIsPreserved = cyruntime.cudaLibraryOption.cudaLibraryBinaryIsPreserved _dict_cudaLibraryOption = dict(((int(v), v) for k, v in cudaLibraryOption.__members__.items())) class cudaJit_CacheMode(IntEnum): """ Caching modes for dlcm """ #: Compile with no -dlcm flag specified cudaJitCacheOptionNone = cyruntime.cudaJit_CacheMode.cudaJitCacheOptionNone #: Compile with L1 cache disabled cudaJitCacheOptionCG = cyruntime.cudaJit_CacheMode.cudaJitCacheOptionCG #: Compile with L1 cache enabled cudaJitCacheOptionCA = cyruntime.cudaJit_CacheMode.cudaJitCacheOptionCA _dict_cudaJit_CacheMode = dict(((int(v), v) for k, v in cudaJit_CacheMode.__members__.items())) class cudaJit_Fallback(IntEnum): """ Cubin matching fallback strategies """ #: Prefer to compile ptx if exact binary match not found cudaPreferPtx = cyruntime.cudaJit_Fallback.cudaPreferPtx #: Prefer to fall back to compatible binary code if exact match not #: found cudaPreferBinary = cyruntime.cudaJit_Fallback.cudaPreferBinary _dict_cudaJit_Fallback = dict(((int(v), v) for k, v in cudaJit_Fallback.__members__.items())) class cudaCGScope(IntEnum): """ CUDA cooperative group scope """ #: Invalid cooperative group scope cudaCGScopeInvalid = cyruntime.cudaCGScope.cudaCGScopeInvalid #: Scope represented by a grid_group cudaCGScopeGrid = cyruntime.cudaCGScope.cudaCGScopeGrid #: Scope represented by a multi_grid_group cudaCGScopeMultiGrid = cyruntime.cudaCGScope.cudaCGScopeMultiGrid _dict_cudaCGScope = dict(((int(v), v) for k, v in cudaCGScope.__members__.items())) class cudaGraphConditionalHandleFlags(IntEnum): """ """ #: Apply default handle value when graph is launched. cudaGraphCondAssignDefault = cyruntime.cudaGraphConditionalHandleFlags.cudaGraphCondAssignDefault _dict_cudaGraphConditionalHandleFlags = dict(((int(v), v) for k, v in cudaGraphConditionalHandleFlags.__members__.items())) class cudaGraphConditionalNodeType(IntEnum): """ CUDA conditional node types """ #: Conditional 'if/else' Node. Body[0] executed if condition is non- #: zero. If `size` == 2, an optional ELSE graph is created and this is #: executed if the condition is zero. cudaGraphCondTypeIf = cyruntime.cudaGraphConditionalNodeType.cudaGraphCondTypeIf #: Conditional 'while' Node. Body executed repeatedly while condition #: value is non-zero. cudaGraphCondTypeWhile = cyruntime.cudaGraphConditionalNodeType.cudaGraphCondTypeWhile #: Conditional 'switch' Node. Body[n] is executed once, where 'n' is #: the value of the condition. If the condition does not match a body #: index, no body is launched. cudaGraphCondTypeSwitch = cyruntime.cudaGraphConditionalNodeType.cudaGraphCondTypeSwitch _dict_cudaGraphConditionalNodeType = dict(((int(v), v) for k, v in cudaGraphConditionalNodeType.__members__.items())) class cudaGraphNodeType(IntEnum): """ CUDA Graph node types """ #: GPU kernel node cudaGraphNodeTypeKernel = cyruntime.cudaGraphNodeType.cudaGraphNodeTypeKernel #: Memcpy node cudaGraphNodeTypeMemcpy = cyruntime.cudaGraphNodeType.cudaGraphNodeTypeMemcpy #: Memset node cudaGraphNodeTypeMemset = cyruntime.cudaGraphNodeType.cudaGraphNodeTypeMemset #: Host (executable) node cudaGraphNodeTypeHost = cyruntime.cudaGraphNodeType.cudaGraphNodeTypeHost #: Node which executes an embedded graph cudaGraphNodeTypeGraph = cyruntime.cudaGraphNodeType.cudaGraphNodeTypeGraph #: Empty (no-op) node cudaGraphNodeTypeEmpty = cyruntime.cudaGraphNodeType.cudaGraphNodeTypeEmpty #: External event wait node cudaGraphNodeTypeWaitEvent = cyruntime.cudaGraphNodeType.cudaGraphNodeTypeWaitEvent #: External event record node cudaGraphNodeTypeEventRecord = cyruntime.cudaGraphNodeType.cudaGraphNodeTypeEventRecord #: External semaphore signal node cudaGraphNodeTypeExtSemaphoreSignal = cyruntime.cudaGraphNodeType.cudaGraphNodeTypeExtSemaphoreSignal #: External semaphore wait node cudaGraphNodeTypeExtSemaphoreWait = cyruntime.cudaGraphNodeType.cudaGraphNodeTypeExtSemaphoreWait #: Memory allocation node cudaGraphNodeTypeMemAlloc = cyruntime.cudaGraphNodeType.cudaGraphNodeTypeMemAlloc #: Memory free node cudaGraphNodeTypeMemFree = cyruntime.cudaGraphNodeType.cudaGraphNodeTypeMemFree #: Conditional node May be used to #: implement a conditional execution path or loop #: inside of a graph. The graph(s) #: contained within the body of the conditional node #: can be selectively executed or #: iterated upon based on the value of a conditional #: variable. #: #: Handles must be created in #: advance of creating the node #: using #: :py:obj:`~.cudaGraphConditionalHandleCreate`. #: #: The following restrictions apply #: to graphs which contain conditional nodes: #: The graph cannot be used in a #: child node. #: Only one instantiation of the #: graph may exist at any point in time. #: The graph cannot be cloned. #: #: To set the control value, supply #: a default value when creating the handle and/or #: call #: :py:obj:`~.cudaGraphSetConditional` from device code. cudaGraphNodeTypeConditional = cyruntime.cudaGraphNodeType.cudaGraphNodeTypeConditional cudaGraphNodeTypeCount = cyruntime.cudaGraphNodeType.cudaGraphNodeTypeCount _dict_cudaGraphNodeType = dict(((int(v), v) for k, v in cudaGraphNodeType.__members__.items())) class cudaGraphChildGraphNodeOwnership(IntEnum): """ Child graph node ownership """ #: Default behavior for a child graph node. Child graph is cloned into #: the parent and memory allocation/free nodes can't be present in the #: child graph. cudaGraphChildGraphOwnershipClone = cyruntime.cudaGraphChildGraphNodeOwnership.cudaGraphChildGraphOwnershipClone #: The child graph is moved to the parent. The handle to the child #: graph is owned by the parent and will be destroyed when the parent #: is destroyed. #: #: The following restrictions apply to child graphs after they have #: been moved: Cannot be independently instantiated or destroyed; #: Cannot be added as a child graph of a separate parent graph; Cannot #: be used as an argument to cudaGraphExecUpdate; Cannot have #: additional memory allocation or free nodes added. cudaGraphChildGraphOwnershipMove = cyruntime.cudaGraphChildGraphNodeOwnership.cudaGraphChildGraphOwnershipMove _dict_cudaGraphChildGraphNodeOwnership = dict(((int(v), v) for k, v in cudaGraphChildGraphNodeOwnership.__members__.items())) class cudaGraphExecUpdateResult(IntEnum): """ CUDA Graph Update error types """ #: The update succeeded cudaGraphExecUpdateSuccess = cyruntime.cudaGraphExecUpdateResult.cudaGraphExecUpdateSuccess #: The update failed for an unexpected reason which is described in the #: return value of the function cudaGraphExecUpdateError = cyruntime.cudaGraphExecUpdateResult.cudaGraphExecUpdateError #: The update failed because the topology changed cudaGraphExecUpdateErrorTopologyChanged = cyruntime.cudaGraphExecUpdateResult.cudaGraphExecUpdateErrorTopologyChanged #: The update failed because a node type changed cudaGraphExecUpdateErrorNodeTypeChanged = cyruntime.cudaGraphExecUpdateResult.cudaGraphExecUpdateErrorNodeTypeChanged #: The update failed because the function of a kernel node changed #: (CUDA driver < 11.2) cudaGraphExecUpdateErrorFunctionChanged = cyruntime.cudaGraphExecUpdateResult.cudaGraphExecUpdateErrorFunctionChanged #: The update failed because the parameters changed in a way that is #: not supported cudaGraphExecUpdateErrorParametersChanged = cyruntime.cudaGraphExecUpdateResult.cudaGraphExecUpdateErrorParametersChanged #: The update failed because something about the node is not supported cudaGraphExecUpdateErrorNotSupported = cyruntime.cudaGraphExecUpdateResult.cudaGraphExecUpdateErrorNotSupported #: The update failed because the function of a kernel node changed in #: an unsupported way cudaGraphExecUpdateErrorUnsupportedFunctionChange = cyruntime.cudaGraphExecUpdateResult.cudaGraphExecUpdateErrorUnsupportedFunctionChange #: The update failed because the node attributes changed in a way that #: is not supported cudaGraphExecUpdateErrorAttributesChanged = cyruntime.cudaGraphExecUpdateResult.cudaGraphExecUpdateErrorAttributesChanged _dict_cudaGraphExecUpdateResult = dict(((int(v), v) for k, v in cudaGraphExecUpdateResult.__members__.items())) class cudaGraphKernelNodeField(IntEnum): """ Specifies the field to update when performing multiple node updates from the device """ #: Invalid field cudaGraphKernelNodeFieldInvalid = cyruntime.cudaGraphKernelNodeField.cudaGraphKernelNodeFieldInvalid #: Grid dimension update cudaGraphKernelNodeFieldGridDim = cyruntime.cudaGraphKernelNodeField.cudaGraphKernelNodeFieldGridDim #: Kernel parameter update cudaGraphKernelNodeFieldParam = cyruntime.cudaGraphKernelNodeField.cudaGraphKernelNodeFieldParam #: Node enable/disable cudaGraphKernelNodeFieldEnabled = cyruntime.cudaGraphKernelNodeField.cudaGraphKernelNodeFieldEnabled _dict_cudaGraphKernelNodeField = dict(((int(v), v) for k, v in cudaGraphKernelNodeField.__members__.items())) class cudaGetDriverEntryPointFlags(IntEnum): """ Flags to specify search options to be used with :py:obj:`~.cudaGetDriverEntryPoint` For more details see :py:obj:`~.cuGetProcAddress` """ #: Default search mode for driver symbols. cudaEnableDefault = cyruntime.cudaGetDriverEntryPointFlags.cudaEnableDefault #: Search for legacy versions of driver symbols. cudaEnableLegacyStream = cyruntime.cudaGetDriverEntryPointFlags.cudaEnableLegacyStream #: Search for per-thread versions of driver symbols. cudaEnablePerThreadDefaultStream = cyruntime.cudaGetDriverEntryPointFlags.cudaEnablePerThreadDefaultStream _dict_cudaGetDriverEntryPointFlags = dict(((int(v), v) for k, v in cudaGetDriverEntryPointFlags.__members__.items())) class cudaDriverEntryPointQueryResult(IntEnum): """ Enum for status from obtaining driver entry points, used with :py:obj:`~.cudaApiGetDriverEntryPoint` """ #: Search for symbol found a match cudaDriverEntryPointSuccess = cyruntime.cudaDriverEntryPointQueryResult.cudaDriverEntryPointSuccess #: Search for symbol was not found cudaDriverEntryPointSymbolNotFound = cyruntime.cudaDriverEntryPointQueryResult.cudaDriverEntryPointSymbolNotFound #: Search for symbol was found but version wasn't great enough cudaDriverEntryPointVersionNotSufficent = cyruntime.cudaDriverEntryPointQueryResult.cudaDriverEntryPointVersionNotSufficent _dict_cudaDriverEntryPointQueryResult = dict(((int(v), v) for k, v in cudaDriverEntryPointQueryResult.__members__.items())) class cudaGraphDebugDotFlags(IntEnum): """ CUDA Graph debug write options """ #: Output all debug data as if every debug flag is enabled cudaGraphDebugDotFlagsVerbose = cyruntime.cudaGraphDebugDotFlags.cudaGraphDebugDotFlagsVerbose #: Adds :py:obj:`~.cudaKernelNodeParams` to output cudaGraphDebugDotFlagsKernelNodeParams = cyruntime.cudaGraphDebugDotFlags.cudaGraphDebugDotFlagsKernelNodeParams #: Adds :py:obj:`~.cudaMemcpy3DParms` to output cudaGraphDebugDotFlagsMemcpyNodeParams = cyruntime.cudaGraphDebugDotFlags.cudaGraphDebugDotFlagsMemcpyNodeParams #: Adds :py:obj:`~.cudaMemsetParams` to output cudaGraphDebugDotFlagsMemsetNodeParams = cyruntime.cudaGraphDebugDotFlags.cudaGraphDebugDotFlagsMemsetNodeParams #: Adds :py:obj:`~.cudaHostNodeParams` to output cudaGraphDebugDotFlagsHostNodeParams = cyruntime.cudaGraphDebugDotFlags.cudaGraphDebugDotFlagsHostNodeParams #: Adds cudaEvent_t handle from record and wait nodes to output cudaGraphDebugDotFlagsEventNodeParams = cyruntime.cudaGraphDebugDotFlags.cudaGraphDebugDotFlagsEventNodeParams #: Adds :py:obj:`~.cudaExternalSemaphoreSignalNodeParams` values to #: output cudaGraphDebugDotFlagsExtSemasSignalNodeParams = cyruntime.cudaGraphDebugDotFlags.cudaGraphDebugDotFlagsExtSemasSignalNodeParams #: Adds :py:obj:`~.cudaExternalSemaphoreWaitNodeParams` to output cudaGraphDebugDotFlagsExtSemasWaitNodeParams = cyruntime.cudaGraphDebugDotFlags.cudaGraphDebugDotFlagsExtSemasWaitNodeParams #: Adds cudaKernelNodeAttrID values to output cudaGraphDebugDotFlagsKernelNodeAttributes = cyruntime.cudaGraphDebugDotFlags.cudaGraphDebugDotFlagsKernelNodeAttributes #: Adds node handles and every kernel function handle to output cudaGraphDebugDotFlagsHandles = cyruntime.cudaGraphDebugDotFlags.cudaGraphDebugDotFlagsHandles #: Adds :py:obj:`~.cudaConditionalNodeParams` to output cudaGraphDebugDotFlagsConditionalNodeParams = cyruntime.cudaGraphDebugDotFlags.cudaGraphDebugDotFlagsConditionalNodeParams _dict_cudaGraphDebugDotFlags = dict(((int(v), v) for k, v in cudaGraphDebugDotFlags.__members__.items())) class cudaGraphInstantiateFlags(IntEnum): """ Flags for instantiating a graph """ #: Automatically free memory allocated in a graph before relaunching. cudaGraphInstantiateFlagAutoFreeOnLaunch = cyruntime.cudaGraphInstantiateFlags.cudaGraphInstantiateFlagAutoFreeOnLaunch #: Automatically upload the graph after instantiation. Only supported #: by #: :py:obj:`~.cudaGraphInstantiateWithParams`. The upload will be #: performed using the #: stream provided in `instantiateParams`. cudaGraphInstantiateFlagUpload = cyruntime.cudaGraphInstantiateFlags.cudaGraphInstantiateFlagUpload #: Instantiate the graph to be launchable from the device. This flag #: can only #: be used on platforms which support unified addressing. This flag #: cannot be #: used in conjunction with cudaGraphInstantiateFlagAutoFreeOnLaunch. cudaGraphInstantiateFlagDeviceLaunch = cyruntime.cudaGraphInstantiateFlags.cudaGraphInstantiateFlagDeviceLaunch #: Run the graph using the per-node priority attributes rather than the #: priority of the stream it is launched into. cudaGraphInstantiateFlagUseNodePriority = cyruntime.cudaGraphInstantiateFlags.cudaGraphInstantiateFlagUseNodePriority _dict_cudaGraphInstantiateFlags = dict(((int(v), v) for k, v in cudaGraphInstantiateFlags.__members__.items())) class cudaDeviceNumaConfig(IntEnum): """ CUDA device NUMA config """ #: The GPU is not a NUMA node cudaDeviceNumaConfigNone = cyruntime.cudaDeviceNumaConfig.cudaDeviceNumaConfigNone #: The GPU is a NUMA node, cudaDevAttrNumaId contains its NUMA ID cudaDeviceNumaConfigNumaNode = cyruntime.cudaDeviceNumaConfig.cudaDeviceNumaConfigNumaNode _dict_cudaDeviceNumaConfig = dict(((int(v), v) for k, v in cudaDeviceNumaConfig.__members__.items())) class cudaSurfaceBoundaryMode(IntEnum): """ CUDA Surface boundary modes """ #: Zero boundary mode cudaBoundaryModeZero = cyruntime.cudaSurfaceBoundaryMode.cudaBoundaryModeZero #: Clamp boundary mode cudaBoundaryModeClamp = cyruntime.cudaSurfaceBoundaryMode.cudaBoundaryModeClamp #: Trap boundary mode cudaBoundaryModeTrap = cyruntime.cudaSurfaceBoundaryMode.cudaBoundaryModeTrap _dict_cudaSurfaceBoundaryMode = dict(((int(v), v) for k, v in cudaSurfaceBoundaryMode.__members__.items())) class cudaSurfaceFormatMode(IntEnum): """ CUDA Surface format modes """ #: Forced format mode cudaFormatModeForced = cyruntime.cudaSurfaceFormatMode.cudaFormatModeForced #: Auto format mode cudaFormatModeAuto = cyruntime.cudaSurfaceFormatMode.cudaFormatModeAuto _dict_cudaSurfaceFormatMode = dict(((int(v), v) for k, v in cudaSurfaceFormatMode.__members__.items())) class cudaTextureAddressMode(IntEnum): """ CUDA texture address modes """ #: Wrapping address mode cudaAddressModeWrap = cyruntime.cudaTextureAddressMode.cudaAddressModeWrap #: Clamp to edge address mode cudaAddressModeClamp = cyruntime.cudaTextureAddressMode.cudaAddressModeClamp #: Mirror address mode cudaAddressModeMirror = cyruntime.cudaTextureAddressMode.cudaAddressModeMirror #: Border address mode cudaAddressModeBorder = cyruntime.cudaTextureAddressMode.cudaAddressModeBorder _dict_cudaTextureAddressMode = dict(((int(v), v) for k, v in cudaTextureAddressMode.__members__.items())) class cudaTextureFilterMode(IntEnum): """ CUDA texture filter modes """ #: Point filter mode cudaFilterModePoint = cyruntime.cudaTextureFilterMode.cudaFilterModePoint #: Linear filter mode cudaFilterModeLinear = cyruntime.cudaTextureFilterMode.cudaFilterModeLinear _dict_cudaTextureFilterMode = dict(((int(v), v) for k, v in cudaTextureFilterMode.__members__.items())) class cudaTextureReadMode(IntEnum): """ CUDA texture read modes """ #: Read texture as specified element type cudaReadModeElementType = cyruntime.cudaTextureReadMode.cudaReadModeElementType #: Read texture as normalized float cudaReadModeNormalizedFloat = cyruntime.cudaTextureReadMode.cudaReadModeNormalizedFloat _dict_cudaTextureReadMode = dict(((int(v), v) for k, v in cudaTextureReadMode.__members__.items())) class cudaRoundMode(IntEnum): """""" cudaRoundNearest = cyruntime.cudaRoundMode.cudaRoundNearest cudaRoundZero = cyruntime.cudaRoundMode.cudaRoundZero cudaRoundPosInf = cyruntime.cudaRoundMode.cudaRoundPosInf cudaRoundMinInf = cyruntime.cudaRoundMode.cudaRoundMinInf _dict_cudaRoundMode = dict(((int(v), v) for k, v in cudaRoundMode.__members__.items())) class cudaGLDeviceList(IntEnum): """ CUDA devices corresponding to the current OpenGL context """ #: The CUDA devices for all GPUs used by the current OpenGL context cudaGLDeviceListAll = cyruntime.cudaGLDeviceList.cudaGLDeviceListAll #: The CUDA devices for the GPUs used by the current OpenGL context in #: its currently rendering frame cudaGLDeviceListCurrentFrame = cyruntime.cudaGLDeviceList.cudaGLDeviceListCurrentFrame #: The CUDA devices for the GPUs to be used by the current OpenGL #: context in the next frame cudaGLDeviceListNextFrame = cyruntime.cudaGLDeviceList.cudaGLDeviceListNextFrame _dict_cudaGLDeviceList = dict(((int(v), v) for k, v in cudaGLDeviceList.__members__.items())) class cudaGLMapFlags(IntEnum): """ CUDA GL Map Flags """ #: Default; Assume resource can be read/written cudaGLMapFlagsNone = cyruntime.cudaGLMapFlags.cudaGLMapFlagsNone #: CUDA kernels will not write to this resource cudaGLMapFlagsReadOnly = cyruntime.cudaGLMapFlags.cudaGLMapFlagsReadOnly #: CUDA kernels will only write to and will not read from this resource cudaGLMapFlagsWriteDiscard = cyruntime.cudaGLMapFlags.cudaGLMapFlagsWriteDiscard _dict_cudaGLMapFlags = dict(((int(v), v) for k, v in cudaGLMapFlags.__members__.items())) class cudaStreamAttrID(IntEnum): """ Launch attributes enum; used as id field of :py:obj:`~.cudaLaunchAttribute` """ #: Ignored entry, for convenient composition cudaLaunchAttributeIgnore = cyruntime.cudaLaunchAttributeID.cudaLaunchAttributeIgnore #: Valid for streams, graph nodes, launches. See #: :py:obj:`~.cudaLaunchAttributeValue.accessPolicyWindow`. cudaLaunchAttributeAccessPolicyWindow = cyruntime.cudaLaunchAttributeID.cudaLaunchAttributeAccessPolicyWindow #: Valid for graph nodes, launches. See #: :py:obj:`~.cudaLaunchAttributeValue.cooperative`. cudaLaunchAttributeCooperative = cyruntime.cudaLaunchAttributeID.cudaLaunchAttributeCooperative #: Valid for streams. See #: :py:obj:`~.cudaLaunchAttributeValue.syncPolicy`. cudaLaunchAttributeSynchronizationPolicy = cyruntime.cudaLaunchAttributeID.cudaLaunchAttributeSynchronizationPolicy #: Valid for graph nodes, launches. See #: :py:obj:`~.cudaLaunchAttributeValue.clusterDim`. cudaLaunchAttributeClusterDimension = cyruntime.cudaLaunchAttributeID.cudaLaunchAttributeClusterDimension #: Valid for graph nodes, launches. See #: :py:obj:`~.cudaLaunchAttributeValue.clusterSchedulingPolicyPreference`. cudaLaunchAttributeClusterSchedulingPolicyPreference = cyruntime.cudaLaunchAttributeID.cudaLaunchAttributeClusterSchedulingPolicyPreference #: Valid for launches. Setting #: :py:obj:`~.cudaLaunchAttributeValue.programmaticStreamSerializationAllowed` #: to non-0 signals that the kernel will use programmatic means to #: resolve its stream dependency, so that the CUDA runtime should #: opportunistically allow the grid's execution to overlap with the #: previous kernel in the stream, if that kernel requests the overlap. #: The dependent launches can choose to wait on the dependency using #: the programmatic sync (cudaGridDependencySynchronize() or equivalent #: PTX instructions). cudaLaunchAttributeProgrammaticStreamSerialization = cyruntime.cudaLaunchAttributeID.cudaLaunchAttributeProgrammaticStreamSerialization #: Valid for launches. Set #: :py:obj:`~.cudaLaunchAttributeValue.programmaticEvent` to record the #: event. Event recorded through this launch attribute is guaranteed to #: only trigger after all block in the associated kernel trigger the #: event. A block can trigger the event programmatically in a future #: CUDA release. A trigger can also be inserted at the beginning of #: each block's execution if triggerAtBlockStart is set to non-0. The #: dependent launches can choose to wait on the dependency using the #: programmatic sync (cudaGridDependencySynchronize() or equivalent PTX #: instructions). Note that dependents (including the CPU thread #: calling :py:obj:`~.cudaEventSynchronize()`) are not guaranteed to #: observe the release precisely when it is released. For example, #: :py:obj:`~.cudaEventSynchronize()` may only observe the event #: trigger long after the associated kernel has completed. This #: recording type is primarily meant for establishing programmatic #: dependency between device tasks. Note also this type of dependency #: allows, but does not guarantee, concurrent execution of tasks. #: The event supplied must not be an interprocess or interop event. #: The event must disable timing (i.e. must be created with the #: :py:obj:`~.cudaEventDisableTiming` flag set). cudaLaunchAttributeProgrammaticEvent = cyruntime.cudaLaunchAttributeID.cudaLaunchAttributeProgrammaticEvent #: Valid for streams, graph nodes, launches. See #: :py:obj:`~.cudaLaunchAttributeValue.priority`. cudaLaunchAttributePriority = cyruntime.cudaLaunchAttributeID.cudaLaunchAttributePriority #: Valid for streams, graph nodes, launches. See #: :py:obj:`~.cudaLaunchAttributeValue.memSyncDomainMap`. cudaLaunchAttributeMemSyncDomainMap = cyruntime.cudaLaunchAttributeID.cudaLaunchAttributeMemSyncDomainMap #: Valid for streams, graph nodes, launches. See #: :py:obj:`~.cudaLaunchAttributeValue.memSyncDomain`. cudaLaunchAttributeMemSyncDomain = cyruntime.cudaLaunchAttributeID.cudaLaunchAttributeMemSyncDomain #: Valid for graph nodes and launches. Set #: :py:obj:`~.cudaLaunchAttributeValue.preferredClusterDim` to allow #: the kernel launch to specify a preferred substitute cluster #: dimension. Blocks may be grouped according to either the dimensions #: specified with this attribute (grouped into a "preferred substitute #: cluster"), or the one specified with #: :py:obj:`~.cudaLaunchAttributeClusterDimension` attribute (grouped #: into a "regular cluster"). The cluster dimensions of a "preferred #: substitute cluster" shall be an integer multiple greater than zero #: of the regular cluster dimensions. The device will attempt - on a #: best-effort basis - to group thread blocks into preferred clusters #: over grouping them into regular clusters. When it deems necessary #: (primarily when the device temporarily runs out of physical #: resources to launch the larger preferred clusters), the device may #: switch to launch the regular clusters instead to attempt to utilize #: as much of the physical device resources as possible. #: Each type of cluster will have its enumeration / coordinate setup #: as if the grid consists solely of its type of cluster. For example, #: if the preferred substitute cluster dimensions double the regular #: cluster dimensions, there might be simultaneously a regular cluster #: indexed at (1,0,0), and a preferred cluster indexed at (1,0,0). In #: this example, the preferred substitute cluster (1,0,0) replaces #: regular clusters (2,0,0) and (3,0,0) and groups their blocks. #: This attribute will only take effect when a regular cluster #: dimension has been specified. The preferred substitute cluster #: dimension must be an integer multiple greater than zero of the #: regular cluster dimension and must divide the grid. It must also be #: no more than `maxBlocksPerCluster`, if it is set in the kernel's #: `__launch_bounds__`. Otherwise it must be less than the maximum #: value the driver can support. Otherwise, setting this attribute to a #: value physically unable to fit on any particular device is #: permitted. cudaLaunchAttributePreferredClusterDimension = cyruntime.cudaLaunchAttributeID.cudaLaunchAttributePreferredClusterDimension #: Valid for launches. Set #: :py:obj:`~.cudaLaunchAttributeValue.launchCompletionEvent` to record #: the event. #: Nominally, the event is triggered once all blocks of the kernel #: have begun execution. Currently this is a best effort. If a kernel B #: has a launch completion dependency on a kernel A, B may wait until A #: is complete. Alternatively, blocks of B may begin before all blocks #: of A have begun, for example if B can claim execution resources #: unavailable to A (e.g. they run on different GPUs) or if B is a #: higher priority than A. Exercise caution if such an ordering #: inversion could lead to deadlock. #: A launch completion event is nominally similar to a programmatic #: event with `triggerAtBlockStart` set except that it is not visible #: to `cudaGridDependencySynchronize()` and can be used with compute #: capability less than 9.0. #: The event supplied must not be an interprocess or interop event. #: The event must disable timing (i.e. must be created with the #: :py:obj:`~.cudaEventDisableTiming` flag set). cudaLaunchAttributeLaunchCompletionEvent = cyruntime.cudaLaunchAttributeID.cudaLaunchAttributeLaunchCompletionEvent #: Valid for graph nodes, launches. This attribute is graphs-only, and #: passing it to a launch in a non-capturing stream will result in an #: error. #: :cudaLaunchAttributeValue::deviceUpdatableKernelNode::deviceUpdatable #: can only be set to 0 or 1. Setting the field to 1 indicates that the #: corresponding kernel node should be device-updatable. On success, a #: handle will be returned via #: :py:obj:`~.cudaLaunchAttributeValue`::deviceUpdatableKernelNode::devNode #: which can be passed to the various device-side update functions to #: update the node's kernel parameters from within another kernel. For #: more information on the types of device updates that can be made, as #: well as the relevant limitations thereof, see #: :py:obj:`~.cudaGraphKernelNodeUpdatesApply`. #: Nodes which are device-updatable have additional restrictions #: compared to regular kernel nodes. Firstly, device-updatable nodes #: cannot be removed from their graph via #: :py:obj:`~.cudaGraphDestroyNode`. Additionally, once opted-in to #: this functionality, a node cannot opt out, and any attempt to set #: the deviceUpdatable attribute to 0 will result in an error. Device- #: updatable kernel nodes also cannot have their attributes copied #: to/from another kernel node via #: :py:obj:`~.cudaGraphKernelNodeCopyAttributes`. Graphs containing one #: or more device-updatable nodes also do not allow multiple #: instantiation, and neither the graph nor its instantiated version #: can be passed to :py:obj:`~.cudaGraphExecUpdate`. #: If a graph contains device-updatable nodes and updates those nodes #: from the device from within the graph, the graph must be uploaded #: with :py:obj:`~.cuGraphUpload` before it is launched. For such a #: graph, if host-side executable graph updates are made to the device- #: updatable nodes, the graph must be uploaded before it is launched #: again. cudaLaunchAttributeDeviceUpdatableKernelNode = cyruntime.cudaLaunchAttributeID.cudaLaunchAttributeDeviceUpdatableKernelNode #: Valid for launches. On devices where the L1 cache and shared memory #: use the same hardware resources, setting #: :py:obj:`~.cudaLaunchAttributeValue.sharedMemCarveout` to a #: percentage between 0-100 signals sets the shared memory carveout #: preference in percent of the total shared memory for that kernel #: launch. This attribute takes precedence over #: :py:obj:`~.cudaFuncAttributePreferredSharedMemoryCarveout`. This is #: only a hint, and the driver can choose a different configuration if #: required for the launch. cudaLaunchAttributePreferredSharedMemoryCarveout = cyruntime.cudaLaunchAttributeID.cudaLaunchAttributePreferredSharedMemoryCarveout _dict_cudaLaunchAttributeID = dict(((int(v), v) for k, v in cudaLaunchAttributeID.__members__.items())) class cudaKernelNodeAttrID(IntEnum): """ Launch attributes enum; used as id field of :py:obj:`~.cudaLaunchAttribute` """ #: Ignored entry, for convenient composition cudaLaunchAttributeIgnore = cyruntime.cudaLaunchAttributeID.cudaLaunchAttributeIgnore #: Valid for streams, graph nodes, launches. See #: :py:obj:`~.cudaLaunchAttributeValue.accessPolicyWindow`. cudaLaunchAttributeAccessPolicyWindow = cyruntime.cudaLaunchAttributeID.cudaLaunchAttributeAccessPolicyWindow #: Valid for graph nodes, launches. See #: :py:obj:`~.cudaLaunchAttributeValue.cooperative`. cudaLaunchAttributeCooperative = cyruntime.cudaLaunchAttributeID.cudaLaunchAttributeCooperative #: Valid for streams. See #: :py:obj:`~.cudaLaunchAttributeValue.syncPolicy`. cudaLaunchAttributeSynchronizationPolicy = cyruntime.cudaLaunchAttributeID.cudaLaunchAttributeSynchronizationPolicy #: Valid for graph nodes, launches. See #: :py:obj:`~.cudaLaunchAttributeValue.clusterDim`. cudaLaunchAttributeClusterDimension = cyruntime.cudaLaunchAttributeID.cudaLaunchAttributeClusterDimension #: Valid for graph nodes, launches. See #: :py:obj:`~.cudaLaunchAttributeValue.clusterSchedulingPolicyPreference`. cudaLaunchAttributeClusterSchedulingPolicyPreference = cyruntime.cudaLaunchAttributeID.cudaLaunchAttributeClusterSchedulingPolicyPreference #: Valid for launches. Setting #: :py:obj:`~.cudaLaunchAttributeValue.programmaticStreamSerializationAllowed` #: to non-0 signals that the kernel will use programmatic means to #: resolve its stream dependency, so that the CUDA runtime should #: opportunistically allow the grid's execution to overlap with the #: previous kernel in the stream, if that kernel requests the overlap. #: The dependent launches can choose to wait on the dependency using #: the programmatic sync (cudaGridDependencySynchronize() or equivalent #: PTX instructions). cudaLaunchAttributeProgrammaticStreamSerialization = cyruntime.cudaLaunchAttributeID.cudaLaunchAttributeProgrammaticStreamSerialization #: Valid for launches. Set #: :py:obj:`~.cudaLaunchAttributeValue.programmaticEvent` to record the #: event. Event recorded through this launch attribute is guaranteed to #: only trigger after all block in the associated kernel trigger the #: event. A block can trigger the event programmatically in a future #: CUDA release. A trigger can also be inserted at the beginning of #: each block's execution if triggerAtBlockStart is set to non-0. The #: dependent launches can choose to wait on the dependency using the #: programmatic sync (cudaGridDependencySynchronize() or equivalent PTX #: instructions). Note that dependents (including the CPU thread #: calling :py:obj:`~.cudaEventSynchronize()`) are not guaranteed to #: observe the release precisely when it is released. For example, #: :py:obj:`~.cudaEventSynchronize()` may only observe the event #: trigger long after the associated kernel has completed. This #: recording type is primarily meant for establishing programmatic #: dependency between device tasks. Note also this type of dependency #: allows, but does not guarantee, concurrent execution of tasks. #: The event supplied must not be an interprocess or interop event. #: The event must disable timing (i.e. must be created with the #: :py:obj:`~.cudaEventDisableTiming` flag set). cudaLaunchAttributeProgrammaticEvent = cyruntime.cudaLaunchAttributeID.cudaLaunchAttributeProgrammaticEvent #: Valid for streams, graph nodes, launches. See #: :py:obj:`~.cudaLaunchAttributeValue.priority`. cudaLaunchAttributePriority = cyruntime.cudaLaunchAttributeID.cudaLaunchAttributePriority #: Valid for streams, graph nodes, launches. See #: :py:obj:`~.cudaLaunchAttributeValue.memSyncDomainMap`. cudaLaunchAttributeMemSyncDomainMap = cyruntime.cudaLaunchAttributeID.cudaLaunchAttributeMemSyncDomainMap #: Valid for streams, graph nodes, launches. See #: :py:obj:`~.cudaLaunchAttributeValue.memSyncDomain`. cudaLaunchAttributeMemSyncDomain = cyruntime.cudaLaunchAttributeID.cudaLaunchAttributeMemSyncDomain #: Valid for graph nodes and launches. Set #: :py:obj:`~.cudaLaunchAttributeValue.preferredClusterDim` to allow #: the kernel launch to specify a preferred substitute cluster #: dimension. Blocks may be grouped according to either the dimensions #: specified with this attribute (grouped into a "preferred substitute #: cluster"), or the one specified with #: :py:obj:`~.cudaLaunchAttributeClusterDimension` attribute (grouped #: into a "regular cluster"). The cluster dimensions of a "preferred #: substitute cluster" shall be an integer multiple greater than zero #: of the regular cluster dimensions. The device will attempt - on a #: best-effort basis - to group thread blocks into preferred clusters #: over grouping them into regular clusters. When it deems necessary #: (primarily when the device temporarily runs out of physical #: resources to launch the larger preferred clusters), the device may #: switch to launch the regular clusters instead to attempt to utilize #: as much of the physical device resources as possible. #: Each type of cluster will have its enumeration / coordinate setup #: as if the grid consists solely of its type of cluster. For example, #: if the preferred substitute cluster dimensions double the regular #: cluster dimensions, there might be simultaneously a regular cluster #: indexed at (1,0,0), and a preferred cluster indexed at (1,0,0). In #: this example, the preferred substitute cluster (1,0,0) replaces #: regular clusters (2,0,0) and (3,0,0) and groups their blocks. #: This attribute will only take effect when a regular cluster #: dimension has been specified. The preferred substitute cluster #: dimension must be an integer multiple greater than zero of the #: regular cluster dimension and must divide the grid. It must also be #: no more than `maxBlocksPerCluster`, if it is set in the kernel's #: `__launch_bounds__`. Otherwise it must be less than the maximum #: value the driver can support. Otherwise, setting this attribute to a #: value physically unable to fit on any particular device is #: permitted. cudaLaunchAttributePreferredClusterDimension = cyruntime.cudaLaunchAttributeID.cudaLaunchAttributePreferredClusterDimension #: Valid for launches. Set #: :py:obj:`~.cudaLaunchAttributeValue.launchCompletionEvent` to record #: the event. #: Nominally, the event is triggered once all blocks of the kernel #: have begun execution. Currently this is a best effort. If a kernel B #: has a launch completion dependency on a kernel A, B may wait until A #: is complete. Alternatively, blocks of B may begin before all blocks #: of A have begun, for example if B can claim execution resources #: unavailable to A (e.g. they run on different GPUs) or if B is a #: higher priority than A. Exercise caution if such an ordering #: inversion could lead to deadlock. #: A launch completion event is nominally similar to a programmatic #: event with `triggerAtBlockStart` set except that it is not visible #: to `cudaGridDependencySynchronize()` and can be used with compute #: capability less than 9.0. #: The event supplied must not be an interprocess or interop event. #: The event must disable timing (i.e. must be created with the #: :py:obj:`~.cudaEventDisableTiming` flag set). cudaLaunchAttributeLaunchCompletionEvent = cyruntime.cudaLaunchAttributeID.cudaLaunchAttributeLaunchCompletionEvent #: Valid for graph nodes, launches. This attribute is graphs-only, and #: passing it to a launch in a non-capturing stream will result in an #: error. #: :cudaLaunchAttributeValue::deviceUpdatableKernelNode::deviceUpdatable #: can only be set to 0 or 1. Setting the field to 1 indicates that the #: corresponding kernel node should be device-updatable. On success, a #: handle will be returned via #: :py:obj:`~.cudaLaunchAttributeValue`::deviceUpdatableKernelNode::devNode #: which can be passed to the various device-side update functions to #: update the node's kernel parameters from within another kernel. For #: more information on the types of device updates that can be made, as #: well as the relevant limitations thereof, see #: :py:obj:`~.cudaGraphKernelNodeUpdatesApply`. #: Nodes which are device-updatable have additional restrictions #: compared to regular kernel nodes. Firstly, device-updatable nodes #: cannot be removed from their graph via #: :py:obj:`~.cudaGraphDestroyNode`. Additionally, once opted-in to #: this functionality, a node cannot opt out, and any attempt to set #: the deviceUpdatable attribute to 0 will result in an error. Device- #: updatable kernel nodes also cannot have their attributes copied #: to/from another kernel node via #: :py:obj:`~.cudaGraphKernelNodeCopyAttributes`. Graphs containing one #: or more device-updatable nodes also do not allow multiple #: instantiation, and neither the graph nor its instantiated version #: can be passed to :py:obj:`~.cudaGraphExecUpdate`. #: If a graph contains device-updatable nodes and updates those nodes #: from the device from within the graph, the graph must be uploaded #: with :py:obj:`~.cuGraphUpload` before it is launched. For such a #: graph, if host-side executable graph updates are made to the device- #: updatable nodes, the graph must be uploaded before it is launched #: again. cudaLaunchAttributeDeviceUpdatableKernelNode = cyruntime.cudaLaunchAttributeID.cudaLaunchAttributeDeviceUpdatableKernelNode #: Valid for launches. On devices where the L1 cache and shared memory #: use the same hardware resources, setting #: :py:obj:`~.cudaLaunchAttributeValue.sharedMemCarveout` to a #: percentage between 0-100 signals sets the shared memory carveout #: preference in percent of the total shared memory for that kernel #: launch. This attribute takes precedence over #: :py:obj:`~.cudaFuncAttributePreferredSharedMemoryCarveout`. This is #: only a hint, and the driver can choose a different configuration if #: required for the launch. cudaLaunchAttributePreferredSharedMemoryCarveout = cyruntime.cudaLaunchAttributeID.cudaLaunchAttributePreferredSharedMemoryCarveout _dict_cudaLaunchAttributeID = dict(((int(v), v) for k, v in cudaLaunchAttributeID.__members__.items())) cdef class cudaArray_t: """ CUDA array Methods ------- getPtr() Get memory address of class instance """ def __cinit__(self, void_ptr init_value = 0, void_ptr _ptr = 0): if _ptr == 0: self._pvt_ptr = &self._pvt_val self._pvt_ptr[0] = init_value else: self._pvt_ptr = _ptr def __init__(self, *args, **kwargs): pass def __repr__(self): return '' def __index__(self): return self.__int__() def __eq__(self, other): if not isinstance(other, cudaArray_t): return False return self._pvt_ptr[0] == (other)._pvt_ptr[0] def __hash__(self): return hash((self._pvt_ptr[0])) def __int__(self): return self._pvt_ptr[0] def getPtr(self): return self._pvt_ptr cdef class cudaArray_const_t: """ CUDA array (as source copy argument) Methods ------- getPtr() Get memory address of class instance """ def __cinit__(self, void_ptr init_value = 0, void_ptr _ptr = 0): if _ptr == 0: self._pvt_ptr = &self._pvt_val self._pvt_ptr[0] = init_value else: self._pvt_ptr = _ptr def __init__(self, *args, **kwargs): pass def __repr__(self): return '' def __index__(self): return self.__int__() def __eq__(self, other): if not isinstance(other, cudaArray_const_t): return False return self._pvt_ptr[0] == (other)._pvt_ptr[0] def __hash__(self): return hash((self._pvt_ptr[0])) def __int__(self): return self._pvt_ptr[0] def getPtr(self): return self._pvt_ptr cdef class cudaMipmappedArray_t: """ CUDA mipmapped array Methods ------- getPtr() Get memory address of class instance """ def __cinit__(self, void_ptr init_value = 0, void_ptr _ptr = 0): if _ptr == 0: self._pvt_ptr = &self._pvt_val self._pvt_ptr[0] = init_value else: self._pvt_ptr = _ptr def __init__(self, *args, **kwargs): pass def __repr__(self): return '' def __index__(self): return self.__int__() def __eq__(self, other): if not isinstance(other, cudaMipmappedArray_t): return False return self._pvt_ptr[0] == (other)._pvt_ptr[0] def __hash__(self): return hash((self._pvt_ptr[0])) def __int__(self): return self._pvt_ptr[0] def getPtr(self): return self._pvt_ptr cdef class cudaMipmappedArray_const_t: """ CUDA mipmapped array (as source argument) Methods ------- getPtr() Get memory address of class instance """ def __cinit__(self, void_ptr init_value = 0, void_ptr _ptr = 0): if _ptr == 0: self._pvt_ptr = &self._pvt_val self._pvt_ptr[0] = init_value else: self._pvt_ptr = _ptr def __init__(self, *args, **kwargs): pass def __repr__(self): return '' def __index__(self): return self.__int__() def __eq__(self, other): if not isinstance(other, cudaMipmappedArray_const_t): return False return self._pvt_ptr[0] == (other)._pvt_ptr[0] def __hash__(self): return hash((self._pvt_ptr[0])) def __int__(self): return self._pvt_ptr[0] def getPtr(self): return self._pvt_ptr cdef class cudaGraphicsResource_t: """ CUDA graphics resource types Methods ------- getPtr() Get memory address of class instance """ def __cinit__(self, void_ptr init_value = 0, void_ptr _ptr = 0): if _ptr == 0: self._pvt_ptr = &self._pvt_val self._pvt_ptr[0] = init_value else: self._pvt_ptr = _ptr def __init__(self, *args, **kwargs): pass def __repr__(self): return '' def __index__(self): return self.__int__() def __eq__(self, other): if not isinstance(other, cudaGraphicsResource_t): return False return self._pvt_ptr[0] == (other)._pvt_ptr[0] def __hash__(self): return hash((self._pvt_ptr[0])) def __int__(self): return self._pvt_ptr[0] def getPtr(self): return self._pvt_ptr cdef class cudaExternalMemory_t: """ CUDA external memory Methods ------- getPtr() Get memory address of class instance """ def __cinit__(self, void_ptr init_value = 0, void_ptr _ptr = 0): if _ptr == 0: self._pvt_ptr = &self._pvt_val self._pvt_ptr[0] = init_value else: self._pvt_ptr = _ptr def __init__(self, *args, **kwargs): pass def __repr__(self): return '' def __index__(self): return self.__int__() def __eq__(self, other): if not isinstance(other, cudaExternalMemory_t): return False return self._pvt_ptr[0] == (other)._pvt_ptr[0] def __hash__(self): return hash((self._pvt_ptr[0])) def __int__(self): return self._pvt_ptr[0] def getPtr(self): return self._pvt_ptr cdef class cudaExternalSemaphore_t: """ CUDA external semaphore Methods ------- getPtr() Get memory address of class instance """ def __cinit__(self, void_ptr init_value = 0, void_ptr _ptr = 0): if _ptr == 0: self._pvt_ptr = &self._pvt_val self._pvt_ptr[0] = init_value else: self._pvt_ptr = _ptr def __init__(self, *args, **kwargs): pass def __repr__(self): return '' def __index__(self): return self.__int__() def __eq__(self, other): if not isinstance(other, cudaExternalSemaphore_t): return False return self._pvt_ptr[0] == (other)._pvt_ptr[0] def __hash__(self): return hash((self._pvt_ptr[0])) def __int__(self): return self._pvt_ptr[0] def getPtr(self): return self._pvt_ptr cdef class cudaKernel_t: """ CUDA kernel Methods ------- getPtr() Get memory address of class instance """ def __cinit__(self, void_ptr init_value = 0, void_ptr _ptr = 0): if _ptr == 0: self._pvt_ptr = &self._pvt_val self._pvt_ptr[0] = init_value else: self._pvt_ptr = _ptr def __init__(self, *args, **kwargs): pass def __repr__(self): return '' def __index__(self): return self.__int__() def __eq__(self, other): if not isinstance(other, cudaKernel_t): return False return self._pvt_ptr[0] == (other)._pvt_ptr[0] def __hash__(self): return hash((self._pvt_ptr[0])) def __int__(self): return self._pvt_ptr[0] def getPtr(self): return self._pvt_ptr cdef class cudaLibrary_t: """ CUDA library Methods ------- getPtr() Get memory address of class instance """ def __cinit__(self, void_ptr init_value = 0, void_ptr _ptr = 0): if _ptr == 0: self._pvt_ptr = &self._pvt_val self._pvt_ptr[0] = init_value else: self._pvt_ptr = _ptr def __init__(self, *args, **kwargs): pass def __repr__(self): return '' def __index__(self): return self.__int__() def __eq__(self, other): if not isinstance(other, cudaLibrary_t): return False return self._pvt_ptr[0] == (other)._pvt_ptr[0] def __hash__(self): return hash((self._pvt_ptr[0])) def __int__(self): return self._pvt_ptr[0] def getPtr(self): return self._pvt_ptr cdef class cudaGraphDeviceNode_t: """ CUDA device node handle for device-side node update Methods ------- getPtr() Get memory address of class instance """ def __cinit__(self, void_ptr init_value = 0, void_ptr _ptr = 0): if _ptr == 0: self._pvt_ptr = &self._pvt_val self._pvt_ptr[0] = init_value else: self._pvt_ptr = _ptr def __init__(self, *args, **kwargs): pass def __repr__(self): return '' def __index__(self): return self.__int__() def __eq__(self, other): if not isinstance(other, cudaGraphDeviceNode_t): return False return self._pvt_ptr[0] == (other)._pvt_ptr[0] def __hash__(self): return hash((self._pvt_ptr[0])) def __int__(self): return self._pvt_ptr[0] def getPtr(self): return self._pvt_ptr cdef class cudaAsyncCallbackHandle_t: """ CUDA async callback handle Methods ------- getPtr() Get memory address of class instance """ def __cinit__(self, void_ptr init_value = 0, void_ptr _ptr = 0): if _ptr == 0: self._pvt_ptr = &self._pvt_val self._pvt_ptr[0] = init_value else: self._pvt_ptr = _ptr def __init__(self, *args, **kwargs): pass def __repr__(self): return '' def __index__(self): return self.__int__() def __eq__(self, other): if not isinstance(other, cudaAsyncCallbackHandle_t): return False return self._pvt_ptr[0] == (other)._pvt_ptr[0] def __hash__(self): return hash((self._pvt_ptr[0])) def __int__(self): return self._pvt_ptr[0] def getPtr(self): return self._pvt_ptr cdef class EGLImageKHR: """ Methods ------- getPtr() Get memory address of class instance """ def __cinit__(self, void_ptr init_value = 0, void_ptr _ptr = 0): if _ptr == 0: self._pvt_ptr = &self._pvt_val self._pvt_ptr[0] = init_value else: self._pvt_ptr = _ptr def __init__(self, *args, **kwargs): pass def __repr__(self): return '' def __index__(self): return self.__int__() def __eq__(self, other): if not isinstance(other, EGLImageKHR): return False return self._pvt_ptr[0] == (other)._pvt_ptr[0] def __hash__(self): return hash((self._pvt_ptr[0])) def __int__(self): return self._pvt_ptr[0] def getPtr(self): return self._pvt_ptr cdef class EGLStreamKHR: """ Methods ------- getPtr() Get memory address of class instance """ def __cinit__(self, void_ptr init_value = 0, void_ptr _ptr = 0): if _ptr == 0: self._pvt_ptr = &self._pvt_val self._pvt_ptr[0] = init_value else: self._pvt_ptr = _ptr def __init__(self, *args, **kwargs): pass def __repr__(self): return '' def __index__(self): return self.__int__() def __eq__(self, other): if not isinstance(other, EGLStreamKHR): return False return self._pvt_ptr[0] == (other)._pvt_ptr[0] def __hash__(self): return hash((self._pvt_ptr[0])) def __int__(self): return self._pvt_ptr[0] def getPtr(self): return self._pvt_ptr cdef class EGLSyncKHR: """ Methods ------- getPtr() Get memory address of class instance """ def __cinit__(self, void_ptr init_value = 0, void_ptr _ptr = 0): if _ptr == 0: self._pvt_ptr = &self._pvt_val self._pvt_ptr[0] = init_value else: self._pvt_ptr = _ptr def __init__(self, *args, **kwargs): pass def __repr__(self): return '' def __index__(self): return self.__int__() def __eq__(self, other): if not isinstance(other, EGLSyncKHR): return False return self._pvt_ptr[0] == (other)._pvt_ptr[0] def __hash__(self): return hash((self._pvt_ptr[0])) def __int__(self): return self._pvt_ptr[0] def getPtr(self): return self._pvt_ptr cdef class cudaHostFn_t: """ Methods ------- getPtr() Get memory address of class instance """ def __cinit__(self, void_ptr init_value = 0, void_ptr _ptr = 0): if _ptr == 0: self._pvt_ptr = &self._pvt_val self._pvt_ptr[0] = init_value else: self._pvt_ptr = _ptr def __init__(self, *args, **kwargs): pass def __repr__(self): return '' def __index__(self): return self.__int__() def __int__(self): return self._pvt_ptr[0] def getPtr(self): return self._pvt_ptr cdef class cudaAsyncCallback: """ Methods ------- getPtr() Get memory address of class instance """ def __cinit__(self, void_ptr init_value = 0, void_ptr _ptr = 0): if _ptr == 0: self._pvt_ptr = &self._pvt_val self._pvt_ptr[0] = init_value else: self._pvt_ptr = _ptr def __init__(self, *args, **kwargs): pass def __repr__(self): return '' def __index__(self): return self.__int__() def __int__(self): return self._pvt_ptr[0] def getPtr(self): return self._pvt_ptr cdef class cudaStreamCallback_t: """ Methods ------- getPtr() Get memory address of class instance """ def __cinit__(self, void_ptr init_value = 0, void_ptr _ptr = 0): if _ptr == 0: self._pvt_ptr = &self._pvt_val self._pvt_ptr[0] = init_value else: self._pvt_ptr = _ptr def __init__(self, *args, **kwargs): pass def __repr__(self): return '' def __index__(self): return self.__int__() def __int__(self): return self._pvt_ptr[0] def getPtr(self): return self._pvt_ptr cdef class dim3: """ Attributes ---------- x : unsigned int y : unsigned int z : unsigned int Methods ------- getPtr() Get memory address of class instance """ def __cinit__(self, void_ptr _ptr = 0): if _ptr == 0: self._pvt_ptr = &self._pvt_val else: self._pvt_ptr = _ptr def __init__(self, void_ptr _ptr = 0): pass def __dealloc__(self): pass def getPtr(self): return self._pvt_ptr def __repr__(self): if self._pvt_ptr is not NULL: str_list = [] try: str_list += ['x : ' + str(self.x)] except ValueError: str_list += ['x : '] try: str_list += ['y : ' + str(self.y)] except ValueError: str_list += ['y : '] try: str_list += ['z : ' + str(self.z)] except ValueError: str_list += ['z : '] return '\n'.join(str_list) else: return '' @property def x(self): return self._pvt_ptr[0].x @x.setter def x(self, unsigned int x): self._pvt_ptr[0].x = x @property def y(self): return self._pvt_ptr[0].y @y.setter def y(self, unsigned int y): self._pvt_ptr[0].y = y @property def z(self): return self._pvt_ptr[0].z @z.setter def z(self, unsigned int z): self._pvt_ptr[0].z = z cdef class cudaChannelFormatDesc: """ CUDA Channel format descriptor Attributes ---------- x : int x y : int y z : int z w : int w f : cudaChannelFormatKind Channel format kind Methods ------- getPtr() Get memory address of class instance """ def __cinit__(self, void_ptr _ptr = 0): if _ptr == 0: self._pvt_ptr = &self._pvt_val else: self._pvt_ptr = _ptr def __init__(self, void_ptr _ptr = 0): pass def __dealloc__(self): pass def getPtr(self): return self._pvt_ptr def __repr__(self): if self._pvt_ptr is not NULL: str_list = [] try: str_list += ['x : ' + str(self.x)] except ValueError: str_list += ['x : '] try: str_list += ['y : ' + str(self.y)] except ValueError: str_list += ['y : '] try: str_list += ['z : ' + str(self.z)] except ValueError: str_list += ['z : '] try: str_list += ['w : ' + str(self.w)] except ValueError: str_list += ['w : '] try: str_list += ['f : ' + str(self.f)] except ValueError: str_list += ['f : '] return '\n'.join(str_list) else: return '' @property def x(self): return self._pvt_ptr[0].x @x.setter def x(self, int x): self._pvt_ptr[0].x = x @property def y(self): return self._pvt_ptr[0].y @y.setter def y(self, int y): self._pvt_ptr[0].y = y @property def z(self): return self._pvt_ptr[0].z @z.setter def z(self, int z): self._pvt_ptr[0].z = z @property def w(self): return self._pvt_ptr[0].w @w.setter def w(self, int w): self._pvt_ptr[0].w = w @property def f(self): if self._pvt_ptr[0].f not in _dict_cudaChannelFormatKind: return None return _dict_cudaChannelFormatKind[self._pvt_ptr[0].f] @f.setter def f(self, f not None : cudaChannelFormatKind): self._pvt_ptr[0].f = f.value cdef class anon_struct0: """ Attributes ---------- width : unsigned int height : unsigned int depth : unsigned int Methods ------- getPtr() Get memory address of class instance """ def __cinit__(self, void_ptr _ptr): self._pvt_ptr = _ptr def __init__(self, void_ptr _ptr): pass def __dealloc__(self): pass def getPtr(self): return &self._pvt_ptr[0].tileExtent def __repr__(self): if self._pvt_ptr is not NULL: str_list = [] try: str_list += ['width : ' + str(self.width)] except ValueError: str_list += ['width : '] try: str_list += ['height : ' + str(self.height)] except ValueError: str_list += ['height : '] try: str_list += ['depth : ' + str(self.depth)] except ValueError: str_list += ['depth : '] return '\n'.join(str_list) else: return '' @property def width(self): return self._pvt_ptr[0].tileExtent.width @width.setter def width(self, unsigned int width): self._pvt_ptr[0].tileExtent.width = width @property def height(self): return self._pvt_ptr[0].tileExtent.height @height.setter def height(self, unsigned int height): self._pvt_ptr[0].tileExtent.height = height @property def depth(self): return self._pvt_ptr[0].tileExtent.depth @depth.setter def depth(self, unsigned int depth): self._pvt_ptr[0].tileExtent.depth = depth cdef class cudaArraySparseProperties: """ Sparse CUDA array and CUDA mipmapped array properties Attributes ---------- tileExtent : anon_struct0 miptailFirstLevel : unsigned int First mip level at which the mip tail begins miptailSize : unsigned long long Total size of the mip tail. flags : unsigned int Flags will either be zero or cudaArraySparsePropertiesSingleMipTail reserved : list[unsigned int] Methods ------- getPtr() Get memory address of class instance """ def __cinit__(self, void_ptr _ptr = 0): if _ptr == 0: self._pvt_ptr = &self._pvt_val else: self._pvt_ptr = _ptr def __init__(self, void_ptr _ptr = 0): pass self._tileExtent = anon_struct0(_ptr=self._pvt_ptr) def __dealloc__(self): pass def getPtr(self): return self._pvt_ptr def __repr__(self): if self._pvt_ptr is not NULL: str_list = [] try: str_list += ['tileExtent :\n' + '\n'.join([' ' + line for line in str(self.tileExtent).splitlines()])] except ValueError: str_list += ['tileExtent : '] try: str_list += ['miptailFirstLevel : ' + str(self.miptailFirstLevel)] except ValueError: str_list += ['miptailFirstLevel : '] try: str_list += ['miptailSize : ' + str(self.miptailSize)] except ValueError: str_list += ['miptailSize : '] try: str_list += ['flags : ' + str(self.flags)] except ValueError: str_list += ['flags : '] try: str_list += ['reserved : ' + str(self.reserved)] except ValueError: str_list += ['reserved : '] return '\n'.join(str_list) else: return '' @property def tileExtent(self): return self._tileExtent @tileExtent.setter def tileExtent(self, tileExtent not None : anon_struct0): string.memcpy(&self._pvt_ptr[0].tileExtent, tileExtent.getPtr(), sizeof(self._pvt_ptr[0].tileExtent)) @property def miptailFirstLevel(self): return self._pvt_ptr[0].miptailFirstLevel @miptailFirstLevel.setter def miptailFirstLevel(self, unsigned int miptailFirstLevel): self._pvt_ptr[0].miptailFirstLevel = miptailFirstLevel @property def miptailSize(self): return self._pvt_ptr[0].miptailSize @miptailSize.setter def miptailSize(self, unsigned long long miptailSize): self._pvt_ptr[0].miptailSize = miptailSize @property def flags(self): return self._pvt_ptr[0].flags @flags.setter def flags(self, unsigned int flags): self._pvt_ptr[0].flags = flags @property def reserved(self): return self._pvt_ptr[0].reserved @reserved.setter def reserved(self, reserved): self._pvt_ptr[0].reserved = reserved cdef class cudaArrayMemoryRequirements: """ CUDA array and CUDA mipmapped array memory requirements Attributes ---------- size : size_t Total size of the array. alignment : size_t Alignment necessary for mapping the array. reserved : list[unsigned int] Methods ------- getPtr() Get memory address of class instance """ def __cinit__(self, void_ptr _ptr = 0): if _ptr == 0: self._pvt_ptr = &self._pvt_val else: self._pvt_ptr = _ptr def __init__(self, void_ptr _ptr = 0): pass def __dealloc__(self): pass def getPtr(self): return self._pvt_ptr def __repr__(self): if self._pvt_ptr is not NULL: str_list = [] try: str_list += ['size : ' + str(self.size)] except ValueError: str_list += ['size : '] try: str_list += ['alignment : ' + str(self.alignment)] except ValueError: str_list += ['alignment : '] try: str_list += ['reserved : ' + str(self.reserved)] except ValueError: str_list += ['reserved : '] return '\n'.join(str_list) else: return '' @property def size(self): return self._pvt_ptr[0].size @size.setter def size(self, size_t size): self._pvt_ptr[0].size = size @property def alignment(self): return self._pvt_ptr[0].alignment @alignment.setter def alignment(self, size_t alignment): self._pvt_ptr[0].alignment = alignment @property def reserved(self): return self._pvt_ptr[0].reserved @reserved.setter def reserved(self, reserved): self._pvt_ptr[0].reserved = reserved cdef class cudaPitchedPtr: """ CUDA Pitched memory pointer ::make_cudaPitchedPtr Attributes ---------- ptr : Any Pointer to allocated memory pitch : size_t Pitch of allocated memory in bytes xsize : size_t Logical width of allocation in elements ysize : size_t Logical height of allocation in elements Methods ------- getPtr() Get memory address of class instance """ def __cinit__(self, void_ptr _ptr = 0): if _ptr == 0: self._pvt_ptr = &self._pvt_val else: self._pvt_ptr = _ptr def __init__(self, void_ptr _ptr = 0): pass def __dealloc__(self): pass def getPtr(self): return self._pvt_ptr def __repr__(self): if self._pvt_ptr is not NULL: str_list = [] try: str_list += ['ptr : ' + hex(self.ptr)] except ValueError: str_list += ['ptr : '] try: str_list += ['pitch : ' + str(self.pitch)] except ValueError: str_list += ['pitch : '] try: str_list += ['xsize : ' + str(self.xsize)] except ValueError: str_list += ['xsize : '] try: str_list += ['ysize : ' + str(self.ysize)] except ValueError: str_list += ['ysize : '] return '\n'.join(str_list) else: return '' @property def ptr(self): return self._pvt_ptr[0].ptr @ptr.setter def ptr(self, ptr): _cptr = _HelperInputVoidPtr(ptr) self._pvt_ptr[0].ptr = _cptr.cptr @property def pitch(self): return self._pvt_ptr[0].pitch @pitch.setter def pitch(self, size_t pitch): self._pvt_ptr[0].pitch = pitch @property def xsize(self): return self._pvt_ptr[0].xsize @xsize.setter def xsize(self, size_t xsize): self._pvt_ptr[0].xsize = xsize @property def ysize(self): return self._pvt_ptr[0].ysize @ysize.setter def ysize(self, size_t ysize): self._pvt_ptr[0].ysize = ysize cdef class cudaExtent: """ CUDA extent ::make_cudaExtent Attributes ---------- width : size_t Width in elements when referring to array memory, in bytes when referring to linear memory height : size_t Height in elements depth : size_t Depth in elements Methods ------- getPtr() Get memory address of class instance """ def __cinit__(self, void_ptr _ptr = 0): if _ptr == 0: self._pvt_ptr = &self._pvt_val else: self._pvt_ptr = _ptr def __init__(self, void_ptr _ptr = 0): pass def __dealloc__(self): pass def getPtr(self): return self._pvt_ptr def __repr__(self): if self._pvt_ptr is not NULL: str_list = [] try: str_list += ['width : ' + str(self.width)] except ValueError: str_list += ['width : '] try: str_list += ['height : ' + str(self.height)] except ValueError: str_list += ['height : '] try: str_list += ['depth : ' + str(self.depth)] except ValueError: str_list += ['depth : '] return '\n'.join(str_list) else: return '' @property def width(self): return self._pvt_ptr[0].width @width.setter def width(self, size_t width): self._pvt_ptr[0].width = width @property def height(self): return self._pvt_ptr[0].height @height.setter def height(self, size_t height): self._pvt_ptr[0].height = height @property def depth(self): return self._pvt_ptr[0].depth @depth.setter def depth(self, size_t depth): self._pvt_ptr[0].depth = depth cdef class cudaPos: """ CUDA 3D position ::make_cudaPos Attributes ---------- x : size_t x y : size_t y z : size_t z Methods ------- getPtr() Get memory address of class instance """ def __cinit__(self, void_ptr _ptr = 0): if _ptr == 0: self._pvt_ptr = &self._pvt_val else: self._pvt_ptr = _ptr def __init__(self, void_ptr _ptr = 0): pass def __dealloc__(self): pass def getPtr(self): return self._pvt_ptr def __repr__(self): if self._pvt_ptr is not NULL: str_list = [] try: str_list += ['x : ' + str(self.x)] except ValueError: str_list += ['x : '] try: str_list += ['y : ' + str(self.y)] except ValueError: str_list += ['y : '] try: str_list += ['z : ' + str(self.z)] except ValueError: str_list += ['z : '] return '\n'.join(str_list) else: return '' @property def x(self): return self._pvt_ptr[0].x @x.setter def x(self, size_t x): self._pvt_ptr[0].x = x @property def y(self): return self._pvt_ptr[0].y @y.setter def y(self, size_t y): self._pvt_ptr[0].y = y @property def z(self): return self._pvt_ptr[0].z @z.setter def z(self, size_t z): self._pvt_ptr[0].z = z cdef class cudaMemcpy3DParms: """ CUDA 3D memory copying parameters Attributes ---------- srcArray : cudaArray_t Source memory address srcPos : cudaPos Source position offset srcPtr : cudaPitchedPtr Pitched source memory address dstArray : cudaArray_t Destination memory address dstPos : cudaPos Destination position offset dstPtr : cudaPitchedPtr Pitched destination memory address extent : cudaExtent Requested memory copy size kind : cudaMemcpyKind Type of transfer Methods ------- getPtr() Get memory address of class instance """ def __cinit__(self, void_ptr _ptr = 0): if _ptr == 0: self._pvt_ptr = &self._pvt_val else: self._pvt_ptr = _ptr def __init__(self, void_ptr _ptr = 0): pass self._srcArray = cudaArray_t(_ptr=&self._pvt_ptr[0].srcArray) self._srcPos = cudaPos(_ptr=&self._pvt_ptr[0].srcPos) self._srcPtr = cudaPitchedPtr(_ptr=&self._pvt_ptr[0].srcPtr) self._dstArray = cudaArray_t(_ptr=&self._pvt_ptr[0].dstArray) self._dstPos = cudaPos(_ptr=&self._pvt_ptr[0].dstPos) self._dstPtr = cudaPitchedPtr(_ptr=&self._pvt_ptr[0].dstPtr) self._extent = cudaExtent(_ptr=&self._pvt_ptr[0].extent) def __dealloc__(self): pass def getPtr(self): return self._pvt_ptr def __repr__(self): if self._pvt_ptr is not NULL: str_list = [] try: str_list += ['srcArray : ' + str(self.srcArray)] except ValueError: str_list += ['srcArray : '] try: str_list += ['srcPos :\n' + '\n'.join([' ' + line for line in str(self.srcPos).splitlines()])] except ValueError: str_list += ['srcPos : '] try: str_list += ['srcPtr :\n' + '\n'.join([' ' + line for line in str(self.srcPtr).splitlines()])] except ValueError: str_list += ['srcPtr : '] try: str_list += ['dstArray : ' + str(self.dstArray)] except ValueError: str_list += ['dstArray : '] try: str_list += ['dstPos :\n' + '\n'.join([' ' + line for line in str(self.dstPos).splitlines()])] except ValueError: str_list += ['dstPos : '] try: str_list += ['dstPtr :\n' + '\n'.join([' ' + line for line in str(self.dstPtr).splitlines()])] except ValueError: str_list += ['dstPtr : '] try: str_list += ['extent :\n' + '\n'.join([' ' + line for line in str(self.extent).splitlines()])] except ValueError: str_list += ['extent : '] try: str_list += ['kind : ' + str(self.kind)] except ValueError: str_list += ['kind : '] return '\n'.join(str_list) else: return '' @property def srcArray(self): return self._srcArray @srcArray.setter def srcArray(self, srcArray): cdef cyruntime.cudaArray_t cysrcArray if srcArray is None: cysrcArray = 0 elif isinstance(srcArray, (cudaArray_t,)): psrcArray = int(srcArray) cysrcArray = psrcArray else: psrcArray = int(cudaArray_t(srcArray)) cysrcArray = psrcArray self._srcArray._pvt_ptr[0] = cysrcArray @property def srcPos(self): return self._srcPos @srcPos.setter def srcPos(self, srcPos not None : cudaPos): string.memcpy(&self._pvt_ptr[0].srcPos, srcPos.getPtr(), sizeof(self._pvt_ptr[0].srcPos)) @property def srcPtr(self): return self._srcPtr @srcPtr.setter def srcPtr(self, srcPtr not None : cudaPitchedPtr): string.memcpy(&self._pvt_ptr[0].srcPtr, srcPtr.getPtr(), sizeof(self._pvt_ptr[0].srcPtr)) @property def dstArray(self): return self._dstArray @dstArray.setter def dstArray(self, dstArray): cdef cyruntime.cudaArray_t cydstArray if dstArray is None: cydstArray = 0 elif isinstance(dstArray, (cudaArray_t,)): pdstArray = int(dstArray) cydstArray = pdstArray else: pdstArray = int(cudaArray_t(dstArray)) cydstArray = pdstArray self._dstArray._pvt_ptr[0] = cydstArray @property def dstPos(self): return self._dstPos @dstPos.setter def dstPos(self, dstPos not None : cudaPos): string.memcpy(&self._pvt_ptr[0].dstPos, dstPos.getPtr(), sizeof(self._pvt_ptr[0].dstPos)) @property def dstPtr(self): return self._dstPtr @dstPtr.setter def dstPtr(self, dstPtr not None : cudaPitchedPtr): string.memcpy(&self._pvt_ptr[0].dstPtr, dstPtr.getPtr(), sizeof(self._pvt_ptr[0].dstPtr)) @property def extent(self): return self._extent @extent.setter def extent(self, extent not None : cudaExtent): string.memcpy(&self._pvt_ptr[0].extent, extent.getPtr(), sizeof(self._pvt_ptr[0].extent)) @property def kind(self): if self._pvt_ptr[0].kind not in _dict_cudaMemcpyKind: return None return _dict_cudaMemcpyKind[self._pvt_ptr[0].kind] @kind.setter def kind(self, kind not None : cudaMemcpyKind): self._pvt_ptr[0].kind = kind.value cdef class cudaMemcpyNodeParams: """ Memcpy node parameters Attributes ---------- flags : int Must be zero reserved : list[int] Must be zero copyParams : cudaMemcpy3DParms Parameters for the memory copy Methods ------- getPtr() Get memory address of class instance """ def __cinit__(self, void_ptr _ptr = 0): if _ptr == 0: self._pvt_ptr = &self._pvt_val else: self._pvt_ptr = _ptr def __init__(self, void_ptr _ptr = 0): pass self._copyParams = cudaMemcpy3DParms(_ptr=&self._pvt_ptr[0].copyParams) def __dealloc__(self): pass def getPtr(self): return self._pvt_ptr def __repr__(self): if self._pvt_ptr is not NULL: str_list = [] try: str_list += ['flags : ' + str(self.flags)] except ValueError: str_list += ['flags : '] try: str_list += ['reserved : ' + str(self.reserved)] except ValueError: str_list += ['reserved : '] try: str_list += ['copyParams :\n' + '\n'.join([' ' + line for line in str(self.copyParams).splitlines()])] except ValueError: str_list += ['copyParams : '] return '\n'.join(str_list) else: return '' @property def flags(self): return self._pvt_ptr[0].flags @flags.setter def flags(self, int flags): self._pvt_ptr[0].flags = flags @property def reserved(self): return self._pvt_ptr[0].reserved @reserved.setter def reserved(self, reserved): self._pvt_ptr[0].reserved = reserved @property def copyParams(self): return self._copyParams @copyParams.setter def copyParams(self, copyParams not None : cudaMemcpy3DParms): string.memcpy(&self._pvt_ptr[0].copyParams, copyParams.getPtr(), sizeof(self._pvt_ptr[0].copyParams)) cdef class cudaMemcpy3DPeerParms: """ CUDA 3D cross-device memory copying parameters Attributes ---------- srcArray : cudaArray_t Source memory address srcPos : cudaPos Source position offset srcPtr : cudaPitchedPtr Pitched source memory address srcDevice : int Source device dstArray : cudaArray_t Destination memory address dstPos : cudaPos Destination position offset dstPtr : cudaPitchedPtr Pitched destination memory address dstDevice : int Destination device extent : cudaExtent Requested memory copy size Methods ------- getPtr() Get memory address of class instance """ def __cinit__(self, void_ptr _ptr = 0): if _ptr == 0: self._pvt_ptr = &self._pvt_val else: self._pvt_ptr = _ptr def __init__(self, void_ptr _ptr = 0): pass self._srcArray = cudaArray_t(_ptr=&self._pvt_ptr[0].srcArray) self._srcPos = cudaPos(_ptr=&self._pvt_ptr[0].srcPos) self._srcPtr = cudaPitchedPtr(_ptr=&self._pvt_ptr[0].srcPtr) self._dstArray = cudaArray_t(_ptr=&self._pvt_ptr[0].dstArray) self._dstPos = cudaPos(_ptr=&self._pvt_ptr[0].dstPos) self._dstPtr = cudaPitchedPtr(_ptr=&self._pvt_ptr[0].dstPtr) self._extent = cudaExtent(_ptr=&self._pvt_ptr[0].extent) def __dealloc__(self): pass def getPtr(self): return self._pvt_ptr def __repr__(self): if self._pvt_ptr is not NULL: str_list = [] try: str_list += ['srcArray : ' + str(self.srcArray)] except ValueError: str_list += ['srcArray : '] try: str_list += ['srcPos :\n' + '\n'.join([' ' + line for line in str(self.srcPos).splitlines()])] except ValueError: str_list += ['srcPos : '] try: str_list += ['srcPtr :\n' + '\n'.join([' ' + line for line in str(self.srcPtr).splitlines()])] except ValueError: str_list += ['srcPtr : '] try: str_list += ['srcDevice : ' + str(self.srcDevice)] except ValueError: str_list += ['srcDevice : '] try: str_list += ['dstArray : ' + str(self.dstArray)] except ValueError: str_list += ['dstArray : '] try: str_list += ['dstPos :\n' + '\n'.join([' ' + line for line in str(self.dstPos).splitlines()])] except ValueError: str_list += ['dstPos : '] try: str_list += ['dstPtr :\n' + '\n'.join([' ' + line for line in str(self.dstPtr).splitlines()])] except ValueError: str_list += ['dstPtr : '] try: str_list += ['dstDevice : ' + str(self.dstDevice)] except ValueError: str_list += ['dstDevice : '] try: str_list += ['extent :\n' + '\n'.join([' ' + line for line in str(self.extent).splitlines()])] except ValueError: str_list += ['extent : '] return '\n'.join(str_list) else: return '' @property def srcArray(self): return self._srcArray @srcArray.setter def srcArray(self, srcArray): cdef cyruntime.cudaArray_t cysrcArray if srcArray is None: cysrcArray = 0 elif isinstance(srcArray, (cudaArray_t,)): psrcArray = int(srcArray) cysrcArray = psrcArray else: psrcArray = int(cudaArray_t(srcArray)) cysrcArray = psrcArray self._srcArray._pvt_ptr[0] = cysrcArray @property def srcPos(self): return self._srcPos @srcPos.setter def srcPos(self, srcPos not None : cudaPos): string.memcpy(&self._pvt_ptr[0].srcPos, srcPos.getPtr(), sizeof(self._pvt_ptr[0].srcPos)) @property def srcPtr(self): return self._srcPtr @srcPtr.setter def srcPtr(self, srcPtr not None : cudaPitchedPtr): string.memcpy(&self._pvt_ptr[0].srcPtr, srcPtr.getPtr(), sizeof(self._pvt_ptr[0].srcPtr)) @property def srcDevice(self): return self._pvt_ptr[0].srcDevice @srcDevice.setter def srcDevice(self, int srcDevice): self._pvt_ptr[0].srcDevice = srcDevice @property def dstArray(self): return self._dstArray @dstArray.setter def dstArray(self, dstArray): cdef cyruntime.cudaArray_t cydstArray if dstArray is None: cydstArray = 0 elif isinstance(dstArray, (cudaArray_t,)): pdstArray = int(dstArray) cydstArray = pdstArray else: pdstArray = int(cudaArray_t(dstArray)) cydstArray = pdstArray self._dstArray._pvt_ptr[0] = cydstArray @property def dstPos(self): return self._dstPos @dstPos.setter def dstPos(self, dstPos not None : cudaPos): string.memcpy(&self._pvt_ptr[0].dstPos, dstPos.getPtr(), sizeof(self._pvt_ptr[0].dstPos)) @property def dstPtr(self): return self._dstPtr @dstPtr.setter def dstPtr(self, dstPtr not None : cudaPitchedPtr): string.memcpy(&self._pvt_ptr[0].dstPtr, dstPtr.getPtr(), sizeof(self._pvt_ptr[0].dstPtr)) @property def dstDevice(self): return self._pvt_ptr[0].dstDevice @dstDevice.setter def dstDevice(self, int dstDevice): self._pvt_ptr[0].dstDevice = dstDevice @property def extent(self): return self._extent @extent.setter def extent(self, extent not None : cudaExtent): string.memcpy(&self._pvt_ptr[0].extent, extent.getPtr(), sizeof(self._pvt_ptr[0].extent)) cdef class cudaMemsetParams: """ CUDA Memset node parameters Attributes ---------- dst : Any Destination device pointer pitch : size_t Pitch of destination device pointer. Unused if height is 1 value : unsigned int Value to be set elementSize : unsigned int Size of each element in bytes. Must be 1, 2, or 4. width : size_t Width of the row in elements height : size_t Number of rows Methods ------- getPtr() Get memory address of class instance """ def __cinit__(self, void_ptr _ptr = 0): if _ptr == 0: self._pvt_ptr = &self._pvt_val else: self._pvt_ptr = _ptr def __init__(self, void_ptr _ptr = 0): pass def __dealloc__(self): pass def getPtr(self): return self._pvt_ptr def __repr__(self): if self._pvt_ptr is not NULL: str_list = [] try: str_list += ['dst : ' + hex(self.dst)] except ValueError: str_list += ['dst : '] try: str_list += ['pitch : ' + str(self.pitch)] except ValueError: str_list += ['pitch : '] try: str_list += ['value : ' + str(self.value)] except ValueError: str_list += ['value : '] try: str_list += ['elementSize : ' + str(self.elementSize)] except ValueError: str_list += ['elementSize : '] try: str_list += ['width : ' + str(self.width)] except ValueError: str_list += ['width : '] try: str_list += ['height : ' + str(self.height)] except ValueError: str_list += ['height : '] return '\n'.join(str_list) else: return '' @property def dst(self): return self._pvt_ptr[0].dst @dst.setter def dst(self, dst): _cdst = _HelperInputVoidPtr(dst) self._pvt_ptr[0].dst = _cdst.cptr @property def pitch(self): return self._pvt_ptr[0].pitch @pitch.setter def pitch(self, size_t pitch): self._pvt_ptr[0].pitch = pitch @property def value(self): return self._pvt_ptr[0].value @value.setter def value(self, unsigned int value): self._pvt_ptr[0].value = value @property def elementSize(self): return self._pvt_ptr[0].elementSize @elementSize.setter def elementSize(self, unsigned int elementSize): self._pvt_ptr[0].elementSize = elementSize @property def width(self): return self._pvt_ptr[0].width @width.setter def width(self, size_t width): self._pvt_ptr[0].width = width @property def height(self): return self._pvt_ptr[0].height @height.setter def height(self, size_t height): self._pvt_ptr[0].height = height cdef class cudaMemsetParamsV2: """ CUDA Memset node parameters Attributes ---------- dst : Any Destination device pointer pitch : size_t Pitch of destination device pointer. Unused if height is 1 value : unsigned int Value to be set elementSize : unsigned int Size of each element in bytes. Must be 1, 2, or 4. width : size_t Width of the row in elements height : size_t Number of rows Methods ------- getPtr() Get memory address of class instance """ def __cinit__(self, void_ptr _ptr = 0): if _ptr == 0: self._pvt_ptr = &self._pvt_val else: self._pvt_ptr = _ptr def __init__(self, void_ptr _ptr = 0): pass def __dealloc__(self): pass def getPtr(self): return self._pvt_ptr def __repr__(self): if self._pvt_ptr is not NULL: str_list = [] try: str_list += ['dst : ' + hex(self.dst)] except ValueError: str_list += ['dst : '] try: str_list += ['pitch : ' + str(self.pitch)] except ValueError: str_list += ['pitch : '] try: str_list += ['value : ' + str(self.value)] except ValueError: str_list += ['value : '] try: str_list += ['elementSize : ' + str(self.elementSize)] except ValueError: str_list += ['elementSize : '] try: str_list += ['width : ' + str(self.width)] except ValueError: str_list += ['width : '] try: str_list += ['height : ' + str(self.height)] except ValueError: str_list += ['height : '] return '\n'.join(str_list) else: return '' @property def dst(self): return self._pvt_ptr[0].dst @dst.setter def dst(self, dst): _cdst = _HelperInputVoidPtr(dst) self._pvt_ptr[0].dst = _cdst.cptr @property def pitch(self): return self._pvt_ptr[0].pitch @pitch.setter def pitch(self, size_t pitch): self._pvt_ptr[0].pitch = pitch @property def value(self): return self._pvt_ptr[0].value @value.setter def value(self, unsigned int value): self._pvt_ptr[0].value = value @property def elementSize(self): return self._pvt_ptr[0].elementSize @elementSize.setter def elementSize(self, unsigned int elementSize): self._pvt_ptr[0].elementSize = elementSize @property def width(self): return self._pvt_ptr[0].width @width.setter def width(self, size_t width): self._pvt_ptr[0].width = width @property def height(self): return self._pvt_ptr[0].height @height.setter def height(self, size_t height): self._pvt_ptr[0].height = height cdef class cudaAccessPolicyWindow: """ Specifies an access policy for a window, a contiguous extent of memory beginning at base_ptr and ending at base_ptr + num_bytes. Partition into many segments and assign segments such that. sum of "hit segments" / window == approx. ratio. sum of "miss segments" / window == approx 1-ratio. Segments and ratio specifications are fitted to the capabilities of the architecture. Accesses in a hit segment apply the hitProp access policy. Accesses in a miss segment apply the missProp access policy. Attributes ---------- base_ptr : Any Starting address of the access policy window. CUDA driver may align it. num_bytes : size_t Size in bytes of the window policy. CUDA driver may restrict the maximum size and alignment. hitRatio : float hitRatio specifies percentage of lines assigned hitProp, rest are assigned missProp. hitProp : cudaAccessProperty ::CUaccessProperty set for hit. missProp : cudaAccessProperty ::CUaccessProperty set for miss. Must be either NORMAL or STREAMING. Methods ------- getPtr() Get memory address of class instance """ def __cinit__(self, void_ptr _ptr = 0): if _ptr == 0: self._pvt_ptr = &self._pvt_val else: self._pvt_ptr = _ptr def __init__(self, void_ptr _ptr = 0): pass def __dealloc__(self): pass def getPtr(self): return self._pvt_ptr def __repr__(self): if self._pvt_ptr is not NULL: str_list = [] try: str_list += ['base_ptr : ' + hex(self.base_ptr)] except ValueError: str_list += ['base_ptr : '] try: str_list += ['num_bytes : ' + str(self.num_bytes)] except ValueError: str_list += ['num_bytes : '] try: str_list += ['hitRatio : ' + str(self.hitRatio)] except ValueError: str_list += ['hitRatio : '] try: str_list += ['hitProp : ' + str(self.hitProp)] except ValueError: str_list += ['hitProp : '] try: str_list += ['missProp : ' + str(self.missProp)] except ValueError: str_list += ['missProp : '] return '\n'.join(str_list) else: return '' @property def base_ptr(self): return self._pvt_ptr[0].base_ptr @base_ptr.setter def base_ptr(self, base_ptr): _cbase_ptr = _HelperInputVoidPtr(base_ptr) self._pvt_ptr[0].base_ptr = _cbase_ptr.cptr @property def num_bytes(self): return self._pvt_ptr[0].num_bytes @num_bytes.setter def num_bytes(self, size_t num_bytes): self._pvt_ptr[0].num_bytes = num_bytes @property def hitRatio(self): return self._pvt_ptr[0].hitRatio @hitRatio.setter def hitRatio(self, float hitRatio): self._pvt_ptr[0].hitRatio = hitRatio @property def hitProp(self): if self._pvt_ptr[0].hitProp not in _dict_cudaAccessProperty: return None return _dict_cudaAccessProperty[self._pvt_ptr[0].hitProp] @hitProp.setter def hitProp(self, hitProp not None : cudaAccessProperty): self._pvt_ptr[0].hitProp = hitProp.value @property def missProp(self): if self._pvt_ptr[0].missProp not in _dict_cudaAccessProperty: return None return _dict_cudaAccessProperty[self._pvt_ptr[0].missProp] @missProp.setter def missProp(self, missProp not None : cudaAccessProperty): self._pvt_ptr[0].missProp = missProp.value cdef class cudaHostNodeParams: """ CUDA host node parameters Attributes ---------- fn : cudaHostFn_t The function to call when the node executes userData : Any Argument to pass to the function Methods ------- getPtr() Get memory address of class instance """ def __cinit__(self, void_ptr _ptr = 0): if _ptr == 0: self._pvt_ptr = &self._pvt_val else: self._pvt_ptr = _ptr def __init__(self, void_ptr _ptr = 0): pass self._fn = cudaHostFn_t(_ptr=&self._pvt_ptr[0].fn) def __dealloc__(self): pass def getPtr(self): return self._pvt_ptr def __repr__(self): if self._pvt_ptr is not NULL: str_list = [] try: str_list += ['fn : ' + str(self.fn)] except ValueError: str_list += ['fn : '] try: str_list += ['userData : ' + hex(self.userData)] except ValueError: str_list += ['userData : '] return '\n'.join(str_list) else: return '' @property def fn(self): return self._fn @fn.setter def fn(self, fn): cdef cyruntime.cudaHostFn_t cyfn if fn is None: cyfn = 0 elif isinstance(fn, (cudaHostFn_t)): pfn = int(fn) cyfn = pfn else: pfn = int(cudaHostFn_t(fn)) cyfn = pfn self._fn._pvt_ptr[0] = cyfn @property def userData(self): return self._pvt_ptr[0].userData @userData.setter def userData(self, userData): _cuserData = _HelperInputVoidPtr(userData) self._pvt_ptr[0].userData = _cuserData.cptr cdef class cudaHostNodeParamsV2: """ CUDA host node parameters Attributes ---------- fn : cudaHostFn_t The function to call when the node executes userData : Any Argument to pass to the function Methods ------- getPtr() Get memory address of class instance """ def __cinit__(self, void_ptr _ptr = 0): if _ptr == 0: self._pvt_ptr = &self._pvt_val else: self._pvt_ptr = _ptr def __init__(self, void_ptr _ptr = 0): pass self._fn = cudaHostFn_t(_ptr=&self._pvt_ptr[0].fn) def __dealloc__(self): pass def getPtr(self): return self._pvt_ptr def __repr__(self): if self._pvt_ptr is not NULL: str_list = [] try: str_list += ['fn : ' + str(self.fn)] except ValueError: str_list += ['fn : '] try: str_list += ['userData : ' + hex(self.userData)] except ValueError: str_list += ['userData : '] return '\n'.join(str_list) else: return '' @property def fn(self): return self._fn @fn.setter def fn(self, fn): cdef cyruntime.cudaHostFn_t cyfn if fn is None: cyfn = 0 elif isinstance(fn, (cudaHostFn_t)): pfn = int(fn) cyfn = pfn else: pfn = int(cudaHostFn_t(fn)) cyfn = pfn self._fn._pvt_ptr[0] = cyfn @property def userData(self): return self._pvt_ptr[0].userData @userData.setter def userData(self, userData): _cuserData = _HelperInputVoidPtr(userData) self._pvt_ptr[0].userData = _cuserData.cptr cdef class anon_struct1: """ Attributes ---------- array : cudaArray_t Methods ------- getPtr() Get memory address of class instance """ def __cinit__(self, void_ptr _ptr): self._pvt_ptr = _ptr def __init__(self, void_ptr _ptr): pass self._array = cudaArray_t(_ptr=&self._pvt_ptr[0].res.array.array) def __dealloc__(self): pass def getPtr(self): return &self._pvt_ptr[0].res.array def __repr__(self): if self._pvt_ptr is not NULL: str_list = [] try: str_list += ['array : ' + str(self.array)] except ValueError: str_list += ['array : '] return '\n'.join(str_list) else: return '' @property def array(self): return self._array @array.setter def array(self, array): cdef cyruntime.cudaArray_t cyarray if array is None: cyarray = 0 elif isinstance(array, (cudaArray_t,)): parray = int(array) cyarray = parray else: parray = int(cudaArray_t(array)) cyarray = parray self._array._pvt_ptr[0] = cyarray cdef class anon_struct2: """ Attributes ---------- mipmap : cudaMipmappedArray_t Methods ------- getPtr() Get memory address of class instance """ def __cinit__(self, void_ptr _ptr): self._pvt_ptr = _ptr def __init__(self, void_ptr _ptr): pass self._mipmap = cudaMipmappedArray_t(_ptr=&self._pvt_ptr[0].res.mipmap.mipmap) def __dealloc__(self): pass def getPtr(self): return &self._pvt_ptr[0].res.mipmap def __repr__(self): if self._pvt_ptr is not NULL: str_list = [] try: str_list += ['mipmap : ' + str(self.mipmap)] except ValueError: str_list += ['mipmap : '] return '\n'.join(str_list) else: return '' @property def mipmap(self): return self._mipmap @mipmap.setter def mipmap(self, mipmap): cdef cyruntime.cudaMipmappedArray_t cymipmap if mipmap is None: cymipmap = 0 elif isinstance(mipmap, (cudaMipmappedArray_t,)): pmipmap = int(mipmap) cymipmap = pmipmap else: pmipmap = int(cudaMipmappedArray_t(mipmap)) cymipmap = pmipmap self._mipmap._pvt_ptr[0] = cymipmap cdef class anon_struct3: """ Attributes ---------- devPtr : Any desc : cudaChannelFormatDesc sizeInBytes : size_t Methods ------- getPtr() Get memory address of class instance """ def __cinit__(self, void_ptr _ptr): self._pvt_ptr = _ptr def __init__(self, void_ptr _ptr): pass self._desc = cudaChannelFormatDesc(_ptr=&self._pvt_ptr[0].res.linear.desc) def __dealloc__(self): pass def getPtr(self): return &self._pvt_ptr[0].res.linear def __repr__(self): if self._pvt_ptr is not NULL: str_list = [] try: str_list += ['devPtr : ' + hex(self.devPtr)] except ValueError: str_list += ['devPtr : '] try: str_list += ['desc :\n' + '\n'.join([' ' + line for line in str(self.desc).splitlines()])] except ValueError: str_list += ['desc : '] try: str_list += ['sizeInBytes : ' + str(self.sizeInBytes)] except ValueError: str_list += ['sizeInBytes : '] return '\n'.join(str_list) else: return '' @property def devPtr(self): return self._pvt_ptr[0].res.linear.devPtr @devPtr.setter def devPtr(self, devPtr): _cdevPtr = _HelperInputVoidPtr(devPtr) self._pvt_ptr[0].res.linear.devPtr = _cdevPtr.cptr @property def desc(self): return self._desc @desc.setter def desc(self, desc not None : cudaChannelFormatDesc): string.memcpy(&self._pvt_ptr[0].res.linear.desc, desc.getPtr(), sizeof(self._pvt_ptr[0].res.linear.desc)) @property def sizeInBytes(self): return self._pvt_ptr[0].res.linear.sizeInBytes @sizeInBytes.setter def sizeInBytes(self, size_t sizeInBytes): self._pvt_ptr[0].res.linear.sizeInBytes = sizeInBytes cdef class anon_struct4: """ Attributes ---------- devPtr : Any desc : cudaChannelFormatDesc width : size_t height : size_t pitchInBytes : size_t Methods ------- getPtr() Get memory address of class instance """ def __cinit__(self, void_ptr _ptr): self._pvt_ptr = _ptr def __init__(self, void_ptr _ptr): pass self._desc = cudaChannelFormatDesc(_ptr=&self._pvt_ptr[0].res.pitch2D.desc) def __dealloc__(self): pass def getPtr(self): return &self._pvt_ptr[0].res.pitch2D def __repr__(self): if self._pvt_ptr is not NULL: str_list = [] try: str_list += ['devPtr : ' + hex(self.devPtr)] except ValueError: str_list += ['devPtr : '] try: str_list += ['desc :\n' + '\n'.join([' ' + line for line in str(self.desc).splitlines()])] except ValueError: str_list += ['desc : '] try: str_list += ['width : ' + str(self.width)] except ValueError: str_list += ['width : '] try: str_list += ['height : ' + str(self.height)] except ValueError: str_list += ['height : '] try: str_list += ['pitchInBytes : ' + str(self.pitchInBytes)] except ValueError: str_list += ['pitchInBytes : '] return '\n'.join(str_list) else: return '' @property def devPtr(self): return self._pvt_ptr[0].res.pitch2D.devPtr @devPtr.setter def devPtr(self, devPtr): _cdevPtr = _HelperInputVoidPtr(devPtr) self._pvt_ptr[0].res.pitch2D.devPtr = _cdevPtr.cptr @property def desc(self): return self._desc @desc.setter def desc(self, desc not None : cudaChannelFormatDesc): string.memcpy(&self._pvt_ptr[0].res.pitch2D.desc, desc.getPtr(), sizeof(self._pvt_ptr[0].res.pitch2D.desc)) @property def width(self): return self._pvt_ptr[0].res.pitch2D.width @width.setter def width(self, size_t width): self._pvt_ptr[0].res.pitch2D.width = width @property def height(self): return self._pvt_ptr[0].res.pitch2D.height @height.setter def height(self, size_t height): self._pvt_ptr[0].res.pitch2D.height = height @property def pitchInBytes(self): return self._pvt_ptr[0].res.pitch2D.pitchInBytes @pitchInBytes.setter def pitchInBytes(self, size_t pitchInBytes): self._pvt_ptr[0].res.pitch2D.pitchInBytes = pitchInBytes cdef class anon_union0: """ Attributes ---------- array : anon_struct1 mipmap : anon_struct2 linear : anon_struct3 pitch2D : anon_struct4 Methods ------- getPtr() Get memory address of class instance """ def __cinit__(self, void_ptr _ptr): self._pvt_ptr = _ptr def __init__(self, void_ptr _ptr): pass self._array = anon_struct1(_ptr=self._pvt_ptr) self._mipmap = anon_struct2(_ptr=self._pvt_ptr) self._linear = anon_struct3(_ptr=self._pvt_ptr) self._pitch2D = anon_struct4(_ptr=self._pvt_ptr) def __dealloc__(self): pass def getPtr(self): return &self._pvt_ptr[0].res def __repr__(self): if self._pvt_ptr is not NULL: str_list = [] try: str_list += ['array :\n' + '\n'.join([' ' + line for line in str(self.array).splitlines()])] except ValueError: str_list += ['array : '] try: str_list += ['mipmap :\n' + '\n'.join([' ' + line for line in str(self.mipmap).splitlines()])] except ValueError: str_list += ['mipmap : '] try: str_list += ['linear :\n' + '\n'.join([' ' + line for line in str(self.linear).splitlines()])] except ValueError: str_list += ['linear : '] try: str_list += ['pitch2D :\n' + '\n'.join([' ' + line for line in str(self.pitch2D).splitlines()])] except ValueError: str_list += ['pitch2D : '] return '\n'.join(str_list) else: return '' @property def array(self): return self._array @array.setter def array(self, array not None : anon_struct1): string.memcpy(&self._pvt_ptr[0].res.array, array.getPtr(), sizeof(self._pvt_ptr[0].res.array)) @property def mipmap(self): return self._mipmap @mipmap.setter def mipmap(self, mipmap not None : anon_struct2): string.memcpy(&self._pvt_ptr[0].res.mipmap, mipmap.getPtr(), sizeof(self._pvt_ptr[0].res.mipmap)) @property def linear(self): return self._linear @linear.setter def linear(self, linear not None : anon_struct3): string.memcpy(&self._pvt_ptr[0].res.linear, linear.getPtr(), sizeof(self._pvt_ptr[0].res.linear)) @property def pitch2D(self): return self._pitch2D @pitch2D.setter def pitch2D(self, pitch2D not None : anon_struct4): string.memcpy(&self._pvt_ptr[0].res.pitch2D, pitch2D.getPtr(), sizeof(self._pvt_ptr[0].res.pitch2D)) cdef class cudaResourceDesc: """ CUDA resource descriptor Attributes ---------- resType : cudaResourceType Resource type res : anon_union0 Methods ------- getPtr() Get memory address of class instance """ def __cinit__(self, void_ptr _ptr = 0): if _ptr == 0: self._val_ptr = calloc(1, sizeof(cyruntime.cudaResourceDesc)) self._pvt_ptr = self._val_ptr else: self._pvt_ptr = _ptr def __init__(self, void_ptr _ptr = 0): pass self._res = anon_union0(_ptr=self._pvt_ptr) def __dealloc__(self): if self._val_ptr is not NULL: free(self._val_ptr) def getPtr(self): return self._pvt_ptr def __repr__(self): if self._pvt_ptr is not NULL: str_list = [] try: str_list += ['resType : ' + str(self.resType)] except ValueError: str_list += ['resType : '] try: str_list += ['res :\n' + '\n'.join([' ' + line for line in str(self.res).splitlines()])] except ValueError: str_list += ['res : '] return '\n'.join(str_list) else: return '' @property def resType(self): if self._pvt_ptr[0].resType not in _dict_cudaResourceType: return None return _dict_cudaResourceType[self._pvt_ptr[0].resType] @resType.setter def resType(self, resType not None : cudaResourceType): self._pvt_ptr[0].resType = resType.value @property def res(self): return self._res @res.setter def res(self, res not None : anon_union0): string.memcpy(&self._pvt_ptr[0].res, res.getPtr(), sizeof(self._pvt_ptr[0].res)) cdef class cudaResourceViewDesc: """ CUDA resource view descriptor Attributes ---------- format : cudaResourceViewFormat Resource view format width : size_t Width of the resource view height : size_t Height of the resource view depth : size_t Depth of the resource view firstMipmapLevel : unsigned int First defined mipmap level lastMipmapLevel : unsigned int Last defined mipmap level firstLayer : unsigned int First layer index lastLayer : unsigned int Last layer index Methods ------- getPtr() Get memory address of class instance """ def __cinit__(self, void_ptr _ptr = 0): if _ptr == 0: self._pvt_ptr = &self._pvt_val else: self._pvt_ptr = _ptr def __init__(self, void_ptr _ptr = 0): pass def __dealloc__(self): pass def getPtr(self): return self._pvt_ptr def __repr__(self): if self._pvt_ptr is not NULL: str_list = [] try: str_list += ['format : ' + str(self.format)] except ValueError: str_list += ['format : '] try: str_list += ['width : ' + str(self.width)] except ValueError: str_list += ['width : '] try: str_list += ['height : ' + str(self.height)] except ValueError: str_list += ['height : '] try: str_list += ['depth : ' + str(self.depth)] except ValueError: str_list += ['depth : '] try: str_list += ['firstMipmapLevel : ' + str(self.firstMipmapLevel)] except ValueError: str_list += ['firstMipmapLevel : '] try: str_list += ['lastMipmapLevel : ' + str(self.lastMipmapLevel)] except ValueError: str_list += ['lastMipmapLevel : '] try: str_list += ['firstLayer : ' + str(self.firstLayer)] except ValueError: str_list += ['firstLayer : '] try: str_list += ['lastLayer : ' + str(self.lastLayer)] except ValueError: str_list += ['lastLayer : '] return '\n'.join(str_list) else: return '' @property def format(self): if self._pvt_ptr[0].format not in _dict_cudaResourceViewFormat: return None return _dict_cudaResourceViewFormat[self._pvt_ptr[0].format] @format.setter def format(self, format not None : cudaResourceViewFormat): self._pvt_ptr[0].format = format.value @property def width(self): return self._pvt_ptr[0].width @width.setter def width(self, size_t width): self._pvt_ptr[0].width = width @property def height(self): return self._pvt_ptr[0].height @height.setter def height(self, size_t height): self._pvt_ptr[0].height = height @property def depth(self): return self._pvt_ptr[0].depth @depth.setter def depth(self, size_t depth): self._pvt_ptr[0].depth = depth @property def firstMipmapLevel(self): return self._pvt_ptr[0].firstMipmapLevel @firstMipmapLevel.setter def firstMipmapLevel(self, unsigned int firstMipmapLevel): self._pvt_ptr[0].firstMipmapLevel = firstMipmapLevel @property def lastMipmapLevel(self): return self._pvt_ptr[0].lastMipmapLevel @lastMipmapLevel.setter def lastMipmapLevel(self, unsigned int lastMipmapLevel): self._pvt_ptr[0].lastMipmapLevel = lastMipmapLevel @property def firstLayer(self): return self._pvt_ptr[0].firstLayer @firstLayer.setter def firstLayer(self, unsigned int firstLayer): self._pvt_ptr[0].firstLayer = firstLayer @property def lastLayer(self): return self._pvt_ptr[0].lastLayer @lastLayer.setter def lastLayer(self, unsigned int lastLayer): self._pvt_ptr[0].lastLayer = lastLayer cdef class cudaPointerAttributes: """ CUDA pointer attributes Attributes ---------- type : cudaMemoryType The type of memory - cudaMemoryTypeUnregistered, cudaMemoryTypeHost, cudaMemoryTypeDevice or cudaMemoryTypeManaged. device : int The device against which the memory was allocated or registered. If the memory type is cudaMemoryTypeDevice then this identifies the device on which the memory referred physically resides. If the memory type is cudaMemoryTypeHost or::cudaMemoryTypeManaged then this identifies the device which was current when the memory was allocated or registered (and if that device is deinitialized then this allocation will vanish with that device's state). devicePointer : Any The address which may be dereferenced on the current device to access the memory or NULL if no such address exists. hostPointer : Any The address which may be dereferenced on the host to access the memory or NULL if no such address exists. CUDA doesn't check if unregistered memory is allocated so this field may contain invalid pointer if an invalid pointer has been passed to CUDA. Methods ------- getPtr() Get memory address of class instance """ def __cinit__(self, void_ptr _ptr = 0): if _ptr == 0: self._pvt_ptr = &self._pvt_val else: self._pvt_ptr = _ptr def __init__(self, void_ptr _ptr = 0): pass def __dealloc__(self): pass def getPtr(self): return self._pvt_ptr def __repr__(self): if self._pvt_ptr is not NULL: str_list = [] try: str_list += ['type : ' + str(self.type)] except ValueError: str_list += ['type : '] try: str_list += ['device : ' + str(self.device)] except ValueError: str_list += ['device : '] try: str_list += ['devicePointer : ' + hex(self.devicePointer)] except ValueError: str_list += ['devicePointer : '] try: str_list += ['hostPointer : ' + hex(self.hostPointer)] except ValueError: str_list += ['hostPointer : '] return '\n'.join(str_list) else: return '' @property def type(self): if self._pvt_ptr[0].type not in _dict_cudaMemoryType: return None return _dict_cudaMemoryType[self._pvt_ptr[0].type] @type.setter def type(self, type not None : cudaMemoryType): self._pvt_ptr[0].type = type.value @property def device(self): return self._pvt_ptr[0].device @device.setter def device(self, int device): self._pvt_ptr[0].device = device @property def devicePointer(self): return self._pvt_ptr[0].devicePointer @devicePointer.setter def devicePointer(self, devicePointer): _cdevicePointer = _HelperInputVoidPtr(devicePointer) self._pvt_ptr[0].devicePointer = _cdevicePointer.cptr @property def hostPointer(self): return self._pvt_ptr[0].hostPointer @hostPointer.setter def hostPointer(self, hostPointer): _chostPointer = _HelperInputVoidPtr(hostPointer) self._pvt_ptr[0].hostPointer = _chostPointer.cptr cdef class cudaFuncAttributes: """ CUDA function attributes Attributes ---------- sharedSizeBytes : size_t The size in bytes of statically-allocated shared memory per block required by this function. This does not include dynamically- allocated shared memory requested by the user at runtime. constSizeBytes : size_t The size in bytes of user-allocated constant memory required by this function. localSizeBytes : size_t The size in bytes of local memory used by each thread of this function. maxThreadsPerBlock : int The maximum number of threads per block, beyond which a launch of the function would fail. This number depends on both the function and the device on which the function is currently loaded. numRegs : int The number of registers used by each thread of this function. ptxVersion : int The PTX virtual architecture version for which the function was compiled. This value is the major PTX version * 10 + the minor PTX version, so a PTX version 1.3 function would return the value 13. binaryVersion : int The binary architecture version for which the function was compiled. This value is the major binary version * 10 + the minor binary version, so a binary version 1.3 function would return the value 13. cacheModeCA : int The attribute to indicate whether the function has been compiled with user specified option "-Xptxas --dlcm=ca" set. maxDynamicSharedSizeBytes : int The maximum size in bytes of dynamic shared memory per block for this function. Any launch must have a dynamic shared memory size smaller than this value. preferredShmemCarveout : int On devices where the L1 cache and shared memory use the same hardware resources, this sets the shared memory carveout preference, in percent of the maximum shared memory. Refer to cudaDevAttrMaxSharedMemoryPerMultiprocessor. This is only a hint, and the driver can choose a different ratio if required to execute the function. See cudaFuncSetAttribute clusterDimMustBeSet : int If this attribute is set, the kernel must launch with a valid cluster dimension specified. requiredClusterWidth : int The required cluster width/height/depth in blocks. The values must either all be 0 or all be positive. The validity of the cluster dimensions is otherwise checked at launch time. If the value is set during compile time, it cannot be set at runtime. Setting it at runtime should return cudaErrorNotPermitted. See cudaFuncSetAttribute requiredClusterHeight : int requiredClusterDepth : int clusterSchedulingPolicyPreference : int The block scheduling policy of a function. See cudaFuncSetAttribute nonPortableClusterSizeAllowed : int Whether the function can be launched with non-portable cluster size. 1 is allowed, 0 is disallowed. A non-portable cluster size may only function on the specific SKUs the program is tested on. The launch might fail if the program is run on a different hardware platform. CUDA API provides cudaOccupancyMaxActiveClusters to assist with checking whether the desired size can be launched on the current device. Portable Cluster Size A portable cluster size is guaranteed to be functional on all compute capabilities higher than the target compute capability. The portable cluster size for sm_90 is 8 blocks per cluster. This value may increase for future compute capabilities. The specific hardware unit may support higher cluster sizes that’s not guaranteed to be portable. See cudaFuncSetAttribute reserved : list[int] Reserved for future use. Methods ------- getPtr() Get memory address of class instance """ def __cinit__(self, void_ptr _ptr = 0): if _ptr == 0: self._pvt_ptr = &self._pvt_val else: self._pvt_ptr = _ptr def __init__(self, void_ptr _ptr = 0): pass def __dealloc__(self): pass def getPtr(self): return self._pvt_ptr def __repr__(self): if self._pvt_ptr is not NULL: str_list = [] try: str_list += ['sharedSizeBytes : ' + str(self.sharedSizeBytes)] except ValueError: str_list += ['sharedSizeBytes : '] try: str_list += ['constSizeBytes : ' + str(self.constSizeBytes)] except ValueError: str_list += ['constSizeBytes : '] try: str_list += ['localSizeBytes : ' + str(self.localSizeBytes)] except ValueError: str_list += ['localSizeBytes : '] try: str_list += ['maxThreadsPerBlock : ' + str(self.maxThreadsPerBlock)] except ValueError: str_list += ['maxThreadsPerBlock : '] try: str_list += ['numRegs : ' + str(self.numRegs)] except ValueError: str_list += ['numRegs : '] try: str_list += ['ptxVersion : ' + str(self.ptxVersion)] except ValueError: str_list += ['ptxVersion : '] try: str_list += ['binaryVersion : ' + str(self.binaryVersion)] except ValueError: str_list += ['binaryVersion : '] try: str_list += ['cacheModeCA : ' + str(self.cacheModeCA)] except ValueError: str_list += ['cacheModeCA : '] try: str_list += ['maxDynamicSharedSizeBytes : ' + str(self.maxDynamicSharedSizeBytes)] except ValueError: str_list += ['maxDynamicSharedSizeBytes : '] try: str_list += ['preferredShmemCarveout : ' + str(self.preferredShmemCarveout)] except ValueError: str_list += ['preferredShmemCarveout : '] try: str_list += ['clusterDimMustBeSet : ' + str(self.clusterDimMustBeSet)] except ValueError: str_list += ['clusterDimMustBeSet : '] try: str_list += ['requiredClusterWidth : ' + str(self.requiredClusterWidth)] except ValueError: str_list += ['requiredClusterWidth : '] try: str_list += ['requiredClusterHeight : ' + str(self.requiredClusterHeight)] except ValueError: str_list += ['requiredClusterHeight : '] try: str_list += ['requiredClusterDepth : ' + str(self.requiredClusterDepth)] except ValueError: str_list += ['requiredClusterDepth : '] try: str_list += ['clusterSchedulingPolicyPreference : ' + str(self.clusterSchedulingPolicyPreference)] except ValueError: str_list += ['clusterSchedulingPolicyPreference : '] try: str_list += ['nonPortableClusterSizeAllowed : ' + str(self.nonPortableClusterSizeAllowed)] except ValueError: str_list += ['nonPortableClusterSizeAllowed : '] try: str_list += ['reserved : ' + str(self.reserved)] except ValueError: str_list += ['reserved : '] return '\n'.join(str_list) else: return '' @property def sharedSizeBytes(self): return self._pvt_ptr[0].sharedSizeBytes @sharedSizeBytes.setter def sharedSizeBytes(self, size_t sharedSizeBytes): self._pvt_ptr[0].sharedSizeBytes = sharedSizeBytes @property def constSizeBytes(self): return self._pvt_ptr[0].constSizeBytes @constSizeBytes.setter def constSizeBytes(self, size_t constSizeBytes): self._pvt_ptr[0].constSizeBytes = constSizeBytes @property def localSizeBytes(self): return self._pvt_ptr[0].localSizeBytes @localSizeBytes.setter def localSizeBytes(self, size_t localSizeBytes): self._pvt_ptr[0].localSizeBytes = localSizeBytes @property def maxThreadsPerBlock(self): return self._pvt_ptr[0].maxThreadsPerBlock @maxThreadsPerBlock.setter def maxThreadsPerBlock(self, int maxThreadsPerBlock): self._pvt_ptr[0].maxThreadsPerBlock = maxThreadsPerBlock @property def numRegs(self): return self._pvt_ptr[0].numRegs @numRegs.setter def numRegs(self, int numRegs): self._pvt_ptr[0].numRegs = numRegs @property def ptxVersion(self): return self._pvt_ptr[0].ptxVersion @ptxVersion.setter def ptxVersion(self, int ptxVersion): self._pvt_ptr[0].ptxVersion = ptxVersion @property def binaryVersion(self): return self._pvt_ptr[0].binaryVersion @binaryVersion.setter def binaryVersion(self, int binaryVersion): self._pvt_ptr[0].binaryVersion = binaryVersion @property def cacheModeCA(self): return self._pvt_ptr[0].cacheModeCA @cacheModeCA.setter def cacheModeCA(self, int cacheModeCA): self._pvt_ptr[0].cacheModeCA = cacheModeCA @property def maxDynamicSharedSizeBytes(self): return self._pvt_ptr[0].maxDynamicSharedSizeBytes @maxDynamicSharedSizeBytes.setter def maxDynamicSharedSizeBytes(self, int maxDynamicSharedSizeBytes): self._pvt_ptr[0].maxDynamicSharedSizeBytes = maxDynamicSharedSizeBytes @property def preferredShmemCarveout(self): return self._pvt_ptr[0].preferredShmemCarveout @preferredShmemCarveout.setter def preferredShmemCarveout(self, int preferredShmemCarveout): self._pvt_ptr[0].preferredShmemCarveout = preferredShmemCarveout @property def clusterDimMustBeSet(self): return self._pvt_ptr[0].clusterDimMustBeSet @clusterDimMustBeSet.setter def clusterDimMustBeSet(self, int clusterDimMustBeSet): self._pvt_ptr[0].clusterDimMustBeSet = clusterDimMustBeSet @property def requiredClusterWidth(self): return self._pvt_ptr[0].requiredClusterWidth @requiredClusterWidth.setter def requiredClusterWidth(self, int requiredClusterWidth): self._pvt_ptr[0].requiredClusterWidth = requiredClusterWidth @property def requiredClusterHeight(self): return self._pvt_ptr[0].requiredClusterHeight @requiredClusterHeight.setter def requiredClusterHeight(self, int requiredClusterHeight): self._pvt_ptr[0].requiredClusterHeight = requiredClusterHeight @property def requiredClusterDepth(self): return self._pvt_ptr[0].requiredClusterDepth @requiredClusterDepth.setter def requiredClusterDepth(self, int requiredClusterDepth): self._pvt_ptr[0].requiredClusterDepth = requiredClusterDepth @property def clusterSchedulingPolicyPreference(self): return self._pvt_ptr[0].clusterSchedulingPolicyPreference @clusterSchedulingPolicyPreference.setter def clusterSchedulingPolicyPreference(self, int clusterSchedulingPolicyPreference): self._pvt_ptr[0].clusterSchedulingPolicyPreference = clusterSchedulingPolicyPreference @property def nonPortableClusterSizeAllowed(self): return self._pvt_ptr[0].nonPortableClusterSizeAllowed @nonPortableClusterSizeAllowed.setter def nonPortableClusterSizeAllowed(self, int nonPortableClusterSizeAllowed): self._pvt_ptr[0].nonPortableClusterSizeAllowed = nonPortableClusterSizeAllowed @property def reserved(self): return self._pvt_ptr[0].reserved @reserved.setter def reserved(self, reserved): self._pvt_ptr[0].reserved = reserved cdef class cudaMemLocation: """ Specifies a memory location. To specify a gpu, set type = cudaMemLocationTypeDevice and set id = the gpu's device ordinal. To specify a cpu NUMA node, set type = cudaMemLocationTypeHostNuma and set id = host NUMA node id. Attributes ---------- type : cudaMemLocationType Specifies the location type, which modifies the meaning of id. id : int identifier for a given this location's ::CUmemLocationType. Methods ------- getPtr() Get memory address of class instance """ def __cinit__(self, void_ptr _ptr = 0): if _ptr == 0: self._pvt_ptr = &self._pvt_val else: self._pvt_ptr = _ptr def __init__(self, void_ptr _ptr = 0): pass def __dealloc__(self): pass def getPtr(self): return self._pvt_ptr def __repr__(self): if self._pvt_ptr is not NULL: str_list = [] try: str_list += ['type : ' + str(self.type)] except ValueError: str_list += ['type : '] try: str_list += ['id : ' + str(self.id)] except ValueError: str_list += ['id : '] return '\n'.join(str_list) else: return '' @property def type(self): if self._pvt_ptr[0].type not in _dict_cudaMemLocationType: return None return _dict_cudaMemLocationType[self._pvt_ptr[0].type] @type.setter def type(self, type not None : cudaMemLocationType): self._pvt_ptr[0].type = type.value @property def id(self): return self._pvt_ptr[0].id @id.setter def id(self, int id): self._pvt_ptr[0].id = id cdef class cudaMemAccessDesc: """ Memory access descriptor Attributes ---------- location : cudaMemLocation Location on which the request is to change it's accessibility flags : cudaMemAccessFlags ::CUmemProt accessibility flags to set on the request Methods ------- getPtr() Get memory address of class instance """ def __cinit__(self, void_ptr _ptr = 0): if _ptr == 0: self._pvt_ptr = &self._pvt_val else: self._pvt_ptr = _ptr def __init__(self, void_ptr _ptr = 0): pass self._location = cudaMemLocation(_ptr=&self._pvt_ptr[0].location) def __dealloc__(self): pass def getPtr(self): return self._pvt_ptr def __repr__(self): if self._pvt_ptr is not NULL: str_list = [] try: str_list += ['location :\n' + '\n'.join([' ' + line for line in str(self.location).splitlines()])] except ValueError: str_list += ['location : '] try: str_list += ['flags : ' + str(self.flags)] except ValueError: str_list += ['flags : '] return '\n'.join(str_list) else: return '' @property def location(self): return self._location @location.setter def location(self, location not None : cudaMemLocation): string.memcpy(&self._pvt_ptr[0].location, location.getPtr(), sizeof(self._pvt_ptr[0].location)) @property def flags(self): if self._pvt_ptr[0].flags not in _dict_cudaMemAccessFlags: return None return _dict_cudaMemAccessFlags[self._pvt_ptr[0].flags] @flags.setter def flags(self, flags not None : cudaMemAccessFlags): self._pvt_ptr[0].flags = flags.value cdef class cudaMemPoolProps: """ Specifies the properties of allocations made from the pool. Attributes ---------- allocType : cudaMemAllocationType Allocation type. Currently must be specified as cudaMemAllocationTypePinned handleTypes : cudaMemAllocationHandleType Handle types that will be supported by allocations from the pool. location : cudaMemLocation Location allocations should reside. win32SecurityAttributes : Any Windows-specific LPSECURITYATTRIBUTES required when cudaMemHandleTypeWin32 is specified. This security attribute defines the scope of which exported allocations may be tranferred to other processes. In all other cases, this field is required to be zero. maxSize : size_t Maximum pool size. When set to 0, defaults to a system dependent value. usage : unsigned short Bitmask indicating intended usage for the pool. reserved : bytes reserved for future use, must be 0 Methods ------- getPtr() Get memory address of class instance """ def __cinit__(self, void_ptr _ptr = 0): if _ptr == 0: self._pvt_ptr = &self._pvt_val else: self._pvt_ptr = _ptr def __init__(self, void_ptr _ptr = 0): pass self._location = cudaMemLocation(_ptr=&self._pvt_ptr[0].location) def __dealloc__(self): pass def getPtr(self): return self._pvt_ptr def __repr__(self): if self._pvt_ptr is not NULL: str_list = [] try: str_list += ['allocType : ' + str(self.allocType)] except ValueError: str_list += ['allocType : '] try: str_list += ['handleTypes : ' + str(self.handleTypes)] except ValueError: str_list += ['handleTypes : '] try: str_list += ['location :\n' + '\n'.join([' ' + line for line in str(self.location).splitlines()])] except ValueError: str_list += ['location : '] try: str_list += ['win32SecurityAttributes : ' + hex(self.win32SecurityAttributes)] except ValueError: str_list += ['win32SecurityAttributes : '] try: str_list += ['maxSize : ' + str(self.maxSize)] except ValueError: str_list += ['maxSize : '] try: str_list += ['usage : ' + str(self.usage)] except ValueError: str_list += ['usage : '] try: str_list += ['reserved : ' + str(self.reserved)] except ValueError: str_list += ['reserved : '] return '\n'.join(str_list) else: return '' @property def allocType(self): if self._pvt_ptr[0].allocType not in _dict_cudaMemAllocationType: return None return _dict_cudaMemAllocationType[self._pvt_ptr[0].allocType] @allocType.setter def allocType(self, allocType not None : cudaMemAllocationType): self._pvt_ptr[0].allocType = allocType.value @property def handleTypes(self): if self._pvt_ptr[0].handleTypes not in _dict_cudaMemAllocationHandleType: return None return _dict_cudaMemAllocationHandleType[self._pvt_ptr[0].handleTypes] @handleTypes.setter def handleTypes(self, handleTypes not None : cudaMemAllocationHandleType): self._pvt_ptr[0].handleTypes = handleTypes.value @property def location(self): return self._location @location.setter def location(self, location not None : cudaMemLocation): string.memcpy(&self._pvt_ptr[0].location, location.getPtr(), sizeof(self._pvt_ptr[0].location)) @property def win32SecurityAttributes(self): return self._pvt_ptr[0].win32SecurityAttributes @win32SecurityAttributes.setter def win32SecurityAttributes(self, win32SecurityAttributes): _cwin32SecurityAttributes = _HelperInputVoidPtr(win32SecurityAttributes) self._pvt_ptr[0].win32SecurityAttributes = _cwin32SecurityAttributes.cptr @property def maxSize(self): return self._pvt_ptr[0].maxSize @maxSize.setter def maxSize(self, size_t maxSize): self._pvt_ptr[0].maxSize = maxSize @property def usage(self): return self._pvt_ptr[0].usage @usage.setter def usage(self, unsigned short usage): self._pvt_ptr[0].usage = usage @property def reserved(self): return PyBytes_FromStringAndSize(self._pvt_ptr[0].reserved, 54) @reserved.setter def reserved(self, reserved): if len(reserved) != 54: raise ValueError("reserved length must be 54, is " + str(len(reserved))) for i, b in enumerate(reserved): self._pvt_ptr[0].reserved[i] = b cdef class cudaMemPoolPtrExportData: """ Opaque data for exporting a pool allocation Attributes ---------- reserved : bytes Methods ------- getPtr() Get memory address of class instance """ def __cinit__(self, void_ptr _ptr = 0): if _ptr == 0: self._pvt_ptr = &self._pvt_val else: self._pvt_ptr = _ptr def __init__(self, void_ptr _ptr = 0): pass def __dealloc__(self): pass def getPtr(self): return self._pvt_ptr def __repr__(self): if self._pvt_ptr is not NULL: str_list = [] try: str_list += ['reserved : ' + str(self.reserved)] except ValueError: str_list += ['reserved : '] return '\n'.join(str_list) else: return '' @property def reserved(self): return PyBytes_FromStringAndSize(self._pvt_ptr[0].reserved, 64) @reserved.setter def reserved(self, reserved): if len(reserved) != 64: raise ValueError("reserved length must be 64, is " + str(len(reserved))) for i, b in enumerate(reserved): self._pvt_ptr[0].reserved[i] = b cdef class cudaMemAllocNodeParams: """ Memory allocation node parameters Attributes ---------- poolProps : cudaMemPoolProps in: location where the allocation should reside (specified in ::location). ::handleTypes must be cudaMemHandleTypeNone. IPC is not supported. in: array of memory access descriptors. Used to describe peer GPU access accessDescs : cudaMemAccessDesc in: number of memory access descriptors. Must not exceed the number of GPUs. accessDescCount : size_t in: Number of `accessDescs`s bytesize : size_t in: size in bytes of the requested allocation dptr : Any out: address of the allocation returned by CUDA Methods ------- getPtr() Get memory address of class instance """ def __cinit__(self, void_ptr _ptr = 0): if _ptr == 0: self._pvt_ptr = &self._pvt_val else: self._pvt_ptr = _ptr def __init__(self, void_ptr _ptr = 0): pass self._poolProps = cudaMemPoolProps(_ptr=&self._pvt_ptr[0].poolProps) def __dealloc__(self): pass if self._accessDescs is not NULL: free(self._accessDescs) def getPtr(self): return self._pvt_ptr def __repr__(self): if self._pvt_ptr is not NULL: str_list = [] try: str_list += ['poolProps :\n' + '\n'.join([' ' + line for line in str(self.poolProps).splitlines()])] except ValueError: str_list += ['poolProps : '] try: str_list += ['accessDescs : ' + str(self.accessDescs)] except ValueError: str_list += ['accessDescs : '] try: str_list += ['accessDescCount : ' + str(self.accessDescCount)] except ValueError: str_list += ['accessDescCount : '] try: str_list += ['bytesize : ' + str(self.bytesize)] except ValueError: str_list += ['bytesize : '] try: str_list += ['dptr : ' + hex(self.dptr)] except ValueError: str_list += ['dptr : '] return '\n'.join(str_list) else: return '' @property def poolProps(self): return self._poolProps @poolProps.setter def poolProps(self, poolProps not None : cudaMemPoolProps): string.memcpy(&self._pvt_ptr[0].poolProps, poolProps.getPtr(), sizeof(self._pvt_ptr[0].poolProps)) @property def accessDescs(self): arrs = [self._pvt_ptr[0].accessDescs + x*sizeof(cyruntime.cudaMemAccessDesc) for x in range(self._accessDescs_length)] return [cudaMemAccessDesc(_ptr=arr) for arr in arrs] @accessDescs.setter def accessDescs(self, val): if len(val) == 0: free(self._accessDescs) self._accessDescs_length = 0 self._pvt_ptr[0].accessDescs = NULL else: if self._accessDescs_length != len(val): free(self._accessDescs) self._accessDescs = calloc(len(val), sizeof(cyruntime.cudaMemAccessDesc)) if self._accessDescs is NULL: raise MemoryError('Failed to allocate length x size memory: ' + str(len(val)) + 'x' + str(sizeof(cyruntime.cudaMemAccessDesc))) self._accessDescs_length = len(val) self._pvt_ptr[0].accessDescs = self._accessDescs for idx in range(len(val)): string.memcpy(&self._accessDescs[idx], (val[idx])._pvt_ptr, sizeof(cyruntime.cudaMemAccessDesc)) @property def accessDescCount(self): return self._pvt_ptr[0].accessDescCount @accessDescCount.setter def accessDescCount(self, size_t accessDescCount): self._pvt_ptr[0].accessDescCount = accessDescCount @property def bytesize(self): return self._pvt_ptr[0].bytesize @bytesize.setter def bytesize(self, size_t bytesize): self._pvt_ptr[0].bytesize = bytesize @property def dptr(self): return self._pvt_ptr[0].dptr @dptr.setter def dptr(self, dptr): _cdptr = _HelperInputVoidPtr(dptr) self._pvt_ptr[0].dptr = _cdptr.cptr cdef class cudaMemAllocNodeParamsV2: """ Memory allocation node parameters Attributes ---------- poolProps : cudaMemPoolProps in: location where the allocation should reside (specified in ::location). ::handleTypes must be cudaMemHandleTypeNone. IPC is not supported. in: array of memory access descriptors. Used to describe peer GPU access accessDescs : cudaMemAccessDesc in: number of memory access descriptors. Must not exceed the number of GPUs. accessDescCount : size_t in: Number of `accessDescs`s bytesize : size_t in: size in bytes of the requested allocation dptr : Any out: address of the allocation returned by CUDA Methods ------- getPtr() Get memory address of class instance """ def __cinit__(self, void_ptr _ptr = 0): if _ptr == 0: self._pvt_ptr = &self._pvt_val else: self._pvt_ptr = _ptr def __init__(self, void_ptr _ptr = 0): pass self._poolProps = cudaMemPoolProps(_ptr=&self._pvt_ptr[0].poolProps) def __dealloc__(self): pass if self._accessDescs is not NULL: free(self._accessDescs) def getPtr(self): return self._pvt_ptr def __repr__(self): if self._pvt_ptr is not NULL: str_list = [] try: str_list += ['poolProps :\n' + '\n'.join([' ' + line for line in str(self.poolProps).splitlines()])] except ValueError: str_list += ['poolProps : '] try: str_list += ['accessDescs : ' + str(self.accessDescs)] except ValueError: str_list += ['accessDescs : '] try: str_list += ['accessDescCount : ' + str(self.accessDescCount)] except ValueError: str_list += ['accessDescCount : '] try: str_list += ['bytesize : ' + str(self.bytesize)] except ValueError: str_list += ['bytesize : '] try: str_list += ['dptr : ' + hex(self.dptr)] except ValueError: str_list += ['dptr : '] return '\n'.join(str_list) else: return '' @property def poolProps(self): return self._poolProps @poolProps.setter def poolProps(self, poolProps not None : cudaMemPoolProps): string.memcpy(&self._pvt_ptr[0].poolProps, poolProps.getPtr(), sizeof(self._pvt_ptr[0].poolProps)) @property def accessDescs(self): arrs = [self._pvt_ptr[0].accessDescs + x*sizeof(cyruntime.cudaMemAccessDesc) for x in range(self._accessDescs_length)] return [cudaMemAccessDesc(_ptr=arr) for arr in arrs] @accessDescs.setter def accessDescs(self, val): if len(val) == 0: free(self._accessDescs) self._accessDescs_length = 0 self._pvt_ptr[0].accessDescs = NULL else: if self._accessDescs_length != len(val): free(self._accessDescs) self._accessDescs = calloc(len(val), sizeof(cyruntime.cudaMemAccessDesc)) if self._accessDescs is NULL: raise MemoryError('Failed to allocate length x size memory: ' + str(len(val)) + 'x' + str(sizeof(cyruntime.cudaMemAccessDesc))) self._accessDescs_length = len(val) self._pvt_ptr[0].accessDescs = self._accessDescs for idx in range(len(val)): string.memcpy(&self._accessDescs[idx], (val[idx])._pvt_ptr, sizeof(cyruntime.cudaMemAccessDesc)) @property def accessDescCount(self): return self._pvt_ptr[0].accessDescCount @accessDescCount.setter def accessDescCount(self, size_t accessDescCount): self._pvt_ptr[0].accessDescCount = accessDescCount @property def bytesize(self): return self._pvt_ptr[0].bytesize @bytesize.setter def bytesize(self, size_t bytesize): self._pvt_ptr[0].bytesize = bytesize @property def dptr(self): return self._pvt_ptr[0].dptr @dptr.setter def dptr(self, dptr): _cdptr = _HelperInputVoidPtr(dptr) self._pvt_ptr[0].dptr = _cdptr.cptr cdef class cudaMemFreeNodeParams: """ Memory free node parameters Attributes ---------- dptr : Any in: the pointer to free Methods ------- getPtr() Get memory address of class instance """ def __cinit__(self, void_ptr _ptr = 0): if _ptr == 0: self._pvt_ptr = &self._pvt_val else: self._pvt_ptr = _ptr def __init__(self, void_ptr _ptr = 0): pass def __dealloc__(self): pass def getPtr(self): return self._pvt_ptr def __repr__(self): if self._pvt_ptr is not NULL: str_list = [] try: str_list += ['dptr : ' + hex(self.dptr)] except ValueError: str_list += ['dptr : '] return '\n'.join(str_list) else: return '' @property def dptr(self): return self._pvt_ptr[0].dptr @dptr.setter def dptr(self, dptr): _cdptr = _HelperInputVoidPtr(dptr) self._pvt_ptr[0].dptr = _cdptr.cptr cdef class cudaMemcpyAttributes: """ Attributes specific to copies within a batch. For more details on usage see cudaMemcpyBatchAsync. Attributes ---------- srcAccessOrder : cudaMemcpySrcAccessOrder Source access ordering to be observed for copies with this attribute. srcLocHint : cudaMemLocation Hint location for the source operand. Ignored when the pointers are not managed memory or memory allocated outside CUDA. dstLocHint : cudaMemLocation Hint location for the destination operand. Ignored when the pointers are not managed memory or memory allocated outside CUDA. flags : unsigned int Additional flags for copies with this attribute. See cudaMemcpyFlags. Methods ------- getPtr() Get memory address of class instance """ def __cinit__(self, void_ptr _ptr = 0): if _ptr == 0: self._pvt_ptr = &self._pvt_val else: self._pvt_ptr = _ptr def __init__(self, void_ptr _ptr = 0): pass self._srcLocHint = cudaMemLocation(_ptr=&self._pvt_ptr[0].srcLocHint) self._dstLocHint = cudaMemLocation(_ptr=&self._pvt_ptr[0].dstLocHint) def __dealloc__(self): pass def getPtr(self): return self._pvt_ptr def __repr__(self): if self._pvt_ptr is not NULL: str_list = [] try: str_list += ['srcAccessOrder : ' + str(self.srcAccessOrder)] except ValueError: str_list += ['srcAccessOrder : '] try: str_list += ['srcLocHint :\n' + '\n'.join([' ' + line for line in str(self.srcLocHint).splitlines()])] except ValueError: str_list += ['srcLocHint : '] try: str_list += ['dstLocHint :\n' + '\n'.join([' ' + line for line in str(self.dstLocHint).splitlines()])] except ValueError: str_list += ['dstLocHint : '] try: str_list += ['flags : ' + str(self.flags)] except ValueError: str_list += ['flags : '] return '\n'.join(str_list) else: return '' @property def srcAccessOrder(self): if self._pvt_ptr[0].srcAccessOrder not in _dict_cudaMemcpySrcAccessOrder: return None return _dict_cudaMemcpySrcAccessOrder[self._pvt_ptr[0].srcAccessOrder] @srcAccessOrder.setter def srcAccessOrder(self, srcAccessOrder not None : cudaMemcpySrcAccessOrder): self._pvt_ptr[0].srcAccessOrder = srcAccessOrder.value @property def srcLocHint(self): return self._srcLocHint @srcLocHint.setter def srcLocHint(self, srcLocHint not None : cudaMemLocation): string.memcpy(&self._pvt_ptr[0].srcLocHint, srcLocHint.getPtr(), sizeof(self._pvt_ptr[0].srcLocHint)) @property def dstLocHint(self): return self._dstLocHint @dstLocHint.setter def dstLocHint(self, dstLocHint not None : cudaMemLocation): string.memcpy(&self._pvt_ptr[0].dstLocHint, dstLocHint.getPtr(), sizeof(self._pvt_ptr[0].dstLocHint)) @property def flags(self): return self._pvt_ptr[0].flags @flags.setter def flags(self, unsigned int flags): self._pvt_ptr[0].flags = flags cdef class cudaOffset3D: """ Struct representing offset into a cudaArray_t in elements Attributes ---------- x : size_t y : size_t z : size_t Methods ------- getPtr() Get memory address of class instance """ def __cinit__(self, void_ptr _ptr = 0): if _ptr == 0: self._pvt_ptr = &self._pvt_val else: self._pvt_ptr = _ptr def __init__(self, void_ptr _ptr = 0): pass def __dealloc__(self): pass def getPtr(self): return self._pvt_ptr def __repr__(self): if self._pvt_ptr is not NULL: str_list = [] try: str_list += ['x : ' + str(self.x)] except ValueError: str_list += ['x : '] try: str_list += ['y : ' + str(self.y)] except ValueError: str_list += ['y : '] try: str_list += ['z : ' + str(self.z)] except ValueError: str_list += ['z : '] return '\n'.join(str_list) else: return '' @property def x(self): return self._pvt_ptr[0].x @x.setter def x(self, size_t x): self._pvt_ptr[0].x = x @property def y(self): return self._pvt_ptr[0].y @y.setter def y(self, size_t y): self._pvt_ptr[0].y = y @property def z(self): return self._pvt_ptr[0].z @z.setter def z(self, size_t z): self._pvt_ptr[0].z = z cdef class anon_struct5: """ Attributes ---------- ptr : Any rowLength : size_t layerHeight : size_t locHint : cudaMemLocation Methods ------- getPtr() Get memory address of class instance """ def __cinit__(self, void_ptr _ptr): self._pvt_ptr = _ptr def __init__(self, void_ptr _ptr): pass self._locHint = cudaMemLocation(_ptr=&self._pvt_ptr[0].op.ptr.locHint) def __dealloc__(self): pass def getPtr(self): return &self._pvt_ptr[0].op.ptr def __repr__(self): if self._pvt_ptr is not NULL: str_list = [] try: str_list += ['ptr : ' + hex(self.ptr)] except ValueError: str_list += ['ptr : '] try: str_list += ['rowLength : ' + str(self.rowLength)] except ValueError: str_list += ['rowLength : '] try: str_list += ['layerHeight : ' + str(self.layerHeight)] except ValueError: str_list += ['layerHeight : '] try: str_list += ['locHint :\n' + '\n'.join([' ' + line for line in str(self.locHint).splitlines()])] except ValueError: str_list += ['locHint : '] return '\n'.join(str_list) else: return '' @property def ptr(self): return self._pvt_ptr[0].op.ptr.ptr @ptr.setter def ptr(self, ptr): _cptr = _HelperInputVoidPtr(ptr) self._pvt_ptr[0].op.ptr.ptr = _cptr.cptr @property def rowLength(self): return self._pvt_ptr[0].op.ptr.rowLength @rowLength.setter def rowLength(self, size_t rowLength): self._pvt_ptr[0].op.ptr.rowLength = rowLength @property def layerHeight(self): return self._pvt_ptr[0].op.ptr.layerHeight @layerHeight.setter def layerHeight(self, size_t layerHeight): self._pvt_ptr[0].op.ptr.layerHeight = layerHeight @property def locHint(self): return self._locHint @locHint.setter def locHint(self, locHint not None : cudaMemLocation): string.memcpy(&self._pvt_ptr[0].op.ptr.locHint, locHint.getPtr(), sizeof(self._pvt_ptr[0].op.ptr.locHint)) cdef class anon_struct6: """ Attributes ---------- array : cudaArray_t offset : cudaOffset3D Methods ------- getPtr() Get memory address of class instance """ def __cinit__(self, void_ptr _ptr): self._pvt_ptr = _ptr def __init__(self, void_ptr _ptr): pass self._array = cudaArray_t(_ptr=&self._pvt_ptr[0].op.array.array) self._offset = cudaOffset3D(_ptr=&self._pvt_ptr[0].op.array.offset) def __dealloc__(self): pass def getPtr(self): return &self._pvt_ptr[0].op.array def __repr__(self): if self._pvt_ptr is not NULL: str_list = [] try: str_list += ['array : ' + str(self.array)] except ValueError: str_list += ['array : '] try: str_list += ['offset :\n' + '\n'.join([' ' + line for line in str(self.offset).splitlines()])] except ValueError: str_list += ['offset : '] return '\n'.join(str_list) else: return '' @property def array(self): return self._array @array.setter def array(self, array): cdef cyruntime.cudaArray_t cyarray if array is None: cyarray = 0 elif isinstance(array, (cudaArray_t,)): parray = int(array) cyarray = parray else: parray = int(cudaArray_t(array)) cyarray = parray self._array._pvt_ptr[0] = cyarray @property def offset(self): return self._offset @offset.setter def offset(self, offset not None : cudaOffset3D): string.memcpy(&self._pvt_ptr[0].op.array.offset, offset.getPtr(), sizeof(self._pvt_ptr[0].op.array.offset)) cdef class anon_union1: """ Attributes ---------- ptr : anon_struct5 array : anon_struct6 Methods ------- getPtr() Get memory address of class instance """ def __cinit__(self, void_ptr _ptr): self._pvt_ptr = _ptr def __init__(self, void_ptr _ptr): pass self._ptr = anon_struct5(_ptr=self._pvt_ptr) self._array = anon_struct6(_ptr=self._pvt_ptr) def __dealloc__(self): pass def getPtr(self): return &self._pvt_ptr[0].op def __repr__(self): if self._pvt_ptr is not NULL: str_list = [] try: str_list += ['ptr :\n' + '\n'.join([' ' + line for line in str(self.ptr).splitlines()])] except ValueError: str_list += ['ptr : '] try: str_list += ['array :\n' + '\n'.join([' ' + line for line in str(self.array).splitlines()])] except ValueError: str_list += ['array : '] return '\n'.join(str_list) else: return '' @property def ptr(self): return self._ptr @ptr.setter def ptr(self, ptr not None : anon_struct5): string.memcpy(&self._pvt_ptr[0].op.ptr, ptr.getPtr(), sizeof(self._pvt_ptr[0].op.ptr)) @property def array(self): return self._array @array.setter def array(self, array not None : anon_struct6): string.memcpy(&self._pvt_ptr[0].op.array, array.getPtr(), sizeof(self._pvt_ptr[0].op.array)) cdef class cudaMemcpy3DOperand: """ Struct representing an operand for copy with cudaMemcpy3DBatchAsync Attributes ---------- type : cudaMemcpy3DOperandType op : anon_union1 Methods ------- getPtr() Get memory address of class instance """ def __cinit__(self, void_ptr _ptr = 0): if _ptr == 0: self._val_ptr = calloc(1, sizeof(cyruntime.cudaMemcpy3DOperand)) self._pvt_ptr = self._val_ptr else: self._pvt_ptr = _ptr def __init__(self, void_ptr _ptr = 0): pass self._op = anon_union1(_ptr=self._pvt_ptr) def __dealloc__(self): if self._val_ptr is not NULL: free(self._val_ptr) def getPtr(self): return self._pvt_ptr def __repr__(self): if self._pvt_ptr is not NULL: str_list = [] try: str_list += ['type : ' + str(self.type)] except ValueError: str_list += ['type : '] try: str_list += ['op :\n' + '\n'.join([' ' + line for line in str(self.op).splitlines()])] except ValueError: str_list += ['op : '] return '\n'.join(str_list) else: return '' @property def type(self): if self._pvt_ptr[0].type not in _dict_cudaMemcpy3DOperandType: return None return _dict_cudaMemcpy3DOperandType[self._pvt_ptr[0].type] @type.setter def type(self, type not None : cudaMemcpy3DOperandType): self._pvt_ptr[0].type = type.value @property def op(self): return self._op @op.setter def op(self, op not None : anon_union1): string.memcpy(&self._pvt_ptr[0].op, op.getPtr(), sizeof(self._pvt_ptr[0].op)) cdef class cudaMemcpy3DBatchOp: """ Attributes ---------- src : cudaMemcpy3DOperand Source memcpy operand. dst : cudaMemcpy3DOperand Destination memcpy operand. extent : cudaExtent Extents of the memcpy between src and dst. The width, height and depth components must not be 0. srcAccessOrder : cudaMemcpySrcAccessOrder Source access ordering to be observed for copy from src to dst. flags : unsigned int Additional flags for copy from src to dst. See cudaMemcpyFlags. Methods ------- getPtr() Get memory address of class instance """ def __cinit__(self, void_ptr _ptr = 0): if _ptr == 0: self._pvt_ptr = &self._pvt_val else: self._pvt_ptr = _ptr def __init__(self, void_ptr _ptr = 0): pass self._src = cudaMemcpy3DOperand(_ptr=&self._pvt_ptr[0].src) self._dst = cudaMemcpy3DOperand(_ptr=&self._pvt_ptr[0].dst) self._extent = cudaExtent(_ptr=&self._pvt_ptr[0].extent) def __dealloc__(self): pass def getPtr(self): return self._pvt_ptr def __repr__(self): if self._pvt_ptr is not NULL: str_list = [] try: str_list += ['src :\n' + '\n'.join([' ' + line for line in str(self.src).splitlines()])] except ValueError: str_list += ['src : '] try: str_list += ['dst :\n' + '\n'.join([' ' + line for line in str(self.dst).splitlines()])] except ValueError: str_list += ['dst : '] try: str_list += ['extent :\n' + '\n'.join([' ' + line for line in str(self.extent).splitlines()])] except ValueError: str_list += ['extent : '] try: str_list += ['srcAccessOrder : ' + str(self.srcAccessOrder)] except ValueError: str_list += ['srcAccessOrder : '] try: str_list += ['flags : ' + str(self.flags)] except ValueError: str_list += ['flags : '] return '\n'.join(str_list) else: return '' @property def src(self): return self._src @src.setter def src(self, src not None : cudaMemcpy3DOperand): string.memcpy(&self._pvt_ptr[0].src, src.getPtr(), sizeof(self._pvt_ptr[0].src)) @property def dst(self): return self._dst @dst.setter def dst(self, dst not None : cudaMemcpy3DOperand): string.memcpy(&self._pvt_ptr[0].dst, dst.getPtr(), sizeof(self._pvt_ptr[0].dst)) @property def extent(self): return self._extent @extent.setter def extent(self, extent not None : cudaExtent): string.memcpy(&self._pvt_ptr[0].extent, extent.getPtr(), sizeof(self._pvt_ptr[0].extent)) @property def srcAccessOrder(self): if self._pvt_ptr[0].srcAccessOrder not in _dict_cudaMemcpySrcAccessOrder: return None return _dict_cudaMemcpySrcAccessOrder[self._pvt_ptr[0].srcAccessOrder] @srcAccessOrder.setter def srcAccessOrder(self, srcAccessOrder not None : cudaMemcpySrcAccessOrder): self._pvt_ptr[0].srcAccessOrder = srcAccessOrder.value @property def flags(self): return self._pvt_ptr[0].flags @flags.setter def flags(self, unsigned int flags): self._pvt_ptr[0].flags = flags cdef class CUuuid_st: """ Attributes ---------- bytes : bytes < CUDA definition of UUID Methods ------- getPtr() Get memory address of class instance """ def __cinit__(self, void_ptr _ptr = 0): if _ptr == 0: self._pvt_ptr = &self._pvt_val else: self._pvt_ptr = _ptr def __init__(self, void_ptr _ptr = 0): pass def __dealloc__(self): pass def getPtr(self): return self._pvt_ptr def __repr__(self): if self._pvt_ptr is not NULL: str_list = [] try: str_list += ['bytes : ' + str(self.bytes.hex())] except ValueError: str_list += ['bytes : '] return '\n'.join(str_list) else: return '' @property def bytes(self): return PyBytes_FromStringAndSize(self._pvt_ptr[0].bytes, 16) cdef class cudaDeviceProp: """ CUDA device properties Attributes ---------- name : bytes ASCII string identifying device uuid : cudaUUID_t 16-byte unique identifier luid : bytes 8-byte locally unique identifier. Value is undefined on TCC and non-Windows platforms luidDeviceNodeMask : unsigned int LUID device node mask. Value is undefined on TCC and non-Windows platforms totalGlobalMem : size_t Global memory available on device in bytes sharedMemPerBlock : size_t Shared memory available per block in bytes regsPerBlock : int 32-bit registers available per block warpSize : int Warp size in threads memPitch : size_t Maximum pitch in bytes allowed by memory copies maxThreadsPerBlock : int Maximum number of threads per block maxThreadsDim : list[int] Maximum size of each dimension of a block maxGridSize : list[int] Maximum size of each dimension of a grid clockRate : int Deprecated, Clock frequency in kilohertz totalConstMem : size_t Constant memory available on device in bytes major : int Major compute capability minor : int Minor compute capability textureAlignment : size_t Alignment requirement for textures texturePitchAlignment : size_t Pitch alignment requirement for texture references bound to pitched memory deviceOverlap : int Device can concurrently copy memory and execute a kernel. Deprecated. Use instead asyncEngineCount. multiProcessorCount : int Number of multiprocessors on device kernelExecTimeoutEnabled : int Deprecated, Specified whether there is a run time limit on kernels integrated : int Device is integrated as opposed to discrete canMapHostMemory : int Device can map host memory with cudaHostAlloc/cudaHostGetDevicePointer computeMode : int Deprecated, Compute mode (See cudaComputeMode) maxTexture1D : int Maximum 1D texture size maxTexture1DMipmap : int Maximum 1D mipmapped texture size maxTexture1DLinear : int Deprecated, do not use. Use cudaDeviceGetTexture1DLinearMaxWidth() or cuDeviceGetTexture1DLinearMaxWidth() instead. maxTexture2D : list[int] Maximum 2D texture dimensions maxTexture2DMipmap : list[int] Maximum 2D mipmapped texture dimensions maxTexture2DLinear : list[int] Maximum dimensions (width, height, pitch) for 2D textures bound to pitched memory maxTexture2DGather : list[int] Maximum 2D texture dimensions if texture gather operations have to be performed maxTexture3D : list[int] Maximum 3D texture dimensions maxTexture3DAlt : list[int] Maximum alternate 3D texture dimensions maxTextureCubemap : int Maximum Cubemap texture dimensions maxTexture1DLayered : list[int] Maximum 1D layered texture dimensions maxTexture2DLayered : list[int] Maximum 2D layered texture dimensions maxTextureCubemapLayered : list[int] Maximum Cubemap layered texture dimensions maxSurface1D : int Maximum 1D surface size maxSurface2D : list[int] Maximum 2D surface dimensions maxSurface3D : list[int] Maximum 3D surface dimensions maxSurface1DLayered : list[int] Maximum 1D layered surface dimensions maxSurface2DLayered : list[int] Maximum 2D layered surface dimensions maxSurfaceCubemap : int Maximum Cubemap surface dimensions maxSurfaceCubemapLayered : list[int] Maximum Cubemap layered surface dimensions surfaceAlignment : size_t Alignment requirements for surfaces concurrentKernels : int Device can possibly execute multiple kernels concurrently ECCEnabled : int Device has ECC support enabled pciBusID : int PCI bus ID of the device pciDeviceID : int PCI device ID of the device pciDomainID : int PCI domain ID of the device tccDriver : int 1 if device is a Tesla device using TCC driver, 0 otherwise asyncEngineCount : int Number of asynchronous engines unifiedAddressing : int Device shares a unified address space with the host memoryClockRate : int Deprecated, Peak memory clock frequency in kilohertz memoryBusWidth : int Global memory bus width in bits l2CacheSize : int Size of L2 cache in bytes persistingL2CacheMaxSize : int Device's maximum l2 persisting lines capacity setting in bytes maxThreadsPerMultiProcessor : int Maximum resident threads per multiprocessor streamPrioritiesSupported : int Device supports stream priorities globalL1CacheSupported : int Device supports caching globals in L1 localL1CacheSupported : int Device supports caching locals in L1 sharedMemPerMultiprocessor : size_t Shared memory available per multiprocessor in bytes regsPerMultiprocessor : int 32-bit registers available per multiprocessor managedMemory : int Device supports allocating managed memory on this system isMultiGpuBoard : int Device is on a multi-GPU board multiGpuBoardGroupID : int Unique identifier for a group of devices on the same multi-GPU board hostNativeAtomicSupported : int Link between the device and the host supports native atomic operations singleToDoublePrecisionPerfRatio : int Deprecated, Ratio of single precision performance (in floating- point operations per second) to double precision performance pageableMemoryAccess : int Device supports coherently accessing pageable memory without calling cudaHostRegister on it concurrentManagedAccess : int Device can coherently access managed memory concurrently with the CPU computePreemptionSupported : int Device supports Compute Preemption canUseHostPointerForRegisteredMem : int Device can access host registered memory at the same virtual address as the CPU cooperativeLaunch : int Device supports launching cooperative kernels via cudaLaunchCooperativeKernel cooperativeMultiDeviceLaunch : int Deprecated, cudaLaunchCooperativeKernelMultiDevice is deprecated. sharedMemPerBlockOptin : size_t Per device maximum shared memory per block usable by special opt in pageableMemoryAccessUsesHostPageTables : int Device accesses pageable memory via the host's page tables directManagedMemAccessFromHost : int Host can directly access managed memory on the device without migration. maxBlocksPerMultiProcessor : int Maximum number of resident blocks per multiprocessor accessPolicyMaxWindowSize : int The maximum value of cudaAccessPolicyWindow::num_bytes. reservedSharedMemPerBlock : size_t Shared memory reserved by CUDA driver per block in bytes hostRegisterSupported : int Device supports host memory registration via cudaHostRegister. sparseCudaArraySupported : int 1 if the device supports sparse CUDA arrays and sparse CUDA mipmapped arrays, 0 otherwise hostRegisterReadOnlySupported : int Device supports using the cudaHostRegister flag cudaHostRegisterReadOnly to register memory that must be mapped as read-only to the GPU timelineSemaphoreInteropSupported : int External timeline semaphore interop is supported on the device memoryPoolsSupported : int 1 if the device supports using the cudaMallocAsync and cudaMemPool family of APIs, 0 otherwise gpuDirectRDMASupported : int 1 if the device supports GPUDirect RDMA APIs, 0 otherwise gpuDirectRDMAFlushWritesOptions : unsigned int Bitmask to be interpreted according to the cudaFlushGPUDirectRDMAWritesOptions enum gpuDirectRDMAWritesOrdering : int See the cudaGPUDirectRDMAWritesOrdering enum for numerical values memoryPoolSupportedHandleTypes : unsigned int Bitmask of handle types supported with mempool-based IPC deferredMappingCudaArraySupported : int 1 if the device supports deferred mapping CUDA arrays and CUDA mipmapped arrays ipcEventSupported : int Device supports IPC Events. clusterLaunch : int Indicates device supports cluster launch unifiedFunctionPointers : int Indicates device supports unified pointers reserved : list[int] Reserved for future use Methods ------- getPtr() Get memory address of class instance """ def __cinit__(self, void_ptr _ptr = 0): if _ptr == 0: self._pvt_ptr = &self._pvt_val else: self._pvt_ptr = _ptr def __init__(self, void_ptr _ptr = 0): pass self._uuid = cudaUUID_t(_ptr=&self._pvt_ptr[0].uuid) def __dealloc__(self): pass def getPtr(self): return self._pvt_ptr def __repr__(self): if self._pvt_ptr is not NULL: str_list = [] try: str_list += ['name : ' + self.name.decode('utf-8')] except ValueError: str_list += ['name : '] try: str_list += ['uuid :\n' + '\n'.join([' ' + line for line in str(self.uuid).splitlines()])] except ValueError: str_list += ['uuid : '] try: str_list += ['luid : ' + self.luid.hex()] except ValueError: str_list += ['luid : '] try: str_list += ['luidDeviceNodeMask : ' + str(self.luidDeviceNodeMask)] except ValueError: str_list += ['luidDeviceNodeMask : '] try: str_list += ['totalGlobalMem : ' + str(self.totalGlobalMem)] except ValueError: str_list += ['totalGlobalMem : '] try: str_list += ['sharedMemPerBlock : ' + str(self.sharedMemPerBlock)] except ValueError: str_list += ['sharedMemPerBlock : '] try: str_list += ['regsPerBlock : ' + str(self.regsPerBlock)] except ValueError: str_list += ['regsPerBlock : '] try: str_list += ['warpSize : ' + str(self.warpSize)] except ValueError: str_list += ['warpSize : '] try: str_list += ['memPitch : ' + str(self.memPitch)] except ValueError: str_list += ['memPitch : '] try: str_list += ['maxThreadsPerBlock : ' + str(self.maxThreadsPerBlock)] except ValueError: str_list += ['maxThreadsPerBlock : '] try: str_list += ['maxThreadsDim : ' + str(self.maxThreadsDim)] except ValueError: str_list += ['maxThreadsDim : '] try: str_list += ['maxGridSize : ' + str(self.maxGridSize)] except ValueError: str_list += ['maxGridSize : '] try: str_list += ['clockRate : ' + str(self.clockRate)] except ValueError: str_list += ['clockRate : '] try: str_list += ['totalConstMem : ' + str(self.totalConstMem)] except ValueError: str_list += ['totalConstMem : '] try: str_list += ['major : ' + str(self.major)] except ValueError: str_list += ['major : '] try: str_list += ['minor : ' + str(self.minor)] except ValueError: str_list += ['minor : '] try: str_list += ['textureAlignment : ' + str(self.textureAlignment)] except ValueError: str_list += ['textureAlignment : '] try: str_list += ['texturePitchAlignment : ' + str(self.texturePitchAlignment)] except ValueError: str_list += ['texturePitchAlignment : '] try: str_list += ['deviceOverlap : ' + str(self.deviceOverlap)] except ValueError: str_list += ['deviceOverlap : '] try: str_list += ['multiProcessorCount : ' + str(self.multiProcessorCount)] except ValueError: str_list += ['multiProcessorCount : '] try: str_list += ['kernelExecTimeoutEnabled : ' + str(self.kernelExecTimeoutEnabled)] except ValueError: str_list += ['kernelExecTimeoutEnabled : '] try: str_list += ['integrated : ' + str(self.integrated)] except ValueError: str_list += ['integrated : '] try: str_list += ['canMapHostMemory : ' + str(self.canMapHostMemory)] except ValueError: str_list += ['canMapHostMemory : '] try: str_list += ['computeMode : ' + str(self.computeMode)] except ValueError: str_list += ['computeMode : '] try: str_list += ['maxTexture1D : ' + str(self.maxTexture1D)] except ValueError: str_list += ['maxTexture1D : '] try: str_list += ['maxTexture1DMipmap : ' + str(self.maxTexture1DMipmap)] except ValueError: str_list += ['maxTexture1DMipmap : '] try: str_list += ['maxTexture1DLinear : ' + str(self.maxTexture1DLinear)] except ValueError: str_list += ['maxTexture1DLinear : '] try: str_list += ['maxTexture2D : ' + str(self.maxTexture2D)] except ValueError: str_list += ['maxTexture2D : '] try: str_list += ['maxTexture2DMipmap : ' + str(self.maxTexture2DMipmap)] except ValueError: str_list += ['maxTexture2DMipmap : '] try: str_list += ['maxTexture2DLinear : ' + str(self.maxTexture2DLinear)] except ValueError: str_list += ['maxTexture2DLinear : '] try: str_list += ['maxTexture2DGather : ' + str(self.maxTexture2DGather)] except ValueError: str_list += ['maxTexture2DGather : '] try: str_list += ['maxTexture3D : ' + str(self.maxTexture3D)] except ValueError: str_list += ['maxTexture3D : '] try: str_list += ['maxTexture3DAlt : ' + str(self.maxTexture3DAlt)] except ValueError: str_list += ['maxTexture3DAlt : '] try: str_list += ['maxTextureCubemap : ' + str(self.maxTextureCubemap)] except ValueError: str_list += ['maxTextureCubemap : '] try: str_list += ['maxTexture1DLayered : ' + str(self.maxTexture1DLayered)] except ValueError: str_list += ['maxTexture1DLayered : '] try: str_list += ['maxTexture2DLayered : ' + str(self.maxTexture2DLayered)] except ValueError: str_list += ['maxTexture2DLayered : '] try: str_list += ['maxTextureCubemapLayered : ' + str(self.maxTextureCubemapLayered)] except ValueError: str_list += ['maxTextureCubemapLayered : '] try: str_list += ['maxSurface1D : ' + str(self.maxSurface1D)] except ValueError: str_list += ['maxSurface1D : '] try: str_list += ['maxSurface2D : ' + str(self.maxSurface2D)] except ValueError: str_list += ['maxSurface2D : '] try: str_list += ['maxSurface3D : ' + str(self.maxSurface3D)] except ValueError: str_list += ['maxSurface3D : '] try: str_list += ['maxSurface1DLayered : ' + str(self.maxSurface1DLayered)] except ValueError: str_list += ['maxSurface1DLayered : '] try: str_list += ['maxSurface2DLayered : ' + str(self.maxSurface2DLayered)] except ValueError: str_list += ['maxSurface2DLayered : '] try: str_list += ['maxSurfaceCubemap : ' + str(self.maxSurfaceCubemap)] except ValueError: str_list += ['maxSurfaceCubemap : '] try: str_list += ['maxSurfaceCubemapLayered : ' + str(self.maxSurfaceCubemapLayered)] except ValueError: str_list += ['maxSurfaceCubemapLayered : '] try: str_list += ['surfaceAlignment : ' + str(self.surfaceAlignment)] except ValueError: str_list += ['surfaceAlignment : '] try: str_list += ['concurrentKernels : ' + str(self.concurrentKernels)] except ValueError: str_list += ['concurrentKernels : '] try: str_list += ['ECCEnabled : ' + str(self.ECCEnabled)] except ValueError: str_list += ['ECCEnabled : '] try: str_list += ['pciBusID : ' + str(self.pciBusID)] except ValueError: str_list += ['pciBusID : '] try: str_list += ['pciDeviceID : ' + str(self.pciDeviceID)] except ValueError: str_list += ['pciDeviceID : '] try: str_list += ['pciDomainID : ' + str(self.pciDomainID)] except ValueError: str_list += ['pciDomainID : '] try: str_list += ['tccDriver : ' + str(self.tccDriver)] except ValueError: str_list += ['tccDriver : '] try: str_list += ['asyncEngineCount : ' + str(self.asyncEngineCount)] except ValueError: str_list += ['asyncEngineCount : '] try: str_list += ['unifiedAddressing : ' + str(self.unifiedAddressing)] except ValueError: str_list += ['unifiedAddressing : '] try: str_list += ['memoryClockRate : ' + str(self.memoryClockRate)] except ValueError: str_list += ['memoryClockRate : '] try: str_list += ['memoryBusWidth : ' + str(self.memoryBusWidth)] except ValueError: str_list += ['memoryBusWidth : '] try: str_list += ['l2CacheSize : ' + str(self.l2CacheSize)] except ValueError: str_list += ['l2CacheSize : '] try: str_list += ['persistingL2CacheMaxSize : ' + str(self.persistingL2CacheMaxSize)] except ValueError: str_list += ['persistingL2CacheMaxSize : '] try: str_list += ['maxThreadsPerMultiProcessor : ' + str(self.maxThreadsPerMultiProcessor)] except ValueError: str_list += ['maxThreadsPerMultiProcessor : '] try: str_list += ['streamPrioritiesSupported : ' + str(self.streamPrioritiesSupported)] except ValueError: str_list += ['streamPrioritiesSupported : '] try: str_list += ['globalL1CacheSupported : ' + str(self.globalL1CacheSupported)] except ValueError: str_list += ['globalL1CacheSupported : '] try: str_list += ['localL1CacheSupported : ' + str(self.localL1CacheSupported)] except ValueError: str_list += ['localL1CacheSupported : '] try: str_list += ['sharedMemPerMultiprocessor : ' + str(self.sharedMemPerMultiprocessor)] except ValueError: str_list += ['sharedMemPerMultiprocessor : '] try: str_list += ['regsPerMultiprocessor : ' + str(self.regsPerMultiprocessor)] except ValueError: str_list += ['regsPerMultiprocessor : '] try: str_list += ['managedMemory : ' + str(self.managedMemory)] except ValueError: str_list += ['managedMemory : '] try: str_list += ['isMultiGpuBoard : ' + str(self.isMultiGpuBoard)] except ValueError: str_list += ['isMultiGpuBoard : '] try: str_list += ['multiGpuBoardGroupID : ' + str(self.multiGpuBoardGroupID)] except ValueError: str_list += ['multiGpuBoardGroupID : '] try: str_list += ['hostNativeAtomicSupported : ' + str(self.hostNativeAtomicSupported)] except ValueError: str_list += ['hostNativeAtomicSupported : '] try: str_list += ['singleToDoublePrecisionPerfRatio : ' + str(self.singleToDoublePrecisionPerfRatio)] except ValueError: str_list += ['singleToDoublePrecisionPerfRatio : '] try: str_list += ['pageableMemoryAccess : ' + str(self.pageableMemoryAccess)] except ValueError: str_list += ['pageableMemoryAccess : '] try: str_list += ['concurrentManagedAccess : ' + str(self.concurrentManagedAccess)] except ValueError: str_list += ['concurrentManagedAccess : '] try: str_list += ['computePreemptionSupported : ' + str(self.computePreemptionSupported)] except ValueError: str_list += ['computePreemptionSupported : '] try: str_list += ['canUseHostPointerForRegisteredMem : ' + str(self.canUseHostPointerForRegisteredMem)] except ValueError: str_list += ['canUseHostPointerForRegisteredMem : '] try: str_list += ['cooperativeLaunch : ' + str(self.cooperativeLaunch)] except ValueError: str_list += ['cooperativeLaunch : '] try: str_list += ['cooperativeMultiDeviceLaunch : ' + str(self.cooperativeMultiDeviceLaunch)] except ValueError: str_list += ['cooperativeMultiDeviceLaunch : '] try: str_list += ['sharedMemPerBlockOptin : ' + str(self.sharedMemPerBlockOptin)] except ValueError: str_list += ['sharedMemPerBlockOptin : '] try: str_list += ['pageableMemoryAccessUsesHostPageTables : ' + str(self.pageableMemoryAccessUsesHostPageTables)] except ValueError: str_list += ['pageableMemoryAccessUsesHostPageTables : '] try: str_list += ['directManagedMemAccessFromHost : ' + str(self.directManagedMemAccessFromHost)] except ValueError: str_list += ['directManagedMemAccessFromHost : '] try: str_list += ['maxBlocksPerMultiProcessor : ' + str(self.maxBlocksPerMultiProcessor)] except ValueError: str_list += ['maxBlocksPerMultiProcessor : '] try: str_list += ['accessPolicyMaxWindowSize : ' + str(self.accessPolicyMaxWindowSize)] except ValueError: str_list += ['accessPolicyMaxWindowSize : '] try: str_list += ['reservedSharedMemPerBlock : ' + str(self.reservedSharedMemPerBlock)] except ValueError: str_list += ['reservedSharedMemPerBlock : '] try: str_list += ['hostRegisterSupported : ' + str(self.hostRegisterSupported)] except ValueError: str_list += ['hostRegisterSupported : '] try: str_list += ['sparseCudaArraySupported : ' + str(self.sparseCudaArraySupported)] except ValueError: str_list += ['sparseCudaArraySupported : '] try: str_list += ['hostRegisterReadOnlySupported : ' + str(self.hostRegisterReadOnlySupported)] except ValueError: str_list += ['hostRegisterReadOnlySupported : '] try: str_list += ['timelineSemaphoreInteropSupported : ' + str(self.timelineSemaphoreInteropSupported)] except ValueError: str_list += ['timelineSemaphoreInteropSupported : '] try: str_list += ['memoryPoolsSupported : ' + str(self.memoryPoolsSupported)] except ValueError: str_list += ['memoryPoolsSupported : '] try: str_list += ['gpuDirectRDMASupported : ' + str(self.gpuDirectRDMASupported)] except ValueError: str_list += ['gpuDirectRDMASupported : '] try: str_list += ['gpuDirectRDMAFlushWritesOptions : ' + str(self.gpuDirectRDMAFlushWritesOptions)] except ValueError: str_list += ['gpuDirectRDMAFlushWritesOptions : '] try: str_list += ['gpuDirectRDMAWritesOrdering : ' + str(self.gpuDirectRDMAWritesOrdering)] except ValueError: str_list += ['gpuDirectRDMAWritesOrdering : '] try: str_list += ['memoryPoolSupportedHandleTypes : ' + str(self.memoryPoolSupportedHandleTypes)] except ValueError: str_list += ['memoryPoolSupportedHandleTypes : '] try: str_list += ['deferredMappingCudaArraySupported : ' + str(self.deferredMappingCudaArraySupported)] except ValueError: str_list += ['deferredMappingCudaArraySupported : '] try: str_list += ['ipcEventSupported : ' + str(self.ipcEventSupported)] except ValueError: str_list += ['ipcEventSupported : '] try: str_list += ['clusterLaunch : ' + str(self.clusterLaunch)] except ValueError: str_list += ['clusterLaunch : '] try: str_list += ['unifiedFunctionPointers : ' + str(self.unifiedFunctionPointers)] except ValueError: str_list += ['unifiedFunctionPointers : '] try: str_list += ['reserved : ' + str(self.reserved)] except ValueError: str_list += ['reserved : '] return '\n'.join(str_list) else: return '' @property def name(self): return self._pvt_ptr[0].name @name.setter def name(self, name): pass self._pvt_ptr[0].name = name @property def uuid(self): return self._uuid @uuid.setter def uuid(self, uuid not None : cudaUUID_t): string.memcpy(&self._pvt_ptr[0].uuid, uuid.getPtr(), sizeof(self._pvt_ptr[0].uuid)) @property def luid(self): return PyBytes_FromStringAndSize(self._pvt_ptr[0].luid, 8) @luid.setter def luid(self, luid): if len(luid) != 8: raise ValueError("luid length must be 8, is " + str(len(luid))) if CHAR_MIN == 0: for i, b in enumerate(luid): if b < 0 and b > -129: b = b + 256 self._pvt_ptr[0].luid[i] = b else: for i, b in enumerate(luid): if b > 127 and b < 256: b = b - 256 self._pvt_ptr[0].luid[i] = b @property def luidDeviceNodeMask(self): return self._pvt_ptr[0].luidDeviceNodeMask @luidDeviceNodeMask.setter def luidDeviceNodeMask(self, unsigned int luidDeviceNodeMask): self._pvt_ptr[0].luidDeviceNodeMask = luidDeviceNodeMask @property def totalGlobalMem(self): return self._pvt_ptr[0].totalGlobalMem @totalGlobalMem.setter def totalGlobalMem(self, size_t totalGlobalMem): self._pvt_ptr[0].totalGlobalMem = totalGlobalMem @property def sharedMemPerBlock(self): return self._pvt_ptr[0].sharedMemPerBlock @sharedMemPerBlock.setter def sharedMemPerBlock(self, size_t sharedMemPerBlock): self._pvt_ptr[0].sharedMemPerBlock = sharedMemPerBlock @property def regsPerBlock(self): return self._pvt_ptr[0].regsPerBlock @regsPerBlock.setter def regsPerBlock(self, int regsPerBlock): self._pvt_ptr[0].regsPerBlock = regsPerBlock @property def warpSize(self): return self._pvt_ptr[0].warpSize @warpSize.setter def warpSize(self, int warpSize): self._pvt_ptr[0].warpSize = warpSize @property def memPitch(self): return self._pvt_ptr[0].memPitch @memPitch.setter def memPitch(self, size_t memPitch): self._pvt_ptr[0].memPitch = memPitch @property def maxThreadsPerBlock(self): return self._pvt_ptr[0].maxThreadsPerBlock @maxThreadsPerBlock.setter def maxThreadsPerBlock(self, int maxThreadsPerBlock): self._pvt_ptr[0].maxThreadsPerBlock = maxThreadsPerBlock @property def maxThreadsDim(self): return self._pvt_ptr[0].maxThreadsDim @maxThreadsDim.setter def maxThreadsDim(self, maxThreadsDim): self._pvt_ptr[0].maxThreadsDim = maxThreadsDim @property def maxGridSize(self): return self._pvt_ptr[0].maxGridSize @maxGridSize.setter def maxGridSize(self, maxGridSize): self._pvt_ptr[0].maxGridSize = maxGridSize @property def clockRate(self): return self._pvt_ptr[0].clockRate @clockRate.setter def clockRate(self, int clockRate): self._pvt_ptr[0].clockRate = clockRate @property def totalConstMem(self): return self._pvt_ptr[0].totalConstMem @totalConstMem.setter def totalConstMem(self, size_t totalConstMem): self._pvt_ptr[0].totalConstMem = totalConstMem @property def major(self): return self._pvt_ptr[0].major @major.setter def major(self, int major): self._pvt_ptr[0].major = major @property def minor(self): return self._pvt_ptr[0].minor @minor.setter def minor(self, int minor): self._pvt_ptr[0].minor = minor @property def textureAlignment(self): return self._pvt_ptr[0].textureAlignment @textureAlignment.setter def textureAlignment(self, size_t textureAlignment): self._pvt_ptr[0].textureAlignment = textureAlignment @property def texturePitchAlignment(self): return self._pvt_ptr[0].texturePitchAlignment @texturePitchAlignment.setter def texturePitchAlignment(self, size_t texturePitchAlignment): self._pvt_ptr[0].texturePitchAlignment = texturePitchAlignment @property def deviceOverlap(self): return self._pvt_ptr[0].deviceOverlap @deviceOverlap.setter def deviceOverlap(self, int deviceOverlap): self._pvt_ptr[0].deviceOverlap = deviceOverlap @property def multiProcessorCount(self): return self._pvt_ptr[0].multiProcessorCount @multiProcessorCount.setter def multiProcessorCount(self, int multiProcessorCount): self._pvt_ptr[0].multiProcessorCount = multiProcessorCount @property def kernelExecTimeoutEnabled(self): return self._pvt_ptr[0].kernelExecTimeoutEnabled @kernelExecTimeoutEnabled.setter def kernelExecTimeoutEnabled(self, int kernelExecTimeoutEnabled): self._pvt_ptr[0].kernelExecTimeoutEnabled = kernelExecTimeoutEnabled @property def integrated(self): return self._pvt_ptr[0].integrated @integrated.setter def integrated(self, int integrated): self._pvt_ptr[0].integrated = integrated @property def canMapHostMemory(self): return self._pvt_ptr[0].canMapHostMemory @canMapHostMemory.setter def canMapHostMemory(self, int canMapHostMemory): self._pvt_ptr[0].canMapHostMemory = canMapHostMemory @property def computeMode(self): return self._pvt_ptr[0].computeMode @computeMode.setter def computeMode(self, int computeMode): self._pvt_ptr[0].computeMode = computeMode @property def maxTexture1D(self): return self._pvt_ptr[0].maxTexture1D @maxTexture1D.setter def maxTexture1D(self, int maxTexture1D): self._pvt_ptr[0].maxTexture1D = maxTexture1D @property def maxTexture1DMipmap(self): return self._pvt_ptr[0].maxTexture1DMipmap @maxTexture1DMipmap.setter def maxTexture1DMipmap(self, int maxTexture1DMipmap): self._pvt_ptr[0].maxTexture1DMipmap = maxTexture1DMipmap @property def maxTexture1DLinear(self): return self._pvt_ptr[0].maxTexture1DLinear @maxTexture1DLinear.setter def maxTexture1DLinear(self, int maxTexture1DLinear): self._pvt_ptr[0].maxTexture1DLinear = maxTexture1DLinear @property def maxTexture2D(self): return self._pvt_ptr[0].maxTexture2D @maxTexture2D.setter def maxTexture2D(self, maxTexture2D): self._pvt_ptr[0].maxTexture2D = maxTexture2D @property def maxTexture2DMipmap(self): return self._pvt_ptr[0].maxTexture2DMipmap @maxTexture2DMipmap.setter def maxTexture2DMipmap(self, maxTexture2DMipmap): self._pvt_ptr[0].maxTexture2DMipmap = maxTexture2DMipmap @property def maxTexture2DLinear(self): return self._pvt_ptr[0].maxTexture2DLinear @maxTexture2DLinear.setter def maxTexture2DLinear(self, maxTexture2DLinear): self._pvt_ptr[0].maxTexture2DLinear = maxTexture2DLinear @property def maxTexture2DGather(self): return self._pvt_ptr[0].maxTexture2DGather @maxTexture2DGather.setter def maxTexture2DGather(self, maxTexture2DGather): self._pvt_ptr[0].maxTexture2DGather = maxTexture2DGather @property def maxTexture3D(self): return self._pvt_ptr[0].maxTexture3D @maxTexture3D.setter def maxTexture3D(self, maxTexture3D): self._pvt_ptr[0].maxTexture3D = maxTexture3D @property def maxTexture3DAlt(self): return self._pvt_ptr[0].maxTexture3DAlt @maxTexture3DAlt.setter def maxTexture3DAlt(self, maxTexture3DAlt): self._pvt_ptr[0].maxTexture3DAlt = maxTexture3DAlt @property def maxTextureCubemap(self): return self._pvt_ptr[0].maxTextureCubemap @maxTextureCubemap.setter def maxTextureCubemap(self, int maxTextureCubemap): self._pvt_ptr[0].maxTextureCubemap = maxTextureCubemap @property def maxTexture1DLayered(self): return self._pvt_ptr[0].maxTexture1DLayered @maxTexture1DLayered.setter def maxTexture1DLayered(self, maxTexture1DLayered): self._pvt_ptr[0].maxTexture1DLayered = maxTexture1DLayered @property def maxTexture2DLayered(self): return self._pvt_ptr[0].maxTexture2DLayered @maxTexture2DLayered.setter def maxTexture2DLayered(self, maxTexture2DLayered): self._pvt_ptr[0].maxTexture2DLayered = maxTexture2DLayered @property def maxTextureCubemapLayered(self): return self._pvt_ptr[0].maxTextureCubemapLayered @maxTextureCubemapLayered.setter def maxTextureCubemapLayered(self, maxTextureCubemapLayered): self._pvt_ptr[0].maxTextureCubemapLayered = maxTextureCubemapLayered @property def maxSurface1D(self): return self._pvt_ptr[0].maxSurface1D @maxSurface1D.setter def maxSurface1D(self, int maxSurface1D): self._pvt_ptr[0].maxSurface1D = maxSurface1D @property def maxSurface2D(self): return self._pvt_ptr[0].maxSurface2D @maxSurface2D.setter def maxSurface2D(self, maxSurface2D): self._pvt_ptr[0].maxSurface2D = maxSurface2D @property def maxSurface3D(self): return self._pvt_ptr[0].maxSurface3D @maxSurface3D.setter def maxSurface3D(self, maxSurface3D): self._pvt_ptr[0].maxSurface3D = maxSurface3D @property def maxSurface1DLayered(self): return self._pvt_ptr[0].maxSurface1DLayered @maxSurface1DLayered.setter def maxSurface1DLayered(self, maxSurface1DLayered): self._pvt_ptr[0].maxSurface1DLayered = maxSurface1DLayered @property def maxSurface2DLayered(self): return self._pvt_ptr[0].maxSurface2DLayered @maxSurface2DLayered.setter def maxSurface2DLayered(self, maxSurface2DLayered): self._pvt_ptr[0].maxSurface2DLayered = maxSurface2DLayered @property def maxSurfaceCubemap(self): return self._pvt_ptr[0].maxSurfaceCubemap @maxSurfaceCubemap.setter def maxSurfaceCubemap(self, int maxSurfaceCubemap): self._pvt_ptr[0].maxSurfaceCubemap = maxSurfaceCubemap @property def maxSurfaceCubemapLayered(self): return self._pvt_ptr[0].maxSurfaceCubemapLayered @maxSurfaceCubemapLayered.setter def maxSurfaceCubemapLayered(self, maxSurfaceCubemapLayered): self._pvt_ptr[0].maxSurfaceCubemapLayered = maxSurfaceCubemapLayered @property def surfaceAlignment(self): return self._pvt_ptr[0].surfaceAlignment @surfaceAlignment.setter def surfaceAlignment(self, size_t surfaceAlignment): self._pvt_ptr[0].surfaceAlignment = surfaceAlignment @property def concurrentKernels(self): return self._pvt_ptr[0].concurrentKernels @concurrentKernels.setter def concurrentKernels(self, int concurrentKernels): self._pvt_ptr[0].concurrentKernels = concurrentKernels @property def ECCEnabled(self): return self._pvt_ptr[0].ECCEnabled @ECCEnabled.setter def ECCEnabled(self, int ECCEnabled): self._pvt_ptr[0].ECCEnabled = ECCEnabled @property def pciBusID(self): return self._pvt_ptr[0].pciBusID @pciBusID.setter def pciBusID(self, int pciBusID): self._pvt_ptr[0].pciBusID = pciBusID @property def pciDeviceID(self): return self._pvt_ptr[0].pciDeviceID @pciDeviceID.setter def pciDeviceID(self, int pciDeviceID): self._pvt_ptr[0].pciDeviceID = pciDeviceID @property def pciDomainID(self): return self._pvt_ptr[0].pciDomainID @pciDomainID.setter def pciDomainID(self, int pciDomainID): self._pvt_ptr[0].pciDomainID = pciDomainID @property def tccDriver(self): return self._pvt_ptr[0].tccDriver @tccDriver.setter def tccDriver(self, int tccDriver): self._pvt_ptr[0].tccDriver = tccDriver @property def asyncEngineCount(self): return self._pvt_ptr[0].asyncEngineCount @asyncEngineCount.setter def asyncEngineCount(self, int asyncEngineCount): self._pvt_ptr[0].asyncEngineCount = asyncEngineCount @property def unifiedAddressing(self): return self._pvt_ptr[0].unifiedAddressing @unifiedAddressing.setter def unifiedAddressing(self, int unifiedAddressing): self._pvt_ptr[0].unifiedAddressing = unifiedAddressing @property def memoryClockRate(self): return self._pvt_ptr[0].memoryClockRate @memoryClockRate.setter def memoryClockRate(self, int memoryClockRate): self._pvt_ptr[0].memoryClockRate = memoryClockRate @property def memoryBusWidth(self): return self._pvt_ptr[0].memoryBusWidth @memoryBusWidth.setter def memoryBusWidth(self, int memoryBusWidth): self._pvt_ptr[0].memoryBusWidth = memoryBusWidth @property def l2CacheSize(self): return self._pvt_ptr[0].l2CacheSize @l2CacheSize.setter def l2CacheSize(self, int l2CacheSize): self._pvt_ptr[0].l2CacheSize = l2CacheSize @property def persistingL2CacheMaxSize(self): return self._pvt_ptr[0].persistingL2CacheMaxSize @persistingL2CacheMaxSize.setter def persistingL2CacheMaxSize(self, int persistingL2CacheMaxSize): self._pvt_ptr[0].persistingL2CacheMaxSize = persistingL2CacheMaxSize @property def maxThreadsPerMultiProcessor(self): return self._pvt_ptr[0].maxThreadsPerMultiProcessor @maxThreadsPerMultiProcessor.setter def maxThreadsPerMultiProcessor(self, int maxThreadsPerMultiProcessor): self._pvt_ptr[0].maxThreadsPerMultiProcessor = maxThreadsPerMultiProcessor @property def streamPrioritiesSupported(self): return self._pvt_ptr[0].streamPrioritiesSupported @streamPrioritiesSupported.setter def streamPrioritiesSupported(self, int streamPrioritiesSupported): self._pvt_ptr[0].streamPrioritiesSupported = streamPrioritiesSupported @property def globalL1CacheSupported(self): return self._pvt_ptr[0].globalL1CacheSupported @globalL1CacheSupported.setter def globalL1CacheSupported(self, int globalL1CacheSupported): self._pvt_ptr[0].globalL1CacheSupported = globalL1CacheSupported @property def localL1CacheSupported(self): return self._pvt_ptr[0].localL1CacheSupported @localL1CacheSupported.setter def localL1CacheSupported(self, int localL1CacheSupported): self._pvt_ptr[0].localL1CacheSupported = localL1CacheSupported @property def sharedMemPerMultiprocessor(self): return self._pvt_ptr[0].sharedMemPerMultiprocessor @sharedMemPerMultiprocessor.setter def sharedMemPerMultiprocessor(self, size_t sharedMemPerMultiprocessor): self._pvt_ptr[0].sharedMemPerMultiprocessor = sharedMemPerMultiprocessor @property def regsPerMultiprocessor(self): return self._pvt_ptr[0].regsPerMultiprocessor @regsPerMultiprocessor.setter def regsPerMultiprocessor(self, int regsPerMultiprocessor): self._pvt_ptr[0].regsPerMultiprocessor = regsPerMultiprocessor @property def managedMemory(self): return self._pvt_ptr[0].managedMemory @managedMemory.setter def managedMemory(self, int managedMemory): self._pvt_ptr[0].managedMemory = managedMemory @property def isMultiGpuBoard(self): return self._pvt_ptr[0].isMultiGpuBoard @isMultiGpuBoard.setter def isMultiGpuBoard(self, int isMultiGpuBoard): self._pvt_ptr[0].isMultiGpuBoard = isMultiGpuBoard @property def multiGpuBoardGroupID(self): return self._pvt_ptr[0].multiGpuBoardGroupID @multiGpuBoardGroupID.setter def multiGpuBoardGroupID(self, int multiGpuBoardGroupID): self._pvt_ptr[0].multiGpuBoardGroupID = multiGpuBoardGroupID @property def hostNativeAtomicSupported(self): return self._pvt_ptr[0].hostNativeAtomicSupported @hostNativeAtomicSupported.setter def hostNativeAtomicSupported(self, int hostNativeAtomicSupported): self._pvt_ptr[0].hostNativeAtomicSupported = hostNativeAtomicSupported @property def singleToDoublePrecisionPerfRatio(self): return self._pvt_ptr[0].singleToDoublePrecisionPerfRatio @singleToDoublePrecisionPerfRatio.setter def singleToDoublePrecisionPerfRatio(self, int singleToDoublePrecisionPerfRatio): self._pvt_ptr[0].singleToDoublePrecisionPerfRatio = singleToDoublePrecisionPerfRatio @property def pageableMemoryAccess(self): return self._pvt_ptr[0].pageableMemoryAccess @pageableMemoryAccess.setter def pageableMemoryAccess(self, int pageableMemoryAccess): self._pvt_ptr[0].pageableMemoryAccess = pageableMemoryAccess @property def concurrentManagedAccess(self): return self._pvt_ptr[0].concurrentManagedAccess @concurrentManagedAccess.setter def concurrentManagedAccess(self, int concurrentManagedAccess): self._pvt_ptr[0].concurrentManagedAccess = concurrentManagedAccess @property def computePreemptionSupported(self): return self._pvt_ptr[0].computePreemptionSupported @computePreemptionSupported.setter def computePreemptionSupported(self, int computePreemptionSupported): self._pvt_ptr[0].computePreemptionSupported = computePreemptionSupported @property def canUseHostPointerForRegisteredMem(self): return self._pvt_ptr[0].canUseHostPointerForRegisteredMem @canUseHostPointerForRegisteredMem.setter def canUseHostPointerForRegisteredMem(self, int canUseHostPointerForRegisteredMem): self._pvt_ptr[0].canUseHostPointerForRegisteredMem = canUseHostPointerForRegisteredMem @property def cooperativeLaunch(self): return self._pvt_ptr[0].cooperativeLaunch @cooperativeLaunch.setter def cooperativeLaunch(self, int cooperativeLaunch): self._pvt_ptr[0].cooperativeLaunch = cooperativeLaunch @property def cooperativeMultiDeviceLaunch(self): return self._pvt_ptr[0].cooperativeMultiDeviceLaunch @cooperativeMultiDeviceLaunch.setter def cooperativeMultiDeviceLaunch(self, int cooperativeMultiDeviceLaunch): self._pvt_ptr[0].cooperativeMultiDeviceLaunch = cooperativeMultiDeviceLaunch @property def sharedMemPerBlockOptin(self): return self._pvt_ptr[0].sharedMemPerBlockOptin @sharedMemPerBlockOptin.setter def sharedMemPerBlockOptin(self, size_t sharedMemPerBlockOptin): self._pvt_ptr[0].sharedMemPerBlockOptin = sharedMemPerBlockOptin @property def pageableMemoryAccessUsesHostPageTables(self): return self._pvt_ptr[0].pageableMemoryAccessUsesHostPageTables @pageableMemoryAccessUsesHostPageTables.setter def pageableMemoryAccessUsesHostPageTables(self, int pageableMemoryAccessUsesHostPageTables): self._pvt_ptr[0].pageableMemoryAccessUsesHostPageTables = pageableMemoryAccessUsesHostPageTables @property def directManagedMemAccessFromHost(self): return self._pvt_ptr[0].directManagedMemAccessFromHost @directManagedMemAccessFromHost.setter def directManagedMemAccessFromHost(self, int directManagedMemAccessFromHost): self._pvt_ptr[0].directManagedMemAccessFromHost = directManagedMemAccessFromHost @property def maxBlocksPerMultiProcessor(self): return self._pvt_ptr[0].maxBlocksPerMultiProcessor @maxBlocksPerMultiProcessor.setter def maxBlocksPerMultiProcessor(self, int maxBlocksPerMultiProcessor): self._pvt_ptr[0].maxBlocksPerMultiProcessor = maxBlocksPerMultiProcessor @property def accessPolicyMaxWindowSize(self): return self._pvt_ptr[0].accessPolicyMaxWindowSize @accessPolicyMaxWindowSize.setter def accessPolicyMaxWindowSize(self, int accessPolicyMaxWindowSize): self._pvt_ptr[0].accessPolicyMaxWindowSize = accessPolicyMaxWindowSize @property def reservedSharedMemPerBlock(self): return self._pvt_ptr[0].reservedSharedMemPerBlock @reservedSharedMemPerBlock.setter def reservedSharedMemPerBlock(self, size_t reservedSharedMemPerBlock): self._pvt_ptr[0].reservedSharedMemPerBlock = reservedSharedMemPerBlock @property def hostRegisterSupported(self): return self._pvt_ptr[0].hostRegisterSupported @hostRegisterSupported.setter def hostRegisterSupported(self, int hostRegisterSupported): self._pvt_ptr[0].hostRegisterSupported = hostRegisterSupported @property def sparseCudaArraySupported(self): return self._pvt_ptr[0].sparseCudaArraySupported @sparseCudaArraySupported.setter def sparseCudaArraySupported(self, int sparseCudaArraySupported): self._pvt_ptr[0].sparseCudaArraySupported = sparseCudaArraySupported @property def hostRegisterReadOnlySupported(self): return self._pvt_ptr[0].hostRegisterReadOnlySupported @hostRegisterReadOnlySupported.setter def hostRegisterReadOnlySupported(self, int hostRegisterReadOnlySupported): self._pvt_ptr[0].hostRegisterReadOnlySupported = hostRegisterReadOnlySupported @property def timelineSemaphoreInteropSupported(self): return self._pvt_ptr[0].timelineSemaphoreInteropSupported @timelineSemaphoreInteropSupported.setter def timelineSemaphoreInteropSupported(self, int timelineSemaphoreInteropSupported): self._pvt_ptr[0].timelineSemaphoreInteropSupported = timelineSemaphoreInteropSupported @property def memoryPoolsSupported(self): return self._pvt_ptr[0].memoryPoolsSupported @memoryPoolsSupported.setter def memoryPoolsSupported(self, int memoryPoolsSupported): self._pvt_ptr[0].memoryPoolsSupported = memoryPoolsSupported @property def gpuDirectRDMASupported(self): return self._pvt_ptr[0].gpuDirectRDMASupported @gpuDirectRDMASupported.setter def gpuDirectRDMASupported(self, int gpuDirectRDMASupported): self._pvt_ptr[0].gpuDirectRDMASupported = gpuDirectRDMASupported @property def gpuDirectRDMAFlushWritesOptions(self): return self._pvt_ptr[0].gpuDirectRDMAFlushWritesOptions @gpuDirectRDMAFlushWritesOptions.setter def gpuDirectRDMAFlushWritesOptions(self, unsigned int gpuDirectRDMAFlushWritesOptions): self._pvt_ptr[0].gpuDirectRDMAFlushWritesOptions = gpuDirectRDMAFlushWritesOptions @property def gpuDirectRDMAWritesOrdering(self): return self._pvt_ptr[0].gpuDirectRDMAWritesOrdering @gpuDirectRDMAWritesOrdering.setter def gpuDirectRDMAWritesOrdering(self, int gpuDirectRDMAWritesOrdering): self._pvt_ptr[0].gpuDirectRDMAWritesOrdering = gpuDirectRDMAWritesOrdering @property def memoryPoolSupportedHandleTypes(self): return self._pvt_ptr[0].memoryPoolSupportedHandleTypes @memoryPoolSupportedHandleTypes.setter def memoryPoolSupportedHandleTypes(self, unsigned int memoryPoolSupportedHandleTypes): self._pvt_ptr[0].memoryPoolSupportedHandleTypes = memoryPoolSupportedHandleTypes @property def deferredMappingCudaArraySupported(self): return self._pvt_ptr[0].deferredMappingCudaArraySupported @deferredMappingCudaArraySupported.setter def deferredMappingCudaArraySupported(self, int deferredMappingCudaArraySupported): self._pvt_ptr[0].deferredMappingCudaArraySupported = deferredMappingCudaArraySupported @property def ipcEventSupported(self): return self._pvt_ptr[0].ipcEventSupported @ipcEventSupported.setter def ipcEventSupported(self, int ipcEventSupported): self._pvt_ptr[0].ipcEventSupported = ipcEventSupported @property def clusterLaunch(self): return self._pvt_ptr[0].clusterLaunch @clusterLaunch.setter def clusterLaunch(self, int clusterLaunch): self._pvt_ptr[0].clusterLaunch = clusterLaunch @property def unifiedFunctionPointers(self): return self._pvt_ptr[0].unifiedFunctionPointers @unifiedFunctionPointers.setter def unifiedFunctionPointers(self, int unifiedFunctionPointers): self._pvt_ptr[0].unifiedFunctionPointers = unifiedFunctionPointers @property def reserved(self): return self._pvt_ptr[0].reserved @reserved.setter def reserved(self, reserved): self._pvt_ptr[0].reserved = reserved cdef class cudaIpcEventHandle_st: """ CUDA IPC event handle Attributes ---------- reserved : bytes Methods ------- getPtr() Get memory address of class instance """ def __cinit__(self, void_ptr _ptr = 0): if _ptr == 0: self._pvt_ptr = &self._pvt_val else: self._pvt_ptr = _ptr def __init__(self, void_ptr _ptr = 0): pass def __dealloc__(self): pass def getPtr(self): return self._pvt_ptr def __repr__(self): if self._pvt_ptr is not NULL: str_list = [] try: str_list += ['reserved : ' + str(self.reserved)] except ValueError: str_list += ['reserved : '] return '\n'.join(str_list) else: return '' @property def reserved(self): return PyBytes_FromStringAndSize(self._pvt_ptr[0].reserved, 64) @reserved.setter def reserved(self, reserved): if len(reserved) != 64: raise ValueError("reserved length must be 64, is " + str(len(reserved))) if CHAR_MIN == 0: for i, b in enumerate(reserved): if b < 0 and b > -129: b = b + 256 self._pvt_ptr[0].reserved[i] = b else: for i, b in enumerate(reserved): if b > 127 and b < 256: b = b - 256 self._pvt_ptr[0].reserved[i] = b cdef class cudaIpcMemHandle_st: """ CUDA IPC memory handle Attributes ---------- reserved : bytes Methods ------- getPtr() Get memory address of class instance """ def __cinit__(self, void_ptr _ptr = 0): if _ptr == 0: self._pvt_ptr = &self._pvt_val else: self._pvt_ptr = _ptr def __init__(self, void_ptr _ptr = 0): pass def __dealloc__(self): pass def getPtr(self): return self._pvt_ptr def __repr__(self): if self._pvt_ptr is not NULL: str_list = [] try: str_list += ['reserved : ' + str(self.reserved)] except ValueError: str_list += ['reserved : '] return '\n'.join(str_list) else: return '' @property def reserved(self): return PyBytes_FromStringAndSize(self._pvt_ptr[0].reserved, 64) @reserved.setter def reserved(self, reserved): if len(reserved) != 64: raise ValueError("reserved length must be 64, is " + str(len(reserved))) if CHAR_MIN == 0: for i, b in enumerate(reserved): if b < 0 and b > -129: b = b + 256 self._pvt_ptr[0].reserved[i] = b else: for i, b in enumerate(reserved): if b > 127 and b < 256: b = b - 256 self._pvt_ptr[0].reserved[i] = b cdef class cudaMemFabricHandle_st: """ Attributes ---------- reserved : bytes Methods ------- getPtr() Get memory address of class instance """ def __cinit__(self, void_ptr _ptr = 0): if _ptr == 0: self._pvt_ptr = &self._pvt_val else: self._pvt_ptr = _ptr def __init__(self, void_ptr _ptr = 0): pass def __dealloc__(self): pass def getPtr(self): return self._pvt_ptr def __repr__(self): if self._pvt_ptr is not NULL: str_list = [] try: str_list += ['reserved : ' + str(self.reserved)] except ValueError: str_list += ['reserved : '] return '\n'.join(str_list) else: return '' @property def reserved(self): return PyBytes_FromStringAndSize(self._pvt_ptr[0].reserved, 64) @reserved.setter def reserved(self, reserved): if len(reserved) != 64: raise ValueError("reserved length must be 64, is " + str(len(reserved))) if CHAR_MIN == 0: for i, b in enumerate(reserved): if b < 0 and b > -129: b = b + 256 self._pvt_ptr[0].reserved[i] = b else: for i, b in enumerate(reserved): if b > 127 and b < 256: b = b - 256 self._pvt_ptr[0].reserved[i] = b cdef class anon_struct7: """ Attributes ---------- handle : Any name : Any Methods ------- getPtr() Get memory address of class instance """ def __cinit__(self, void_ptr _ptr): self._pvt_ptr = _ptr def __init__(self, void_ptr _ptr): pass def __dealloc__(self): pass def getPtr(self): return &self._pvt_ptr[0].handle.win32 def __repr__(self): if self._pvt_ptr is not NULL: str_list = [] try: str_list += ['handle : ' + hex(self.handle)] except ValueError: str_list += ['handle : '] try: str_list += ['name : ' + hex(self.name)] except ValueError: str_list += ['name : '] return '\n'.join(str_list) else: return '' @property def handle(self): return self._pvt_ptr[0].handle.win32.handle @handle.setter def handle(self, handle): _chandle = _HelperInputVoidPtr(handle) self._pvt_ptr[0].handle.win32.handle = _chandle.cptr @property def name(self): return self._pvt_ptr[0].handle.win32.name @name.setter def name(self, name): _cname = _HelperInputVoidPtr(name) self._pvt_ptr[0].handle.win32.name = _cname.cptr cdef class anon_union2: """ Attributes ---------- fd : int win32 : anon_struct7 nvSciBufObject : Any Methods ------- getPtr() Get memory address of class instance """ def __cinit__(self, void_ptr _ptr): self._pvt_ptr = _ptr def __init__(self, void_ptr _ptr): pass self._win32 = anon_struct7(_ptr=self._pvt_ptr) def __dealloc__(self): pass def getPtr(self): return &self._pvt_ptr[0].handle def __repr__(self): if self._pvt_ptr is not NULL: str_list = [] try: str_list += ['fd : ' + str(self.fd)] except ValueError: str_list += ['fd : '] try: str_list += ['win32 :\n' + '\n'.join([' ' + line for line in str(self.win32).splitlines()])] except ValueError: str_list += ['win32 : '] try: str_list += ['nvSciBufObject : ' + hex(self.nvSciBufObject)] except ValueError: str_list += ['nvSciBufObject : '] return '\n'.join(str_list) else: return '' @property def fd(self): return self._pvt_ptr[0].handle.fd @fd.setter def fd(self, int fd): self._pvt_ptr[0].handle.fd = fd @property def win32(self): return self._win32 @win32.setter def win32(self, win32 not None : anon_struct7): string.memcpy(&self._pvt_ptr[0].handle.win32, win32.getPtr(), sizeof(self._pvt_ptr[0].handle.win32)) @property def nvSciBufObject(self): return self._pvt_ptr[0].handle.nvSciBufObject @nvSciBufObject.setter def nvSciBufObject(self, nvSciBufObject): _cnvSciBufObject = _HelperInputVoidPtr(nvSciBufObject) self._pvt_ptr[0].handle.nvSciBufObject = _cnvSciBufObject.cptr cdef class cudaExternalMemoryHandleDesc: """ External memory handle descriptor Attributes ---------- type : cudaExternalMemoryHandleType Type of the handle handle : anon_union2 size : unsigned long long Size of the memory allocation flags : unsigned int Flags must either be zero or cudaExternalMemoryDedicated Methods ------- getPtr() Get memory address of class instance """ def __cinit__(self, void_ptr _ptr = 0): if _ptr == 0: self._val_ptr = calloc(1, sizeof(cyruntime.cudaExternalMemoryHandleDesc)) self._pvt_ptr = self._val_ptr else: self._pvt_ptr = _ptr def __init__(self, void_ptr _ptr = 0): pass self._handle = anon_union2(_ptr=self._pvt_ptr) def __dealloc__(self): if self._val_ptr is not NULL: free(self._val_ptr) def getPtr(self): return self._pvt_ptr def __repr__(self): if self._pvt_ptr is not NULL: str_list = [] try: str_list += ['type : ' + str(self.type)] except ValueError: str_list += ['type : '] try: str_list += ['handle :\n' + '\n'.join([' ' + line for line in str(self.handle).splitlines()])] except ValueError: str_list += ['handle : '] try: str_list += ['size : ' + str(self.size)] except ValueError: str_list += ['size : '] try: str_list += ['flags : ' + str(self.flags)] except ValueError: str_list += ['flags : '] return '\n'.join(str_list) else: return '' @property def type(self): if self._pvt_ptr[0].type not in _dict_cudaExternalMemoryHandleType: return None return _dict_cudaExternalMemoryHandleType[self._pvt_ptr[0].type] @type.setter def type(self, type not None : cudaExternalMemoryHandleType): self._pvt_ptr[0].type = type.value @property def handle(self): return self._handle @handle.setter def handle(self, handle not None : anon_union2): string.memcpy(&self._pvt_ptr[0].handle, handle.getPtr(), sizeof(self._pvt_ptr[0].handle)) @property def size(self): return self._pvt_ptr[0].size @size.setter def size(self, unsigned long long size): self._pvt_ptr[0].size = size @property def flags(self): return self._pvt_ptr[0].flags @flags.setter def flags(self, unsigned int flags): self._pvt_ptr[0].flags = flags cdef class cudaExternalMemoryBufferDesc: """ External memory buffer descriptor Attributes ---------- offset : unsigned long long Offset into the memory object where the buffer's base is size : unsigned long long Size of the buffer flags : unsigned int Flags reserved for future use. Must be zero. Methods ------- getPtr() Get memory address of class instance """ def __cinit__(self, void_ptr _ptr = 0): if _ptr == 0: self._pvt_ptr = &self._pvt_val else: self._pvt_ptr = _ptr def __init__(self, void_ptr _ptr = 0): pass def __dealloc__(self): pass def getPtr(self): return self._pvt_ptr def __repr__(self): if self._pvt_ptr is not NULL: str_list = [] try: str_list += ['offset : ' + str(self.offset)] except ValueError: str_list += ['offset : '] try: str_list += ['size : ' + str(self.size)] except ValueError: str_list += ['size : '] try: str_list += ['flags : ' + str(self.flags)] except ValueError: str_list += ['flags : '] return '\n'.join(str_list) else: return '' @property def offset(self): return self._pvt_ptr[0].offset @offset.setter def offset(self, unsigned long long offset): self._pvt_ptr[0].offset = offset @property def size(self): return self._pvt_ptr[0].size @size.setter def size(self, unsigned long long size): self._pvt_ptr[0].size = size @property def flags(self): return self._pvt_ptr[0].flags @flags.setter def flags(self, unsigned int flags): self._pvt_ptr[0].flags = flags cdef class cudaExternalMemoryMipmappedArrayDesc: """ External memory mipmap descriptor Attributes ---------- offset : unsigned long long Offset into the memory object where the base level of the mipmap chain is. formatDesc : cudaChannelFormatDesc Format of base level of the mipmap chain extent : cudaExtent Dimensions of base level of the mipmap chain flags : unsigned int Flags associated with CUDA mipmapped arrays. See cudaMallocMipmappedArray numLevels : unsigned int Total number of levels in the mipmap chain Methods ------- getPtr() Get memory address of class instance """ def __cinit__(self, void_ptr _ptr = 0): if _ptr == 0: self._pvt_ptr = &self._pvt_val else: self._pvt_ptr = _ptr def __init__(self, void_ptr _ptr = 0): pass self._formatDesc = cudaChannelFormatDesc(_ptr=&self._pvt_ptr[0].formatDesc) self._extent = cudaExtent(_ptr=&self._pvt_ptr[0].extent) def __dealloc__(self): pass def getPtr(self): return self._pvt_ptr def __repr__(self): if self._pvt_ptr is not NULL: str_list = [] try: str_list += ['offset : ' + str(self.offset)] except ValueError: str_list += ['offset : '] try: str_list += ['formatDesc :\n' + '\n'.join([' ' + line for line in str(self.formatDesc).splitlines()])] except ValueError: str_list += ['formatDesc : '] try: str_list += ['extent :\n' + '\n'.join([' ' + line for line in str(self.extent).splitlines()])] except ValueError: str_list += ['extent : '] try: str_list += ['flags : ' + str(self.flags)] except ValueError: str_list += ['flags : '] try: str_list += ['numLevels : ' + str(self.numLevels)] except ValueError: str_list += ['numLevels : '] return '\n'.join(str_list) else: return '' @property def offset(self): return self._pvt_ptr[0].offset @offset.setter def offset(self, unsigned long long offset): self._pvt_ptr[0].offset = offset @property def formatDesc(self): return self._formatDesc @formatDesc.setter def formatDesc(self, formatDesc not None : cudaChannelFormatDesc): string.memcpy(&self._pvt_ptr[0].formatDesc, formatDesc.getPtr(), sizeof(self._pvt_ptr[0].formatDesc)) @property def extent(self): return self._extent @extent.setter def extent(self, extent not None : cudaExtent): string.memcpy(&self._pvt_ptr[0].extent, extent.getPtr(), sizeof(self._pvt_ptr[0].extent)) @property def flags(self): return self._pvt_ptr[0].flags @flags.setter def flags(self, unsigned int flags): self._pvt_ptr[0].flags = flags @property def numLevels(self): return self._pvt_ptr[0].numLevels @numLevels.setter def numLevels(self, unsigned int numLevels): self._pvt_ptr[0].numLevels = numLevels cdef class anon_struct8: """ Attributes ---------- handle : Any name : Any Methods ------- getPtr() Get memory address of class instance """ def __cinit__(self, void_ptr _ptr): self._pvt_ptr = _ptr def __init__(self, void_ptr _ptr): pass def __dealloc__(self): pass def getPtr(self): return &self._pvt_ptr[0].handle.win32 def __repr__(self): if self._pvt_ptr is not NULL: str_list = [] try: str_list += ['handle : ' + hex(self.handle)] except ValueError: str_list += ['handle : '] try: str_list += ['name : ' + hex(self.name)] except ValueError: str_list += ['name : '] return '\n'.join(str_list) else: return '' @property def handle(self): return self._pvt_ptr[0].handle.win32.handle @handle.setter def handle(self, handle): _chandle = _HelperInputVoidPtr(handle) self._pvt_ptr[0].handle.win32.handle = _chandle.cptr @property def name(self): return self._pvt_ptr[0].handle.win32.name @name.setter def name(self, name): _cname = _HelperInputVoidPtr(name) self._pvt_ptr[0].handle.win32.name = _cname.cptr cdef class anon_union3: """ Attributes ---------- fd : int win32 : anon_struct8 nvSciSyncObj : Any Methods ------- getPtr() Get memory address of class instance """ def __cinit__(self, void_ptr _ptr): self._pvt_ptr = _ptr def __init__(self, void_ptr _ptr): pass self._win32 = anon_struct8(_ptr=self._pvt_ptr) def __dealloc__(self): pass def getPtr(self): return &self._pvt_ptr[0].handle def __repr__(self): if self._pvt_ptr is not NULL: str_list = [] try: str_list += ['fd : ' + str(self.fd)] except ValueError: str_list += ['fd : '] try: str_list += ['win32 :\n' + '\n'.join([' ' + line for line in str(self.win32).splitlines()])] except ValueError: str_list += ['win32 : '] try: str_list += ['nvSciSyncObj : ' + hex(self.nvSciSyncObj)] except ValueError: str_list += ['nvSciSyncObj : '] return '\n'.join(str_list) else: return '' @property def fd(self): return self._pvt_ptr[0].handle.fd @fd.setter def fd(self, int fd): self._pvt_ptr[0].handle.fd = fd @property def win32(self): return self._win32 @win32.setter def win32(self, win32 not None : anon_struct8): string.memcpy(&self._pvt_ptr[0].handle.win32, win32.getPtr(), sizeof(self._pvt_ptr[0].handle.win32)) @property def nvSciSyncObj(self): return self._pvt_ptr[0].handle.nvSciSyncObj @nvSciSyncObj.setter def nvSciSyncObj(self, nvSciSyncObj): _cnvSciSyncObj = _HelperInputVoidPtr(nvSciSyncObj) self._pvt_ptr[0].handle.nvSciSyncObj = _cnvSciSyncObj.cptr cdef class cudaExternalSemaphoreHandleDesc: """ External semaphore handle descriptor Attributes ---------- type : cudaExternalSemaphoreHandleType Type of the handle handle : anon_union3 flags : unsigned int Flags reserved for the future. Must be zero. Methods ------- getPtr() Get memory address of class instance """ def __cinit__(self, void_ptr _ptr = 0): if _ptr == 0: self._val_ptr = calloc(1, sizeof(cyruntime.cudaExternalSemaphoreHandleDesc)) self._pvt_ptr = self._val_ptr else: self._pvt_ptr = _ptr def __init__(self, void_ptr _ptr = 0): pass self._handle = anon_union3(_ptr=self._pvt_ptr) def __dealloc__(self): if self._val_ptr is not NULL: free(self._val_ptr) def getPtr(self): return self._pvt_ptr def __repr__(self): if self._pvt_ptr is not NULL: str_list = [] try: str_list += ['type : ' + str(self.type)] except ValueError: str_list += ['type : '] try: str_list += ['handle :\n' + '\n'.join([' ' + line for line in str(self.handle).splitlines()])] except ValueError: str_list += ['handle : '] try: str_list += ['flags : ' + str(self.flags)] except ValueError: str_list += ['flags : '] return '\n'.join(str_list) else: return '' @property def type(self): if self._pvt_ptr[0].type not in _dict_cudaExternalSemaphoreHandleType: return None return _dict_cudaExternalSemaphoreHandleType[self._pvt_ptr[0].type] @type.setter def type(self, type not None : cudaExternalSemaphoreHandleType): self._pvt_ptr[0].type = type.value @property def handle(self): return self._handle @handle.setter def handle(self, handle not None : anon_union3): string.memcpy(&self._pvt_ptr[0].handle, handle.getPtr(), sizeof(self._pvt_ptr[0].handle)) @property def flags(self): return self._pvt_ptr[0].flags @flags.setter def flags(self, unsigned int flags): self._pvt_ptr[0].flags = flags cdef class anon_struct15: """ Attributes ---------- value : unsigned long long Methods ------- getPtr() Get memory address of class instance """ def __cinit__(self, void_ptr _ptr): self._pvt_ptr = _ptr def __init__(self, void_ptr _ptr): pass def __dealloc__(self): pass def getPtr(self): return &self._pvt_ptr[0].params.fence def __repr__(self): if self._pvt_ptr is not NULL: str_list = [] try: str_list += ['value : ' + str(self.value)] except ValueError: str_list += ['value : '] return '\n'.join(str_list) else: return '' @property def value(self): return self._pvt_ptr[0].params.fence.value @value.setter def value(self, unsigned long long value): self._pvt_ptr[0].params.fence.value = value cdef class anon_union6: """ Attributes ---------- fence : Any reserved : unsigned long long Methods ------- getPtr() Get memory address of class instance """ def __cinit__(self, void_ptr _ptr): self._pvt_ptr = _ptr def __init__(self, void_ptr _ptr): pass def __dealloc__(self): pass def getPtr(self): return &self._pvt_ptr[0].params.nvSciSync def __repr__(self): if self._pvt_ptr is not NULL: str_list = [] try: str_list += ['fence : ' + hex(self.fence)] except ValueError: str_list += ['fence : '] try: str_list += ['reserved : ' + str(self.reserved)] except ValueError: str_list += ['reserved : '] return '\n'.join(str_list) else: return '' @property def fence(self): return self._pvt_ptr[0].params.nvSciSync.fence @fence.setter def fence(self, fence): _cfence = _HelperInputVoidPtr(fence) self._pvt_ptr[0].params.nvSciSync.fence = _cfence.cptr @property def reserved(self): return self._pvt_ptr[0].params.nvSciSync.reserved @reserved.setter def reserved(self, unsigned long long reserved): self._pvt_ptr[0].params.nvSciSync.reserved = reserved cdef class anon_struct16: """ Attributes ---------- key : unsigned long long Methods ------- getPtr() Get memory address of class instance """ def __cinit__(self, void_ptr _ptr): self._pvt_ptr = _ptr def __init__(self, void_ptr _ptr): pass def __dealloc__(self): pass def getPtr(self): return &self._pvt_ptr[0].params.keyedMutex def __repr__(self): if self._pvt_ptr is not NULL: str_list = [] try: str_list += ['key : ' + str(self.key)] except ValueError: str_list += ['key : '] return '\n'.join(str_list) else: return '' @property def key(self): return self._pvt_ptr[0].params.keyedMutex.key @key.setter def key(self, unsigned long long key): self._pvt_ptr[0].params.keyedMutex.key = key cdef class anon_struct17: """ Attributes ---------- fence : anon_struct15 nvSciSync : anon_union6 keyedMutex : anon_struct16 reserved : list[unsigned int] Methods ------- getPtr() Get memory address of class instance """ def __cinit__(self, void_ptr _ptr): self._pvt_ptr = _ptr def __init__(self, void_ptr _ptr): pass self._fence = anon_struct15(_ptr=self._pvt_ptr) self._nvSciSync = anon_union6(_ptr=self._pvt_ptr) self._keyedMutex = anon_struct16(_ptr=self._pvt_ptr) def __dealloc__(self): pass def getPtr(self): return &self._pvt_ptr[0].params def __repr__(self): if self._pvt_ptr is not NULL: str_list = [] try: str_list += ['fence :\n' + '\n'.join([' ' + line for line in str(self.fence).splitlines()])] except ValueError: str_list += ['fence : '] try: str_list += ['nvSciSync :\n' + '\n'.join([' ' + line for line in str(self.nvSciSync).splitlines()])] except ValueError: str_list += ['nvSciSync : '] try: str_list += ['keyedMutex :\n' + '\n'.join([' ' + line for line in str(self.keyedMutex).splitlines()])] except ValueError: str_list += ['keyedMutex : '] try: str_list += ['reserved : ' + str(self.reserved)] except ValueError: str_list += ['reserved : '] return '\n'.join(str_list) else: return '' @property def fence(self): return self._fence @fence.setter def fence(self, fence not None : anon_struct15): string.memcpy(&self._pvt_ptr[0].params.fence, fence.getPtr(), sizeof(self._pvt_ptr[0].params.fence)) @property def nvSciSync(self): return self._nvSciSync @nvSciSync.setter def nvSciSync(self, nvSciSync not None : anon_union6): string.memcpy(&self._pvt_ptr[0].params.nvSciSync, nvSciSync.getPtr(), sizeof(self._pvt_ptr[0].params.nvSciSync)) @property def keyedMutex(self): return self._keyedMutex @keyedMutex.setter def keyedMutex(self, keyedMutex not None : anon_struct16): string.memcpy(&self._pvt_ptr[0].params.keyedMutex, keyedMutex.getPtr(), sizeof(self._pvt_ptr[0].params.keyedMutex)) @property def reserved(self): return self._pvt_ptr[0].params.reserved @reserved.setter def reserved(self, reserved): self._pvt_ptr[0].params.reserved = reserved cdef class cudaExternalSemaphoreSignalParams: """ External semaphore signal parameters, compatible with driver type Attributes ---------- params : anon_struct17 flags : unsigned int Only when cudaExternalSemaphoreSignalParams is used to signal a cudaExternalSemaphore_t of type cudaExternalSemaphoreHandleTypeNvSciSync, the valid flag is cudaExternalSemaphoreSignalSkipNvSciBufMemSync: which indicates that while signaling the cudaExternalSemaphore_t, no memory synchronization operations should be performed for any external memory object imported as cudaExternalMemoryHandleTypeNvSciBuf. For all other types of cudaExternalSemaphore_t, flags must be zero. reserved : list[unsigned int] Methods ------- getPtr() Get memory address of class instance """ def __cinit__(self, void_ptr _ptr = 0): if _ptr == 0: self._pvt_ptr = &self._pvt_val else: self._pvt_ptr = _ptr def __init__(self, void_ptr _ptr = 0): pass self._params = anon_struct17(_ptr=self._pvt_ptr) def __dealloc__(self): pass def getPtr(self): return self._pvt_ptr def __repr__(self): if self._pvt_ptr is not NULL: str_list = [] try: str_list += ['params :\n' + '\n'.join([' ' + line for line in str(self.params).splitlines()])] except ValueError: str_list += ['params : '] try: str_list += ['flags : ' + str(self.flags)] except ValueError: str_list += ['flags : '] try: str_list += ['reserved : ' + str(self.reserved)] except ValueError: str_list += ['reserved : '] return '\n'.join(str_list) else: return '' @property def params(self): return self._params @params.setter def params(self, params not None : anon_struct17): string.memcpy(&self._pvt_ptr[0].params, params.getPtr(), sizeof(self._pvt_ptr[0].params)) @property def flags(self): return self._pvt_ptr[0].flags @flags.setter def flags(self, unsigned int flags): self._pvt_ptr[0].flags = flags @property def reserved(self): return self._pvt_ptr[0].reserved @reserved.setter def reserved(self, reserved): self._pvt_ptr[0].reserved = reserved cdef class anon_struct18: """ Attributes ---------- value : unsigned long long Methods ------- getPtr() Get memory address of class instance """ def __cinit__(self, void_ptr _ptr): self._pvt_ptr = _ptr def __init__(self, void_ptr _ptr): pass def __dealloc__(self): pass def getPtr(self): return &self._pvt_ptr[0].params.fence def __repr__(self): if self._pvt_ptr is not NULL: str_list = [] try: str_list += ['value : ' + str(self.value)] except ValueError: str_list += ['value : '] return '\n'.join(str_list) else: return '' @property def value(self): return self._pvt_ptr[0].params.fence.value @value.setter def value(self, unsigned long long value): self._pvt_ptr[0].params.fence.value = value cdef class anon_union7: """ Attributes ---------- fence : Any reserved : unsigned long long Methods ------- getPtr() Get memory address of class instance """ def __cinit__(self, void_ptr _ptr): self._pvt_ptr = _ptr def __init__(self, void_ptr _ptr): pass def __dealloc__(self): pass def getPtr(self): return &self._pvt_ptr[0].params.nvSciSync def __repr__(self): if self._pvt_ptr is not NULL: str_list = [] try: str_list += ['fence : ' + hex(self.fence)] except ValueError: str_list += ['fence : '] try: str_list += ['reserved : ' + str(self.reserved)] except ValueError: str_list += ['reserved : '] return '\n'.join(str_list) else: return '' @property def fence(self): return self._pvt_ptr[0].params.nvSciSync.fence @fence.setter def fence(self, fence): _cfence = _HelperInputVoidPtr(fence) self._pvt_ptr[0].params.nvSciSync.fence = _cfence.cptr @property def reserved(self): return self._pvt_ptr[0].params.nvSciSync.reserved @reserved.setter def reserved(self, unsigned long long reserved): self._pvt_ptr[0].params.nvSciSync.reserved = reserved cdef class anon_struct19: """ Attributes ---------- key : unsigned long long timeoutMs : unsigned int Methods ------- getPtr() Get memory address of class instance """ def __cinit__(self, void_ptr _ptr): self._pvt_ptr = _ptr def __init__(self, void_ptr _ptr): pass def __dealloc__(self): pass def getPtr(self): return &self._pvt_ptr[0].params.keyedMutex def __repr__(self): if self._pvt_ptr is not NULL: str_list = [] try: str_list += ['key : ' + str(self.key)] except ValueError: str_list += ['key : '] try: str_list += ['timeoutMs : ' + str(self.timeoutMs)] except ValueError: str_list += ['timeoutMs : '] return '\n'.join(str_list) else: return '' @property def key(self): return self._pvt_ptr[0].params.keyedMutex.key @key.setter def key(self, unsigned long long key): self._pvt_ptr[0].params.keyedMutex.key = key @property def timeoutMs(self): return self._pvt_ptr[0].params.keyedMutex.timeoutMs @timeoutMs.setter def timeoutMs(self, unsigned int timeoutMs): self._pvt_ptr[0].params.keyedMutex.timeoutMs = timeoutMs cdef class anon_struct20: """ Attributes ---------- fence : anon_struct18 nvSciSync : anon_union7 keyedMutex : anon_struct19 reserved : list[unsigned int] Methods ------- getPtr() Get memory address of class instance """ def __cinit__(self, void_ptr _ptr): self._pvt_ptr = _ptr def __init__(self, void_ptr _ptr): pass self._fence = anon_struct18(_ptr=self._pvt_ptr) self._nvSciSync = anon_union7(_ptr=self._pvt_ptr) self._keyedMutex = anon_struct19(_ptr=self._pvt_ptr) def __dealloc__(self): pass def getPtr(self): return &self._pvt_ptr[0].params def __repr__(self): if self._pvt_ptr is not NULL: str_list = [] try: str_list += ['fence :\n' + '\n'.join([' ' + line for line in str(self.fence).splitlines()])] except ValueError: str_list += ['fence : '] try: str_list += ['nvSciSync :\n' + '\n'.join([' ' + line for line in str(self.nvSciSync).splitlines()])] except ValueError: str_list += ['nvSciSync : '] try: str_list += ['keyedMutex :\n' + '\n'.join([' ' + line for line in str(self.keyedMutex).splitlines()])] except ValueError: str_list += ['keyedMutex : '] try: str_list += ['reserved : ' + str(self.reserved)] except ValueError: str_list += ['reserved : '] return '\n'.join(str_list) else: return '' @property def fence(self): return self._fence @fence.setter def fence(self, fence not None : anon_struct18): string.memcpy(&self._pvt_ptr[0].params.fence, fence.getPtr(), sizeof(self._pvt_ptr[0].params.fence)) @property def nvSciSync(self): return self._nvSciSync @nvSciSync.setter def nvSciSync(self, nvSciSync not None : anon_union7): string.memcpy(&self._pvt_ptr[0].params.nvSciSync, nvSciSync.getPtr(), sizeof(self._pvt_ptr[0].params.nvSciSync)) @property def keyedMutex(self): return self._keyedMutex @keyedMutex.setter def keyedMutex(self, keyedMutex not None : anon_struct19): string.memcpy(&self._pvt_ptr[0].params.keyedMutex, keyedMutex.getPtr(), sizeof(self._pvt_ptr[0].params.keyedMutex)) @property def reserved(self): return self._pvt_ptr[0].params.reserved @reserved.setter def reserved(self, reserved): self._pvt_ptr[0].params.reserved = reserved cdef class cudaExternalSemaphoreWaitParams: """ External semaphore wait parameters, compatible with driver type Attributes ---------- params : anon_struct20 flags : unsigned int Only when cudaExternalSemaphoreSignalParams is used to signal a cudaExternalSemaphore_t of type cudaExternalSemaphoreHandleTypeNvSciSync, the valid flag is cudaExternalSemaphoreSignalSkipNvSciBufMemSync: which indicates that while waiting for the cudaExternalSemaphore_t, no memory synchronization operations should be performed for any external memory object imported as cudaExternalMemoryHandleTypeNvSciBuf. For all other types of cudaExternalSemaphore_t, flags must be zero. reserved : list[unsigned int] Methods ------- getPtr() Get memory address of class instance """ def __cinit__(self, void_ptr _ptr = 0): if _ptr == 0: self._pvt_ptr = &self._pvt_val else: self._pvt_ptr = _ptr def __init__(self, void_ptr _ptr = 0): pass self._params = anon_struct20(_ptr=self._pvt_ptr) def __dealloc__(self): pass def getPtr(self): return self._pvt_ptr def __repr__(self): if self._pvt_ptr is not NULL: str_list = [] try: str_list += ['params :\n' + '\n'.join([' ' + line for line in str(self.params).splitlines()])] except ValueError: str_list += ['params : '] try: str_list += ['flags : ' + str(self.flags)] except ValueError: str_list += ['flags : '] try: str_list += ['reserved : ' + str(self.reserved)] except ValueError: str_list += ['reserved : '] return '\n'.join(str_list) else: return '' @property def params(self): return self._params @params.setter def params(self, params not None : anon_struct20): string.memcpy(&self._pvt_ptr[0].params, params.getPtr(), sizeof(self._pvt_ptr[0].params)) @property def flags(self): return self._pvt_ptr[0].flags @flags.setter def flags(self, unsigned int flags): self._pvt_ptr[0].flags = flags @property def reserved(self): return self._pvt_ptr[0].reserved @reserved.setter def reserved(self, reserved): self._pvt_ptr[0].reserved = reserved cdef class cudalibraryHostUniversalFunctionAndDataTable: """ Attributes ---------- functionTable : Any functionWindowSize : size_t dataTable : Any dataWindowSize : size_t Methods ------- getPtr() Get memory address of class instance """ def __cinit__(self, void_ptr _ptr = 0): if _ptr == 0: self._pvt_ptr = &self._pvt_val else: self._pvt_ptr = _ptr def __init__(self, void_ptr _ptr = 0): pass def __dealloc__(self): pass def getPtr(self): return self._pvt_ptr def __repr__(self): if self._pvt_ptr is not NULL: str_list = [] try: str_list += ['functionTable : ' + hex(self.functionTable)] except ValueError: str_list += ['functionTable : '] try: str_list += ['functionWindowSize : ' + str(self.functionWindowSize)] except ValueError: str_list += ['functionWindowSize : '] try: str_list += ['dataTable : ' + hex(self.dataTable)] except ValueError: str_list += ['dataTable : '] try: str_list += ['dataWindowSize : ' + str(self.dataWindowSize)] except ValueError: str_list += ['dataWindowSize : '] return '\n'.join(str_list) else: return '' @property def functionTable(self): return self._pvt_ptr[0].functionTable @functionTable.setter def functionTable(self, functionTable): _cfunctionTable = _HelperInputVoidPtr(functionTable) self._pvt_ptr[0].functionTable = _cfunctionTable.cptr @property def functionWindowSize(self): return self._pvt_ptr[0].functionWindowSize @functionWindowSize.setter def functionWindowSize(self, size_t functionWindowSize): self._pvt_ptr[0].functionWindowSize = functionWindowSize @property def dataTable(self): return self._pvt_ptr[0].dataTable @dataTable.setter def dataTable(self, dataTable): _cdataTable = _HelperInputVoidPtr(dataTable) self._pvt_ptr[0].dataTable = _cdataTable.cptr @property def dataWindowSize(self): return self._pvt_ptr[0].dataWindowSize @dataWindowSize.setter def dataWindowSize(self, size_t dataWindowSize): self._pvt_ptr[0].dataWindowSize = dataWindowSize cdef class cudaKernelNodeParams: """ CUDA GPU kernel node parameters Attributes ---------- func : Any Kernel to launch gridDim : dim3 Grid dimensions blockDim : dim3 Block dimensions sharedMemBytes : unsigned int Dynamic shared-memory size per thread block in bytes kernelParams : Any Array of pointers to individual kernel arguments extra : Any Pointer to kernel arguments in the "extra" format Methods ------- getPtr() Get memory address of class instance """ def __cinit__(self, void_ptr _ptr = 0): if _ptr == 0: self._pvt_ptr = &self._pvt_val else: self._pvt_ptr = _ptr def __init__(self, void_ptr _ptr = 0): pass self._gridDim = dim3(_ptr=&self._pvt_ptr[0].gridDim) self._blockDim = dim3(_ptr=&self._pvt_ptr[0].blockDim) def __dealloc__(self): pass def getPtr(self): return self._pvt_ptr def __repr__(self): if self._pvt_ptr is not NULL: str_list = [] try: str_list += ['func : ' + hex(self.func)] except ValueError: str_list += ['func : '] try: str_list += ['gridDim :\n' + '\n'.join([' ' + line for line in str(self.gridDim).splitlines()])] except ValueError: str_list += ['gridDim : '] try: str_list += ['blockDim :\n' + '\n'.join([' ' + line for line in str(self.blockDim).splitlines()])] except ValueError: str_list += ['blockDim : '] try: str_list += ['sharedMemBytes : ' + str(self.sharedMemBytes)] except ValueError: str_list += ['sharedMemBytes : '] try: str_list += ['kernelParams : ' + str(self.kernelParams)] except ValueError: str_list += ['kernelParams : '] try: str_list += ['extra : ' + str(self.extra)] except ValueError: str_list += ['extra : '] return '\n'.join(str_list) else: return '' @property def func(self): return self._pvt_ptr[0].func @func.setter def func(self, func): _cfunc = _HelperInputVoidPtr(func) self._pvt_ptr[0].func = _cfunc.cptr @property def gridDim(self): return self._gridDim @gridDim.setter def gridDim(self, gridDim not None : dim3): string.memcpy(&self._pvt_ptr[0].gridDim, gridDim.getPtr(), sizeof(self._pvt_ptr[0].gridDim)) @property def blockDim(self): return self._blockDim @blockDim.setter def blockDim(self, blockDim not None : dim3): string.memcpy(&self._pvt_ptr[0].blockDim, blockDim.getPtr(), sizeof(self._pvt_ptr[0].blockDim)) @property def sharedMemBytes(self): return self._pvt_ptr[0].sharedMemBytes @sharedMemBytes.setter def sharedMemBytes(self, unsigned int sharedMemBytes): self._pvt_ptr[0].sharedMemBytes = sharedMemBytes @property def kernelParams(self): return self._pvt_ptr[0].kernelParams @kernelParams.setter def kernelParams(self, kernelParams): self._cykernelParams = _HelperKernelParams(kernelParams) self._pvt_ptr[0].kernelParams = self._cykernelParams.ckernelParams @property def extra(self): return self._pvt_ptr[0].extra @extra.setter def extra(self, void_ptr extra): self._pvt_ptr[0].extra = extra cdef class cudaKernelNodeParamsV2: """ CUDA GPU kernel node parameters Attributes ---------- func : Any Kernel to launch gridDim : dim3 Grid dimensions blockDim : dim3 Block dimensions sharedMemBytes : unsigned int Dynamic shared-memory size per thread block in bytes kernelParams : Any Array of pointers to individual kernel arguments extra : Any Pointer to kernel arguments in the "extra" format Methods ------- getPtr() Get memory address of class instance """ def __cinit__(self, void_ptr _ptr = 0): if _ptr == 0: self._pvt_ptr = &self._pvt_val else: self._pvt_ptr = _ptr def __init__(self, void_ptr _ptr = 0): pass self._gridDim = dim3(_ptr=&self._pvt_ptr[0].gridDim) self._blockDim = dim3(_ptr=&self._pvt_ptr[0].blockDim) def __dealloc__(self): pass def getPtr(self): return self._pvt_ptr def __repr__(self): if self._pvt_ptr is not NULL: str_list = [] try: str_list += ['func : ' + hex(self.func)] except ValueError: str_list += ['func : '] try: str_list += ['gridDim :\n' + '\n'.join([' ' + line for line in str(self.gridDim).splitlines()])] except ValueError: str_list += ['gridDim : '] try: str_list += ['blockDim :\n' + '\n'.join([' ' + line for line in str(self.blockDim).splitlines()])] except ValueError: str_list += ['blockDim : '] try: str_list += ['sharedMemBytes : ' + str(self.sharedMemBytes)] except ValueError: str_list += ['sharedMemBytes : '] try: str_list += ['kernelParams : ' + str(self.kernelParams)] except ValueError: str_list += ['kernelParams : '] try: str_list += ['extra : ' + str(self.extra)] except ValueError: str_list += ['extra : '] return '\n'.join(str_list) else: return '' @property def func(self): return self._pvt_ptr[0].func @func.setter def func(self, func): _cfunc = _HelperInputVoidPtr(func) self._pvt_ptr[0].func = _cfunc.cptr @property def gridDim(self): return self._gridDim @gridDim.setter def gridDim(self, gridDim not None : dim3): string.memcpy(&self._pvt_ptr[0].gridDim, gridDim.getPtr(), sizeof(self._pvt_ptr[0].gridDim)) @property def blockDim(self): return self._blockDim @blockDim.setter def blockDim(self, blockDim not None : dim3): string.memcpy(&self._pvt_ptr[0].blockDim, blockDim.getPtr(), sizeof(self._pvt_ptr[0].blockDim)) @property def sharedMemBytes(self): return self._pvt_ptr[0].sharedMemBytes @sharedMemBytes.setter def sharedMemBytes(self, unsigned int sharedMemBytes): self._pvt_ptr[0].sharedMemBytes = sharedMemBytes @property def kernelParams(self): return self._pvt_ptr[0].kernelParams @kernelParams.setter def kernelParams(self, kernelParams): self._cykernelParams = _HelperKernelParams(kernelParams) self._pvt_ptr[0].kernelParams = self._cykernelParams.ckernelParams @property def extra(self): return self._pvt_ptr[0].extra @extra.setter def extra(self, void_ptr extra): self._pvt_ptr[0].extra = extra cdef class cudaExternalSemaphoreSignalNodeParams: """ External semaphore signal node parameters Attributes ---------- extSemArray : cudaExternalSemaphore_t Array of external semaphore handles. paramsArray : cudaExternalSemaphoreSignalParams Array of external semaphore signal parameters. numExtSems : unsigned int Number of handles and parameters supplied in extSemArray and paramsArray. Methods ------- getPtr() Get memory address of class instance """ def __cinit__(self, void_ptr _ptr = 0): if _ptr == 0: self._pvt_ptr = &self._pvt_val else: self._pvt_ptr = _ptr def __init__(self, void_ptr _ptr = 0): pass def __dealloc__(self): pass if self._extSemArray is not NULL: free(self._extSemArray) if self._paramsArray is not NULL: free(self._paramsArray) def getPtr(self): return self._pvt_ptr def __repr__(self): if self._pvt_ptr is not NULL: str_list = [] try: str_list += ['extSemArray : ' + str(self.extSemArray)] except ValueError: str_list += ['extSemArray : '] try: str_list += ['paramsArray : ' + str(self.paramsArray)] except ValueError: str_list += ['paramsArray : '] try: str_list += ['numExtSems : ' + str(self.numExtSems)] except ValueError: str_list += ['numExtSems : '] return '\n'.join(str_list) else: return '' @property def extSemArray(self): arrs = [self._pvt_ptr[0].extSemArray + x*sizeof(cyruntime.cudaExternalSemaphore_t) for x in range(self._extSemArray_length)] return [cudaExternalSemaphore_t(_ptr=arr) for arr in arrs] @extSemArray.setter def extSemArray(self, val): if len(val) == 0: free(self._extSemArray) self._extSemArray_length = 0 self._pvt_ptr[0].extSemArray = NULL else: if self._extSemArray_length != len(val): free(self._extSemArray) self._extSemArray = calloc(len(val), sizeof(cyruntime.cudaExternalSemaphore_t)) if self._extSemArray is NULL: raise MemoryError('Failed to allocate length x size memory: ' + str(len(val)) + 'x' + str(sizeof(cyruntime.cudaExternalSemaphore_t))) self._extSemArray_length = len(val) self._pvt_ptr[0].extSemArray = self._extSemArray for idx in range(len(val)): self._extSemArray[idx] = (val[idx])._pvt_ptr[0] @property def paramsArray(self): arrs = [self._pvt_ptr[0].paramsArray + x*sizeof(cyruntime.cudaExternalSemaphoreSignalParams) for x in range(self._paramsArray_length)] return [cudaExternalSemaphoreSignalParams(_ptr=arr) for arr in arrs] @paramsArray.setter def paramsArray(self, val): if len(val) == 0: free(self._paramsArray) self._paramsArray_length = 0 self._pvt_ptr[0].paramsArray = NULL else: if self._paramsArray_length != len(val): free(self._paramsArray) self._paramsArray = calloc(len(val), sizeof(cyruntime.cudaExternalSemaphoreSignalParams)) if self._paramsArray is NULL: raise MemoryError('Failed to allocate length x size memory: ' + str(len(val)) + 'x' + str(sizeof(cyruntime.cudaExternalSemaphoreSignalParams))) self._paramsArray_length = len(val) self._pvt_ptr[0].paramsArray = self._paramsArray for idx in range(len(val)): string.memcpy(&self._paramsArray[idx], (val[idx])._pvt_ptr, sizeof(cyruntime.cudaExternalSemaphoreSignalParams)) @property def numExtSems(self): return self._pvt_ptr[0].numExtSems @numExtSems.setter def numExtSems(self, unsigned int numExtSems): self._pvt_ptr[0].numExtSems = numExtSems cdef class cudaExternalSemaphoreSignalNodeParamsV2: """ External semaphore signal node parameters Attributes ---------- extSemArray : cudaExternalSemaphore_t Array of external semaphore handles. paramsArray : cudaExternalSemaphoreSignalParams Array of external semaphore signal parameters. numExtSems : unsigned int Number of handles and parameters supplied in extSemArray and paramsArray. Methods ------- getPtr() Get memory address of class instance """ def __cinit__(self, void_ptr _ptr = 0): if _ptr == 0: self._pvt_ptr = &self._pvt_val else: self._pvt_ptr = _ptr def __init__(self, void_ptr _ptr = 0): pass def __dealloc__(self): pass if self._extSemArray is not NULL: free(self._extSemArray) if self._paramsArray is not NULL: free(self._paramsArray) def getPtr(self): return self._pvt_ptr def __repr__(self): if self._pvt_ptr is not NULL: str_list = [] try: str_list += ['extSemArray : ' + str(self.extSemArray)] except ValueError: str_list += ['extSemArray : '] try: str_list += ['paramsArray : ' + str(self.paramsArray)] except ValueError: str_list += ['paramsArray : '] try: str_list += ['numExtSems : ' + str(self.numExtSems)] except ValueError: str_list += ['numExtSems : '] return '\n'.join(str_list) else: return '' @property def extSemArray(self): arrs = [self._pvt_ptr[0].extSemArray + x*sizeof(cyruntime.cudaExternalSemaphore_t) for x in range(self._extSemArray_length)] return [cudaExternalSemaphore_t(_ptr=arr) for arr in arrs] @extSemArray.setter def extSemArray(self, val): if len(val) == 0: free(self._extSemArray) self._extSemArray_length = 0 self._pvt_ptr[0].extSemArray = NULL else: if self._extSemArray_length != len(val): free(self._extSemArray) self._extSemArray = calloc(len(val), sizeof(cyruntime.cudaExternalSemaphore_t)) if self._extSemArray is NULL: raise MemoryError('Failed to allocate length x size memory: ' + str(len(val)) + 'x' + str(sizeof(cyruntime.cudaExternalSemaphore_t))) self._extSemArray_length = len(val) self._pvt_ptr[0].extSemArray = self._extSemArray for idx in range(len(val)): self._extSemArray[idx] = (val[idx])._pvt_ptr[0] @property def paramsArray(self): arrs = [self._pvt_ptr[0].paramsArray + x*sizeof(cyruntime.cudaExternalSemaphoreSignalParams) for x in range(self._paramsArray_length)] return [cudaExternalSemaphoreSignalParams(_ptr=arr) for arr in arrs] @paramsArray.setter def paramsArray(self, val): if len(val) == 0: free(self._paramsArray) self._paramsArray_length = 0 self._pvt_ptr[0].paramsArray = NULL else: if self._paramsArray_length != len(val): free(self._paramsArray) self._paramsArray = calloc(len(val), sizeof(cyruntime.cudaExternalSemaphoreSignalParams)) if self._paramsArray is NULL: raise MemoryError('Failed to allocate length x size memory: ' + str(len(val)) + 'x' + str(sizeof(cyruntime.cudaExternalSemaphoreSignalParams))) self._paramsArray_length = len(val) self._pvt_ptr[0].paramsArray = self._paramsArray for idx in range(len(val)): string.memcpy(&self._paramsArray[idx], (val[idx])._pvt_ptr, sizeof(cyruntime.cudaExternalSemaphoreSignalParams)) @property def numExtSems(self): return self._pvt_ptr[0].numExtSems @numExtSems.setter def numExtSems(self, unsigned int numExtSems): self._pvt_ptr[0].numExtSems = numExtSems cdef class cudaExternalSemaphoreWaitNodeParams: """ External semaphore wait node parameters Attributes ---------- extSemArray : cudaExternalSemaphore_t Array of external semaphore handles. paramsArray : cudaExternalSemaphoreWaitParams Array of external semaphore wait parameters. numExtSems : unsigned int Number of handles and parameters supplied in extSemArray and paramsArray. Methods ------- getPtr() Get memory address of class instance """ def __cinit__(self, void_ptr _ptr = 0): if _ptr == 0: self._pvt_ptr = &self._pvt_val else: self._pvt_ptr = _ptr def __init__(self, void_ptr _ptr = 0): pass def __dealloc__(self): pass if self._extSemArray is not NULL: free(self._extSemArray) if self._paramsArray is not NULL: free(self._paramsArray) def getPtr(self): return self._pvt_ptr def __repr__(self): if self._pvt_ptr is not NULL: str_list = [] try: str_list += ['extSemArray : ' + str(self.extSemArray)] except ValueError: str_list += ['extSemArray : '] try: str_list += ['paramsArray : ' + str(self.paramsArray)] except ValueError: str_list += ['paramsArray : '] try: str_list += ['numExtSems : ' + str(self.numExtSems)] except ValueError: str_list += ['numExtSems : '] return '\n'.join(str_list) else: return '' @property def extSemArray(self): arrs = [self._pvt_ptr[0].extSemArray + x*sizeof(cyruntime.cudaExternalSemaphore_t) for x in range(self._extSemArray_length)] return [cudaExternalSemaphore_t(_ptr=arr) for arr in arrs] @extSemArray.setter def extSemArray(self, val): if len(val) == 0: free(self._extSemArray) self._extSemArray_length = 0 self._pvt_ptr[0].extSemArray = NULL else: if self._extSemArray_length != len(val): free(self._extSemArray) self._extSemArray = calloc(len(val), sizeof(cyruntime.cudaExternalSemaphore_t)) if self._extSemArray is NULL: raise MemoryError('Failed to allocate length x size memory: ' + str(len(val)) + 'x' + str(sizeof(cyruntime.cudaExternalSemaphore_t))) self._extSemArray_length = len(val) self._pvt_ptr[0].extSemArray = self._extSemArray for idx in range(len(val)): self._extSemArray[idx] = (val[idx])._pvt_ptr[0] @property def paramsArray(self): arrs = [self._pvt_ptr[0].paramsArray + x*sizeof(cyruntime.cudaExternalSemaphoreWaitParams) for x in range(self._paramsArray_length)] return [cudaExternalSemaphoreWaitParams(_ptr=arr) for arr in arrs] @paramsArray.setter def paramsArray(self, val): if len(val) == 0: free(self._paramsArray) self._paramsArray_length = 0 self._pvt_ptr[0].paramsArray = NULL else: if self._paramsArray_length != len(val): free(self._paramsArray) self._paramsArray = calloc(len(val), sizeof(cyruntime.cudaExternalSemaphoreWaitParams)) if self._paramsArray is NULL: raise MemoryError('Failed to allocate length x size memory: ' + str(len(val)) + 'x' + str(sizeof(cyruntime.cudaExternalSemaphoreWaitParams))) self._paramsArray_length = len(val) self._pvt_ptr[0].paramsArray = self._paramsArray for idx in range(len(val)): string.memcpy(&self._paramsArray[idx], (val[idx])._pvt_ptr, sizeof(cyruntime.cudaExternalSemaphoreWaitParams)) @property def numExtSems(self): return self._pvt_ptr[0].numExtSems @numExtSems.setter def numExtSems(self, unsigned int numExtSems): self._pvt_ptr[0].numExtSems = numExtSems cdef class cudaExternalSemaphoreWaitNodeParamsV2: """ External semaphore wait node parameters Attributes ---------- extSemArray : cudaExternalSemaphore_t Array of external semaphore handles. paramsArray : cudaExternalSemaphoreWaitParams Array of external semaphore wait parameters. numExtSems : unsigned int Number of handles and parameters supplied in extSemArray and paramsArray. Methods ------- getPtr() Get memory address of class instance """ def __cinit__(self, void_ptr _ptr = 0): if _ptr == 0: self._pvt_ptr = &self._pvt_val else: self._pvt_ptr = _ptr def __init__(self, void_ptr _ptr = 0): pass def __dealloc__(self): pass if self._extSemArray is not NULL: free(self._extSemArray) if self._paramsArray is not NULL: free(self._paramsArray) def getPtr(self): return self._pvt_ptr def __repr__(self): if self._pvt_ptr is not NULL: str_list = [] try: str_list += ['extSemArray : ' + str(self.extSemArray)] except ValueError: str_list += ['extSemArray : '] try: str_list += ['paramsArray : ' + str(self.paramsArray)] except ValueError: str_list += ['paramsArray : '] try: str_list += ['numExtSems : ' + str(self.numExtSems)] except ValueError: str_list += ['numExtSems : '] return '\n'.join(str_list) else: return '' @property def extSemArray(self): arrs = [self._pvt_ptr[0].extSemArray + x*sizeof(cyruntime.cudaExternalSemaphore_t) for x in range(self._extSemArray_length)] return [cudaExternalSemaphore_t(_ptr=arr) for arr in arrs] @extSemArray.setter def extSemArray(self, val): if len(val) == 0: free(self._extSemArray) self._extSemArray_length = 0 self._pvt_ptr[0].extSemArray = NULL else: if self._extSemArray_length != len(val): free(self._extSemArray) self._extSemArray = calloc(len(val), sizeof(cyruntime.cudaExternalSemaphore_t)) if self._extSemArray is NULL: raise MemoryError('Failed to allocate length x size memory: ' + str(len(val)) + 'x' + str(sizeof(cyruntime.cudaExternalSemaphore_t))) self._extSemArray_length = len(val) self._pvt_ptr[0].extSemArray = self._extSemArray for idx in range(len(val)): self._extSemArray[idx] = (val[idx])._pvt_ptr[0] @property def paramsArray(self): arrs = [self._pvt_ptr[0].paramsArray + x*sizeof(cyruntime.cudaExternalSemaphoreWaitParams) for x in range(self._paramsArray_length)] return [cudaExternalSemaphoreWaitParams(_ptr=arr) for arr in arrs] @paramsArray.setter def paramsArray(self, val): if len(val) == 0: free(self._paramsArray) self._paramsArray_length = 0 self._pvt_ptr[0].paramsArray = NULL else: if self._paramsArray_length != len(val): free(self._paramsArray) self._paramsArray = calloc(len(val), sizeof(cyruntime.cudaExternalSemaphoreWaitParams)) if self._paramsArray is NULL: raise MemoryError('Failed to allocate length x size memory: ' + str(len(val)) + 'x' + str(sizeof(cyruntime.cudaExternalSemaphoreWaitParams))) self._paramsArray_length = len(val) self._pvt_ptr[0].paramsArray = self._paramsArray for idx in range(len(val)): string.memcpy(&self._paramsArray[idx], (val[idx])._pvt_ptr, sizeof(cyruntime.cudaExternalSemaphoreWaitParams)) @property def numExtSems(self): return self._pvt_ptr[0].numExtSems @numExtSems.setter def numExtSems(self, unsigned int numExtSems): self._pvt_ptr[0].numExtSems = numExtSems cdef class cudaConditionalNodeParams: """ CUDA conditional node parameters Attributes ---------- handle : cudaGraphConditionalHandle Conditional node handle. Handles must be created in advance of creating the node using cudaGraphConditionalHandleCreate. type : cudaGraphConditionalNodeType Type of conditional node. size : unsigned int Size of graph output array. Allowed values are 1 for cudaGraphCondTypeWhile, 1 or 2 for cudaGraphCondTypeWhile, or any value greater than zero for cudaGraphCondTypeSwitch. phGraph_out : cudaGraph_t CUDA-owned array populated with conditional node child graphs during creation of the node. Valid for the lifetime of the conditional node. The contents of the graph(s) are subject to the following constraints: - Allowed node types are kernel nodes, empty nodes, child graphs, memsets, memcopies, and conditionals. This applies recursively to child graphs and conditional bodies. - All kernels, including kernels in nested conditionals or child graphs at any level, must belong to the same CUDA context. These graphs may be populated using graph node creation APIs or cudaStreamBeginCaptureToGraph. cudaGraphCondTypeIf: phGraph_out[0] is executed when the condition is non-zero. If `size` == 2, phGraph_out[1] will be executed when the condition is zero. cudaGraphCondTypeWhile: phGraph_out[0] is executed as long as the condition is non-zero. cudaGraphCondTypeSwitch: phGraph_out[n] is executed when the condition is equal to n. If the condition >= `size`, no body graph is executed. Methods ------- getPtr() Get memory address of class instance """ def __cinit__(self, void_ptr _ptr = 0): if _ptr == 0: self._pvt_ptr = &self._pvt_val else: self._pvt_ptr = _ptr def __init__(self, void_ptr _ptr = 0): pass self._handle = cudaGraphConditionalHandle(_ptr=&self._pvt_ptr[0].handle) def __dealloc__(self): pass def getPtr(self): return self._pvt_ptr def __repr__(self): if self._pvt_ptr is not NULL: str_list = [] try: str_list += ['handle : ' + str(self.handle)] except ValueError: str_list += ['handle : '] try: str_list += ['type : ' + str(self.type)] except ValueError: str_list += ['type : '] try: str_list += ['size : ' + str(self.size)] except ValueError: str_list += ['size : '] try: str_list += ['phGraph_out : ' + str(self.phGraph_out)] except ValueError: str_list += ['phGraph_out : '] return '\n'.join(str_list) else: return '' @property def handle(self): return self._handle @handle.setter def handle(self, handle): cdef cyruntime.cudaGraphConditionalHandle cyhandle if handle is None: cyhandle = 0 elif isinstance(handle, (cudaGraphConditionalHandle)): phandle = int(handle) cyhandle = phandle else: phandle = int(cudaGraphConditionalHandle(handle)) cyhandle = phandle self._handle._pvt_ptr[0] = cyhandle @property def type(self): if self._pvt_ptr[0].type not in _dict_cudaGraphConditionalNodeType: return None return _dict_cudaGraphConditionalNodeType[self._pvt_ptr[0].type] @type.setter def type(self, type not None : cudaGraphConditionalNodeType): self._pvt_ptr[0].type = type.value @property def size(self): return self._pvt_ptr[0].size @size.setter def size(self, unsigned int size): self._pvt_ptr[0].size = size @property def phGraph_out(self): arrs = [self._pvt_ptr[0].phGraph_out + x*sizeof(cyruntime.cudaGraph_t) for x in range(self.size)] return [cudaGraph_t(_ptr=arr) for arr in arrs] cdef class cudaChildGraphNodeParams: """ Child graph node parameters Attributes ---------- graph : cudaGraph_t The child graph to clone into the node for node creation, or a handle to the graph owned by the node for node query. The graph must not contain conditional nodes. Graphs containing memory allocation or memory free nodes must set the ownership to be moved to the parent. ownership : cudaGraphChildGraphNodeOwnership The ownership relationship of the child graph node. Methods ------- getPtr() Get memory address of class instance """ def __cinit__(self, void_ptr _ptr = 0): if _ptr == 0: self._pvt_ptr = &self._pvt_val else: self._pvt_ptr = _ptr def __init__(self, void_ptr _ptr = 0): pass self._graph = cudaGraph_t(_ptr=&self._pvt_ptr[0].graph) def __dealloc__(self): pass def getPtr(self): return self._pvt_ptr def __repr__(self): if self._pvt_ptr is not NULL: str_list = [] try: str_list += ['graph : ' + str(self.graph)] except ValueError: str_list += ['graph : '] try: str_list += ['ownership : ' + str(self.ownership)] except ValueError: str_list += ['ownership : '] return '\n'.join(str_list) else: return '' @property def graph(self): return self._graph @graph.setter def graph(self, graph): cdef cyruntime.cudaGraph_t cygraph if graph is None: cygraph = 0 elif isinstance(graph, (cudaGraph_t,driver.CUgraph)): pgraph = int(graph) cygraph = pgraph else: pgraph = int(cudaGraph_t(graph)) cygraph = pgraph self._graph._pvt_ptr[0] = cygraph @property def ownership(self): if self._pvt_ptr[0].ownership not in _dict_cudaGraphChildGraphNodeOwnership: return None return _dict_cudaGraphChildGraphNodeOwnership[self._pvt_ptr[0].ownership] @ownership.setter def ownership(self, ownership not None : cudaGraphChildGraphNodeOwnership): self._pvt_ptr[0].ownership = ownership.value cdef class cudaEventRecordNodeParams: """ Event record node parameters Attributes ---------- event : cudaEvent_t The event to record when the node executes Methods ------- getPtr() Get memory address of class instance """ def __cinit__(self, void_ptr _ptr = 0): if _ptr == 0: self._pvt_ptr = &self._pvt_val else: self._pvt_ptr = _ptr def __init__(self, void_ptr _ptr = 0): pass self._event = cudaEvent_t(_ptr=&self._pvt_ptr[0].event) def __dealloc__(self): pass def getPtr(self): return self._pvt_ptr def __repr__(self): if self._pvt_ptr is not NULL: str_list = [] try: str_list += ['event : ' + str(self.event)] except ValueError: str_list += ['event : '] return '\n'.join(str_list) else: return '' @property def event(self): return self._event @event.setter def event(self, event): cdef cyruntime.cudaEvent_t cyevent if event is None: cyevent = 0 elif isinstance(event, (cudaEvent_t,driver.CUevent)): pevent = int(event) cyevent = pevent else: pevent = int(cudaEvent_t(event)) cyevent = pevent self._event._pvt_ptr[0] = cyevent cdef class cudaEventWaitNodeParams: """ Event wait node parameters Attributes ---------- event : cudaEvent_t The event to wait on from the node Methods ------- getPtr() Get memory address of class instance """ def __cinit__(self, void_ptr _ptr = 0): if _ptr == 0: self._pvt_ptr = &self._pvt_val else: self._pvt_ptr = _ptr def __init__(self, void_ptr _ptr = 0): pass self._event = cudaEvent_t(_ptr=&self._pvt_ptr[0].event) def __dealloc__(self): pass def getPtr(self): return self._pvt_ptr def __repr__(self): if self._pvt_ptr is not NULL: str_list = [] try: str_list += ['event : ' + str(self.event)] except ValueError: str_list += ['event : '] return '\n'.join(str_list) else: return '' @property def event(self): return self._event @event.setter def event(self, event): cdef cyruntime.cudaEvent_t cyevent if event is None: cyevent = 0 elif isinstance(event, (cudaEvent_t,driver.CUevent)): pevent = int(event) cyevent = pevent else: pevent = int(cudaEvent_t(event)) cyevent = pevent self._event._pvt_ptr[0] = cyevent cdef class cudaGraphNodeParams: """ Graph node parameters. See cudaGraphAddNode. Attributes ---------- type : cudaGraphNodeType Type of the node reserved0 : list[int] Reserved. Must be zero. reserved1 : list[long long] Padding. Unused bytes must be zero. kernel : cudaKernelNodeParamsV2 Kernel node parameters. memcpy : cudaMemcpyNodeParams Memcpy node parameters. memset : cudaMemsetParamsV2 Memset node parameters. host : cudaHostNodeParamsV2 Host node parameters. graph : cudaChildGraphNodeParams Child graph node parameters. eventWait : cudaEventWaitNodeParams Event wait node parameters. eventRecord : cudaEventRecordNodeParams Event record node parameters. extSemSignal : cudaExternalSemaphoreSignalNodeParamsV2 External semaphore signal node parameters. extSemWait : cudaExternalSemaphoreWaitNodeParamsV2 External semaphore wait node parameters. alloc : cudaMemAllocNodeParamsV2 Memory allocation node parameters. free : cudaMemFreeNodeParams Memory free node parameters. conditional : cudaConditionalNodeParams Conditional node parameters. reserved2 : long long Reserved bytes. Must be zero. Methods ------- getPtr() Get memory address of class instance """ def __cinit__(self, void_ptr _ptr = 0): if _ptr == 0: self._val_ptr = calloc(1, sizeof(cyruntime.cudaGraphNodeParams)) self._pvt_ptr = self._val_ptr else: self._pvt_ptr = _ptr def __init__(self, void_ptr _ptr = 0): pass self._kernel = cudaKernelNodeParamsV2(_ptr=&self._pvt_ptr[0].kernel) self._memcpy = cudaMemcpyNodeParams(_ptr=&self._pvt_ptr[0].memcpy) self._memset = cudaMemsetParamsV2(_ptr=&self._pvt_ptr[0].memset) self._host = cudaHostNodeParamsV2(_ptr=&self._pvt_ptr[0].host) self._graph = cudaChildGraphNodeParams(_ptr=&self._pvt_ptr[0].graph) self._eventWait = cudaEventWaitNodeParams(_ptr=&self._pvt_ptr[0].eventWait) self._eventRecord = cudaEventRecordNodeParams(_ptr=&self._pvt_ptr[0].eventRecord) self._extSemSignal = cudaExternalSemaphoreSignalNodeParamsV2(_ptr=&self._pvt_ptr[0].extSemSignal) self._extSemWait = cudaExternalSemaphoreWaitNodeParamsV2(_ptr=&self._pvt_ptr[0].extSemWait) self._alloc = cudaMemAllocNodeParamsV2(_ptr=&self._pvt_ptr[0].alloc) self._free = cudaMemFreeNodeParams(_ptr=&self._pvt_ptr[0].free) self._conditional = cudaConditionalNodeParams(_ptr=&self._pvt_ptr[0].conditional) def __dealloc__(self): if self._val_ptr is not NULL: free(self._val_ptr) def getPtr(self): return self._pvt_ptr def __repr__(self): if self._pvt_ptr is not NULL: str_list = [] try: str_list += ['type : ' + str(self.type)] except ValueError: str_list += ['type : '] try: str_list += ['reserved0 : ' + str(self.reserved0)] except ValueError: str_list += ['reserved0 : '] try: str_list += ['reserved1 : ' + str(self.reserved1)] except ValueError: str_list += ['reserved1 : '] try: str_list += ['kernel :\n' + '\n'.join([' ' + line for line in str(self.kernel).splitlines()])] except ValueError: str_list += ['kernel : '] try: str_list += ['memcpy :\n' + '\n'.join([' ' + line for line in str(self.memcpy).splitlines()])] except ValueError: str_list += ['memcpy : '] try: str_list += ['memset :\n' + '\n'.join([' ' + line for line in str(self.memset).splitlines()])] except ValueError: str_list += ['memset : '] try: str_list += ['host :\n' + '\n'.join([' ' + line for line in str(self.host).splitlines()])] except ValueError: str_list += ['host : '] try: str_list += ['graph :\n' + '\n'.join([' ' + line for line in str(self.graph).splitlines()])] except ValueError: str_list += ['graph : '] try: str_list += ['eventWait :\n' + '\n'.join([' ' + line for line in str(self.eventWait).splitlines()])] except ValueError: str_list += ['eventWait : '] try: str_list += ['eventRecord :\n' + '\n'.join([' ' + line for line in str(self.eventRecord).splitlines()])] except ValueError: str_list += ['eventRecord : '] try: str_list += ['extSemSignal :\n' + '\n'.join([' ' + line for line in str(self.extSemSignal).splitlines()])] except ValueError: str_list += ['extSemSignal : '] try: str_list += ['extSemWait :\n' + '\n'.join([' ' + line for line in str(self.extSemWait).splitlines()])] except ValueError: str_list += ['extSemWait : '] try: str_list += ['alloc :\n' + '\n'.join([' ' + line for line in str(self.alloc).splitlines()])] except ValueError: str_list += ['alloc : '] try: str_list += ['free :\n' + '\n'.join([' ' + line for line in str(self.free).splitlines()])] except ValueError: str_list += ['free : '] try: str_list += ['conditional :\n' + '\n'.join([' ' + line for line in str(self.conditional).splitlines()])] except ValueError: str_list += ['conditional : '] try: str_list += ['reserved2 : ' + str(self.reserved2)] except ValueError: str_list += ['reserved2 : '] return '\n'.join(str_list) else: return '' @property def type(self): if self._pvt_ptr[0].type not in _dict_cudaGraphNodeType: return None return _dict_cudaGraphNodeType[self._pvt_ptr[0].type] @type.setter def type(self, type not None : cudaGraphNodeType): self._pvt_ptr[0].type = type.value @property def reserved0(self): return self._pvt_ptr[0].reserved0 @reserved0.setter def reserved0(self, reserved0): self._pvt_ptr[0].reserved0 = reserved0 @property def reserved1(self): return self._pvt_ptr[0].reserved1 @reserved1.setter def reserved1(self, reserved1): self._pvt_ptr[0].reserved1 = reserved1 @property def kernel(self): return self._kernel @kernel.setter def kernel(self, kernel not None : cudaKernelNodeParamsV2): string.memcpy(&self._pvt_ptr[0].kernel, kernel.getPtr(), sizeof(self._pvt_ptr[0].kernel)) @property def memcpy(self): return self._memcpy @memcpy.setter def memcpy(self, memcpy not None : cudaMemcpyNodeParams): string.memcpy(&self._pvt_ptr[0].memcpy, memcpy.getPtr(), sizeof(self._pvt_ptr[0].memcpy)) @property def memset(self): return self._memset @memset.setter def memset(self, memset not None : cudaMemsetParamsV2): string.memcpy(&self._pvt_ptr[0].memset, memset.getPtr(), sizeof(self._pvt_ptr[0].memset)) @property def host(self): return self._host @host.setter def host(self, host not None : cudaHostNodeParamsV2): string.memcpy(&self._pvt_ptr[0].host, host.getPtr(), sizeof(self._pvt_ptr[0].host)) @property def graph(self): return self._graph @graph.setter def graph(self, graph not None : cudaChildGraphNodeParams): string.memcpy(&self._pvt_ptr[0].graph, graph.getPtr(), sizeof(self._pvt_ptr[0].graph)) @property def eventWait(self): return self._eventWait @eventWait.setter def eventWait(self, eventWait not None : cudaEventWaitNodeParams): string.memcpy(&self._pvt_ptr[0].eventWait, eventWait.getPtr(), sizeof(self._pvt_ptr[0].eventWait)) @property def eventRecord(self): return self._eventRecord @eventRecord.setter def eventRecord(self, eventRecord not None : cudaEventRecordNodeParams): string.memcpy(&self._pvt_ptr[0].eventRecord, eventRecord.getPtr(), sizeof(self._pvt_ptr[0].eventRecord)) @property def extSemSignal(self): return self._extSemSignal @extSemSignal.setter def extSemSignal(self, extSemSignal not None : cudaExternalSemaphoreSignalNodeParamsV2): string.memcpy(&self._pvt_ptr[0].extSemSignal, extSemSignal.getPtr(), sizeof(self._pvt_ptr[0].extSemSignal)) @property def extSemWait(self): return self._extSemWait @extSemWait.setter def extSemWait(self, extSemWait not None : cudaExternalSemaphoreWaitNodeParamsV2): string.memcpy(&self._pvt_ptr[0].extSemWait, extSemWait.getPtr(), sizeof(self._pvt_ptr[0].extSemWait)) @property def alloc(self): return self._alloc @alloc.setter def alloc(self, alloc not None : cudaMemAllocNodeParamsV2): string.memcpy(&self._pvt_ptr[0].alloc, alloc.getPtr(), sizeof(self._pvt_ptr[0].alloc)) @property def free(self): return self._free @free.setter def free(self, free not None : cudaMemFreeNodeParams): string.memcpy(&self._pvt_ptr[0].free, free.getPtr(), sizeof(self._pvt_ptr[0].free)) @property def conditional(self): return self._conditional @conditional.setter def conditional(self, conditional not None : cudaConditionalNodeParams): string.memcpy(&self._pvt_ptr[0].conditional, conditional.getPtr(), sizeof(self._pvt_ptr[0].conditional)) @property def reserved2(self): return self._pvt_ptr[0].reserved2 @reserved2.setter def reserved2(self, long long reserved2): self._pvt_ptr[0].reserved2 = reserved2 cdef class cudaGraphEdgeData_st: """ Optional annotation for edges in a CUDA graph. Note, all edges implicitly have annotations and default to a zero-initialized value if not specified. A zero-initialized struct indicates a standard full serialization of two nodes with memory visibility. Attributes ---------- from_port : bytes This indicates when the dependency is triggered from the upstream node on the edge. The meaning is specfic to the node type. A value of 0 in all cases means full completion of the upstream node, with memory visibility to the downstream node or portion thereof (indicated by `to_port`). Only kernel nodes define non-zero ports. A kernel node can use the following output port types: cudaGraphKernelNodePortDefault, cudaGraphKernelNodePortProgrammatic, or cudaGraphKernelNodePortLaunchCompletion. to_port : bytes This indicates what portion of the downstream node is dependent on the upstream node or portion thereof (indicated by `from_port`). The meaning is specific to the node type. A value of 0 in all cases means the entirety of the downstream node is dependent on the upstream work. Currently no node types define non-zero ports. Accordingly, this field must be set to zero. type : bytes This should be populated with a value from ::cudaGraphDependencyType. (It is typed as char due to compiler- specific layout of bitfields.) See ::cudaGraphDependencyType. reserved : bytes These bytes are unused and must be zeroed. This ensures compatibility if additional fields are added in the future. Methods ------- getPtr() Get memory address of class instance """ def __cinit__(self, void_ptr _ptr = 0): if _ptr == 0: self._pvt_ptr = &self._pvt_val else: self._pvt_ptr = _ptr def __init__(self, void_ptr _ptr = 0): pass def __dealloc__(self): pass def getPtr(self): return self._pvt_ptr def __repr__(self): if self._pvt_ptr is not NULL: str_list = [] try: str_list += ['from_port : ' + str(self.from_port)] except ValueError: str_list += ['from_port : '] try: str_list += ['to_port : ' + str(self.to_port)] except ValueError: str_list += ['to_port : '] try: str_list += ['type : ' + str(self.type)] except ValueError: str_list += ['type : '] try: str_list += ['reserved : ' + str(self.reserved)] except ValueError: str_list += ['reserved : '] return '\n'.join(str_list) else: return '' @property def from_port(self): return self._pvt_ptr[0].from_port @from_port.setter def from_port(self, unsigned char from_port): self._pvt_ptr[0].from_port = from_port @property def to_port(self): return self._pvt_ptr[0].to_port @to_port.setter def to_port(self, unsigned char to_port): self._pvt_ptr[0].to_port = to_port @property def type(self): return self._pvt_ptr[0].type @type.setter def type(self, unsigned char type): self._pvt_ptr[0].type = type @property def reserved(self): return PyBytes_FromStringAndSize(self._pvt_ptr[0].reserved, 5) @reserved.setter def reserved(self, reserved): if len(reserved) != 5: raise ValueError("reserved length must be 5, is " + str(len(reserved))) for i, b in enumerate(reserved): self._pvt_ptr[0].reserved[i] = b cdef class cudaGraphInstantiateParams_st: """ Graph instantiation parameters Attributes ---------- flags : unsigned long long Instantiation flags uploadStream : cudaStream_t Upload stream errNode_out : cudaGraphNode_t The node which caused instantiation to fail, if any result_out : cudaGraphInstantiateResult Whether instantiation was successful. If it failed, the reason why Methods ------- getPtr() Get memory address of class instance """ def __cinit__(self, void_ptr _ptr = 0): if _ptr == 0: self._pvt_ptr = &self._pvt_val else: self._pvt_ptr = _ptr def __init__(self, void_ptr _ptr = 0): pass self._uploadStream = cudaStream_t(_ptr=&self._pvt_ptr[0].uploadStream) self._errNode_out = cudaGraphNode_t(_ptr=&self._pvt_ptr[0].errNode_out) def __dealloc__(self): pass def getPtr(self): return self._pvt_ptr def __repr__(self): if self._pvt_ptr is not NULL: str_list = [] try: str_list += ['flags : ' + str(self.flags)] except ValueError: str_list += ['flags : '] try: str_list += ['uploadStream : ' + str(self.uploadStream)] except ValueError: str_list += ['uploadStream : '] try: str_list += ['errNode_out : ' + str(self.errNode_out)] except ValueError: str_list += ['errNode_out : '] try: str_list += ['result_out : ' + str(self.result_out)] except ValueError: str_list += ['result_out : '] return '\n'.join(str_list) else: return '' @property def flags(self): return self._pvt_ptr[0].flags @flags.setter def flags(self, unsigned long long flags): self._pvt_ptr[0].flags = flags @property def uploadStream(self): return self._uploadStream @uploadStream.setter def uploadStream(self, uploadStream): cdef cyruntime.cudaStream_t cyuploadStream if uploadStream is None: cyuploadStream = 0 elif isinstance(uploadStream, (cudaStream_t,driver.CUstream)): puploadStream = int(uploadStream) cyuploadStream = puploadStream else: puploadStream = int(cudaStream_t(uploadStream)) cyuploadStream = puploadStream self._uploadStream._pvt_ptr[0] = cyuploadStream @property def errNode_out(self): return self._errNode_out @errNode_out.setter def errNode_out(self, errNode_out): cdef cyruntime.cudaGraphNode_t cyerrNode_out if errNode_out is None: cyerrNode_out = 0 elif isinstance(errNode_out, (cudaGraphNode_t,driver.CUgraphNode)): perrNode_out = int(errNode_out) cyerrNode_out = perrNode_out else: perrNode_out = int(cudaGraphNode_t(errNode_out)) cyerrNode_out = perrNode_out self._errNode_out._pvt_ptr[0] = cyerrNode_out @property def result_out(self): if self._pvt_ptr[0].result_out not in _dict_cudaGraphInstantiateResult: return None return _dict_cudaGraphInstantiateResult[self._pvt_ptr[0].result_out] @result_out.setter def result_out(self, result_out not None : cudaGraphInstantiateResult): self._pvt_ptr[0].result_out = result_out.value cdef class cudaGraphExecUpdateResultInfo_st: """ Result information returned by cudaGraphExecUpdate Attributes ---------- result : cudaGraphExecUpdateResult Gives more specific detail when a cuda graph update fails. errorNode : cudaGraphNode_t The "to node" of the error edge when the topologies do not match. The error node when the error is associated with a specific node. NULL when the error is generic. errorFromNode : cudaGraphNode_t The from node of error edge when the topologies do not match. Otherwise NULL. Methods ------- getPtr() Get memory address of class instance """ def __cinit__(self, void_ptr _ptr = 0): if _ptr == 0: self._pvt_ptr = &self._pvt_val else: self._pvt_ptr = _ptr def __init__(self, void_ptr _ptr = 0): pass self._errorNode = cudaGraphNode_t(_ptr=&self._pvt_ptr[0].errorNode) self._errorFromNode = cudaGraphNode_t(_ptr=&self._pvt_ptr[0].errorFromNode) def __dealloc__(self): pass def getPtr(self): return self._pvt_ptr def __repr__(self): if self._pvt_ptr is not NULL: str_list = [] try: str_list += ['result : ' + str(self.result)] except ValueError: str_list += ['result : '] try: str_list += ['errorNode : ' + str(self.errorNode)] except ValueError: str_list += ['errorNode : '] try: str_list += ['errorFromNode : ' + str(self.errorFromNode)] except ValueError: str_list += ['errorFromNode : '] return '\n'.join(str_list) else: return '' @property def result(self): if self._pvt_ptr[0].result not in _dict_cudaGraphExecUpdateResult: return None return _dict_cudaGraphExecUpdateResult[self._pvt_ptr[0].result] @result.setter def result(self, result not None : cudaGraphExecUpdateResult): self._pvt_ptr[0].result = result.value @property def errorNode(self): return self._errorNode @errorNode.setter def errorNode(self, errorNode): cdef cyruntime.cudaGraphNode_t cyerrorNode if errorNode is None: cyerrorNode = 0 elif isinstance(errorNode, (cudaGraphNode_t,driver.CUgraphNode)): perrorNode = int(errorNode) cyerrorNode = perrorNode else: perrorNode = int(cudaGraphNode_t(errorNode)) cyerrorNode = perrorNode self._errorNode._pvt_ptr[0] = cyerrorNode @property def errorFromNode(self): return self._errorFromNode @errorFromNode.setter def errorFromNode(self, errorFromNode): cdef cyruntime.cudaGraphNode_t cyerrorFromNode if errorFromNode is None: cyerrorFromNode = 0 elif isinstance(errorFromNode, (cudaGraphNode_t,driver.CUgraphNode)): perrorFromNode = int(errorFromNode) cyerrorFromNode = perrorFromNode else: perrorFromNode = int(cudaGraphNode_t(errorFromNode)) cyerrorFromNode = perrorFromNode self._errorFromNode._pvt_ptr[0] = cyerrorFromNode cdef class anon_struct21: """ Attributes ---------- pValue : Any offset : size_t size : size_t Methods ------- getPtr() Get memory address of class instance """ def __cinit__(self, void_ptr _ptr): self._pvt_ptr = _ptr def __init__(self, void_ptr _ptr): pass def __dealloc__(self): pass def getPtr(self): return &self._pvt_ptr[0].updateData.param def __repr__(self): if self._pvt_ptr is not NULL: str_list = [] try: str_list += ['pValue : ' + hex(self.pValue)] except ValueError: str_list += ['pValue : '] try: str_list += ['offset : ' + str(self.offset)] except ValueError: str_list += ['offset : '] try: str_list += ['size : ' + str(self.size)] except ValueError: str_list += ['size : '] return '\n'.join(str_list) else: return '' @property def pValue(self): return self._pvt_ptr[0].updateData.param.pValue @pValue.setter def pValue(self, pValue): _cpValue = _HelperInputVoidPtr(pValue) self._pvt_ptr[0].updateData.param.pValue = _cpValue.cptr @property def offset(self): return self._pvt_ptr[0].updateData.param.offset @offset.setter def offset(self, size_t offset): self._pvt_ptr[0].updateData.param.offset = offset @property def size(self): return self._pvt_ptr[0].updateData.param.size @size.setter def size(self, size_t size): self._pvt_ptr[0].updateData.param.size = size cdef class anon_union9: """ Attributes ---------- gridDim : dim3 param : anon_struct21 isEnabled : unsigned int Methods ------- getPtr() Get memory address of class instance """ def __cinit__(self, void_ptr _ptr): self._pvt_ptr = _ptr def __init__(self, void_ptr _ptr): pass self._gridDim = dim3(_ptr=&self._pvt_ptr[0].updateData.gridDim) self._param = anon_struct21(_ptr=self._pvt_ptr) def __dealloc__(self): pass def getPtr(self): return &self._pvt_ptr[0].updateData def __repr__(self): if self._pvt_ptr is not NULL: str_list = [] try: str_list += ['gridDim :\n' + '\n'.join([' ' + line for line in str(self.gridDim).splitlines()])] except ValueError: str_list += ['gridDim : '] try: str_list += ['param :\n' + '\n'.join([' ' + line for line in str(self.param).splitlines()])] except ValueError: str_list += ['param : '] try: str_list += ['isEnabled : ' + str(self.isEnabled)] except ValueError: str_list += ['isEnabled : '] return '\n'.join(str_list) else: return '' @property def gridDim(self): return self._gridDim @gridDim.setter def gridDim(self, gridDim not None : dim3): string.memcpy(&self._pvt_ptr[0].updateData.gridDim, gridDim.getPtr(), sizeof(self._pvt_ptr[0].updateData.gridDim)) @property def param(self): return self._param @param.setter def param(self, param not None : anon_struct21): string.memcpy(&self._pvt_ptr[0].updateData.param, param.getPtr(), sizeof(self._pvt_ptr[0].updateData.param)) @property def isEnabled(self): return self._pvt_ptr[0].updateData.isEnabled @isEnabled.setter def isEnabled(self, unsigned int isEnabled): self._pvt_ptr[0].updateData.isEnabled = isEnabled cdef class cudaGraphKernelNodeUpdate: """ Struct to specify a single node update to pass as part of a larger array to ::cudaGraphKernelNodeUpdatesApply Attributes ---------- node : cudaGraphDeviceNode_t Node to update field : cudaGraphKernelNodeField Which type of update to apply. Determines how updateData is interpreted updateData : anon_union9 Update data to apply. Which field is used depends on field's value Methods ------- getPtr() Get memory address of class instance """ def __cinit__(self, void_ptr _ptr = 0): if _ptr == 0: self._val_ptr = calloc(1, sizeof(cyruntime.cudaGraphKernelNodeUpdate)) self._pvt_ptr = self._val_ptr else: self._pvt_ptr = _ptr def __init__(self, void_ptr _ptr = 0): pass self._node = cudaGraphDeviceNode_t(_ptr=&self._pvt_ptr[0].node) self._updateData = anon_union9(_ptr=self._pvt_ptr) def __dealloc__(self): if self._val_ptr is not NULL: free(self._val_ptr) def getPtr(self): return self._pvt_ptr def __repr__(self): if self._pvt_ptr is not NULL: str_list = [] try: str_list += ['node : ' + str(self.node)] except ValueError: str_list += ['node : '] try: str_list += ['field : ' + str(self.field)] except ValueError: str_list += ['field : '] try: str_list += ['updateData :\n' + '\n'.join([' ' + line for line in str(self.updateData).splitlines()])] except ValueError: str_list += ['updateData : '] return '\n'.join(str_list) else: return '' @property def node(self): return self._node @node.setter def node(self, node): cdef cyruntime.cudaGraphDeviceNode_t cynode if node is None: cynode = 0 elif isinstance(node, (cudaGraphDeviceNode_t,)): pnode = int(node) cynode = pnode else: pnode = int(cudaGraphDeviceNode_t(node)) cynode = pnode self._node._pvt_ptr[0] = cynode @property def field(self): if self._pvt_ptr[0].field not in _dict_cudaGraphKernelNodeField: return None return _dict_cudaGraphKernelNodeField[self._pvt_ptr[0].field] @field.setter def field(self, field not None : cudaGraphKernelNodeField): self._pvt_ptr[0].field = field.value @property def updateData(self): return self._updateData @updateData.setter def updateData(self, updateData not None : anon_union9): string.memcpy(&self._pvt_ptr[0].updateData, updateData.getPtr(), sizeof(self._pvt_ptr[0].updateData)) cdef class cudaLaunchMemSyncDomainMap_st: """ Memory Synchronization Domain map See cudaLaunchMemSyncDomain. By default, kernels are launched in domain 0. Kernel launched with cudaLaunchMemSyncDomainRemote will have a different domain ID. User may also alter the domain ID with ::cudaLaunchMemSyncDomainMap for a specific stream / graph node / kernel launch. See cudaLaunchAttributeMemSyncDomainMap. Domain ID range is available through cudaDevAttrMemSyncDomainCount. Attributes ---------- default_ : bytes The default domain ID to use for designated kernels remote : bytes The remote domain ID to use for designated kernels Methods ------- getPtr() Get memory address of class instance """ def __cinit__(self, void_ptr _ptr = 0): if _ptr == 0: self._pvt_ptr = &self._pvt_val else: self._pvt_ptr = _ptr def __init__(self, void_ptr _ptr = 0): pass def __dealloc__(self): pass def getPtr(self): return self._pvt_ptr def __repr__(self): if self._pvt_ptr is not NULL: str_list = [] try: str_list += ['default_ : ' + str(self.default_)] except ValueError: str_list += ['default_ : '] try: str_list += ['remote : ' + str(self.remote)] except ValueError: str_list += ['remote : '] return '\n'.join(str_list) else: return '' @property def default_(self): return self._pvt_ptr[0].default_ @default_.setter def default_(self, unsigned char default_): self._pvt_ptr[0].default_ = default_ @property def remote(self): return self._pvt_ptr[0].remote @remote.setter def remote(self, unsigned char remote): self._pvt_ptr[0].remote = remote cdef class anon_struct22: """ Attributes ---------- x : unsigned int y : unsigned int z : unsigned int Methods ------- getPtr() Get memory address of class instance """ def __cinit__(self, void_ptr _ptr): self._pvt_ptr = _ptr def __init__(self, void_ptr _ptr): pass def __dealloc__(self): pass def getPtr(self): return &self._pvt_ptr[0].clusterDim def __repr__(self): if self._pvt_ptr is not NULL: str_list = [] try: str_list += ['x : ' + str(self.x)] except ValueError: str_list += ['x : '] try: str_list += ['y : ' + str(self.y)] except ValueError: str_list += ['y : '] try: str_list += ['z : ' + str(self.z)] except ValueError: str_list += ['z : '] return '\n'.join(str_list) else: return '' @property def x(self): return self._pvt_ptr[0].clusterDim.x @x.setter def x(self, unsigned int x): self._pvt_ptr[0].clusterDim.x = x @property def y(self): return self._pvt_ptr[0].clusterDim.y @y.setter def y(self, unsigned int y): self._pvt_ptr[0].clusterDim.y = y @property def z(self): return self._pvt_ptr[0].clusterDim.z @z.setter def z(self, unsigned int z): self._pvt_ptr[0].clusterDim.z = z cdef class anon_struct23: """ Attributes ---------- event : cudaEvent_t flags : int triggerAtBlockStart : int Methods ------- getPtr() Get memory address of class instance """ def __cinit__(self, void_ptr _ptr): self._pvt_ptr = _ptr def __init__(self, void_ptr _ptr): pass self._event = cudaEvent_t(_ptr=&self._pvt_ptr[0].programmaticEvent.event) def __dealloc__(self): pass def getPtr(self): return &self._pvt_ptr[0].programmaticEvent def __repr__(self): if self._pvt_ptr is not NULL: str_list = [] try: str_list += ['event : ' + str(self.event)] except ValueError: str_list += ['event : '] try: str_list += ['flags : ' + str(self.flags)] except ValueError: str_list += ['flags : '] try: str_list += ['triggerAtBlockStart : ' + str(self.triggerAtBlockStart)] except ValueError: str_list += ['triggerAtBlockStart : '] return '\n'.join(str_list) else: return '' @property def event(self): return self._event @event.setter def event(self, event): cdef cyruntime.cudaEvent_t cyevent if event is None: cyevent = 0 elif isinstance(event, (cudaEvent_t,driver.CUevent)): pevent = int(event) cyevent = pevent else: pevent = int(cudaEvent_t(event)) cyevent = pevent self._event._pvt_ptr[0] = cyevent @property def flags(self): return self._pvt_ptr[0].programmaticEvent.flags @flags.setter def flags(self, int flags): self._pvt_ptr[0].programmaticEvent.flags = flags @property def triggerAtBlockStart(self): return self._pvt_ptr[0].programmaticEvent.triggerAtBlockStart @triggerAtBlockStart.setter def triggerAtBlockStart(self, int triggerAtBlockStart): self._pvt_ptr[0].programmaticEvent.triggerAtBlockStart = triggerAtBlockStart cdef class anon_struct24: """ Attributes ---------- x : unsigned int y : unsigned int z : unsigned int Methods ------- getPtr() Get memory address of class instance """ def __cinit__(self, void_ptr _ptr): self._pvt_ptr = _ptr def __init__(self, void_ptr _ptr): pass def __dealloc__(self): pass def getPtr(self): return &self._pvt_ptr[0].preferredClusterDim def __repr__(self): if self._pvt_ptr is not NULL: str_list = [] try: str_list += ['x : ' + str(self.x)] except ValueError: str_list += ['x : '] try: str_list += ['y : ' + str(self.y)] except ValueError: str_list += ['y : '] try: str_list += ['z : ' + str(self.z)] except ValueError: str_list += ['z : '] return '\n'.join(str_list) else: return '' @property def x(self): return self._pvt_ptr[0].preferredClusterDim.x @x.setter def x(self, unsigned int x): self._pvt_ptr[0].preferredClusterDim.x = x @property def y(self): return self._pvt_ptr[0].preferredClusterDim.y @y.setter def y(self, unsigned int y): self._pvt_ptr[0].preferredClusterDim.y = y @property def z(self): return self._pvt_ptr[0].preferredClusterDim.z @z.setter def z(self, unsigned int z): self._pvt_ptr[0].preferredClusterDim.z = z cdef class anon_struct25: """ Attributes ---------- event : cudaEvent_t flags : int Methods ------- getPtr() Get memory address of class instance """ def __cinit__(self, void_ptr _ptr): self._pvt_ptr = _ptr def __init__(self, void_ptr _ptr): pass self._event = cudaEvent_t(_ptr=&self._pvt_ptr[0].launchCompletionEvent.event) def __dealloc__(self): pass def getPtr(self): return &self._pvt_ptr[0].launchCompletionEvent def __repr__(self): if self._pvt_ptr is not NULL: str_list = [] try: str_list += ['event : ' + str(self.event)] except ValueError: str_list += ['event : '] try: str_list += ['flags : ' + str(self.flags)] except ValueError: str_list += ['flags : '] return '\n'.join(str_list) else: return '' @property def event(self): return self._event @event.setter def event(self, event): cdef cyruntime.cudaEvent_t cyevent if event is None: cyevent = 0 elif isinstance(event, (cudaEvent_t,driver.CUevent)): pevent = int(event) cyevent = pevent else: pevent = int(cudaEvent_t(event)) cyevent = pevent self._event._pvt_ptr[0] = cyevent @property def flags(self): return self._pvt_ptr[0].launchCompletionEvent.flags @flags.setter def flags(self, int flags): self._pvt_ptr[0].launchCompletionEvent.flags = flags cdef class anon_struct26: """ Attributes ---------- deviceUpdatable : int devNode : cudaGraphDeviceNode_t Methods ------- getPtr() Get memory address of class instance """ def __cinit__(self, void_ptr _ptr): self._pvt_ptr = _ptr def __init__(self, void_ptr _ptr): pass self._devNode = cudaGraphDeviceNode_t(_ptr=&self._pvt_ptr[0].deviceUpdatableKernelNode.devNode) def __dealloc__(self): pass def getPtr(self): return &self._pvt_ptr[0].deviceUpdatableKernelNode def __repr__(self): if self._pvt_ptr is not NULL: str_list = [] try: str_list += ['deviceUpdatable : ' + str(self.deviceUpdatable)] except ValueError: str_list += ['deviceUpdatable : '] try: str_list += ['devNode : ' + str(self.devNode)] except ValueError: str_list += ['devNode : '] return '\n'.join(str_list) else: return '' @property def deviceUpdatable(self): return self._pvt_ptr[0].deviceUpdatableKernelNode.deviceUpdatable @deviceUpdatable.setter def deviceUpdatable(self, int deviceUpdatable): self._pvt_ptr[0].deviceUpdatableKernelNode.deviceUpdatable = deviceUpdatable @property def devNode(self): return self._devNode @devNode.setter def devNode(self, devNode): cdef cyruntime.cudaGraphDeviceNode_t cydevNode if devNode is None: cydevNode = 0 elif isinstance(devNode, (cudaGraphDeviceNode_t,)): pdevNode = int(devNode) cydevNode = pdevNode else: pdevNode = int(cudaGraphDeviceNode_t(devNode)) cydevNode = pdevNode self._devNode._pvt_ptr[0] = cydevNode cdef class cudaLaunchAttributeValue: """ Launch attributes union; used as value field of ::cudaLaunchAttribute Attributes ---------- pad : bytes accessPolicyWindow : cudaAccessPolicyWindow Value of launch attribute cudaLaunchAttributeAccessPolicyWindow. cooperative : int Value of launch attribute cudaLaunchAttributeCooperative. Nonzero indicates a cooperative kernel (see cudaLaunchCooperativeKernel). syncPolicy : cudaSynchronizationPolicy Value of launch attribute cudaLaunchAttributeSynchronizationPolicy. ::cudaSynchronizationPolicy for work queued up in this stream. clusterDim : anon_struct22 Value of launch attribute cudaLaunchAttributeClusterDimension that represents the desired cluster dimensions for the kernel. Opaque type with the following fields: - `x` - The X dimension of the cluster, in blocks. Must be a divisor of the grid X dimension. - `y` - The Y dimension of the cluster, in blocks. Must be a divisor of the grid Y dimension. - `z` - The Z dimension of the cluster, in blocks. Must be a divisor of the grid Z dimension. clusterSchedulingPolicyPreference : cudaClusterSchedulingPolicy Value of launch attribute cudaLaunchAttributeClusterSchedulingPolicyPreference. Cluster scheduling policy preference for the kernel. programmaticStreamSerializationAllowed : int Value of launch attribute cudaLaunchAttributeProgrammaticStreamSerialization. programmaticEvent : anon_struct23 Value of launch attribute cudaLaunchAttributeProgrammaticEvent with the following fields: - `cudaEvent_t` event - Event to fire when all blocks trigger it. - `int` flags; - Event record flags, see cudaEventRecordWithFlags. Does not accept cudaEventRecordExternal. - `int` triggerAtBlockStart - If this is set to non-0, each block launch will automatically trigger the event. priority : int Value of launch attribute cudaLaunchAttributePriority. Execution priority of the kernel. memSyncDomainMap : cudaLaunchMemSyncDomainMap Value of launch attribute cudaLaunchAttributeMemSyncDomainMap. See ::cudaLaunchMemSyncDomainMap. memSyncDomain : cudaLaunchMemSyncDomain Value of launch attribute cudaLaunchAttributeMemSyncDomain. See cudaLaunchMemSyncDomain. preferredClusterDim : anon_struct24 Value of launch attribute cudaLaunchAttributePreferredClusterDimension that represents the desired preferred cluster dimensions for the kernel. Opaque type with the following fields: - `x` - The X dimension of the preferred cluster, in blocks. Must be a divisor of the grid X dimension, and must be a multiple of the `x` field of cudaLaunchAttributeValue::clusterDim. - `y` - The Y dimension of the preferred cluster, in blocks. Must be a divisor of the grid Y dimension, and must be a multiple of the `y` field of cudaLaunchAttributeValue::clusterDim. - `z` - The Z dimension of the preferred cluster, in blocks. Must be equal to the `z` field of cudaLaunchAttributeValue::clusterDim. launchCompletionEvent : anon_struct25 Value of launch attribute cudaLaunchAttributeLaunchCompletionEvent with the following fields: - `cudaEvent_t` event - Event to fire when the last block launches. - `int` flags - Event record flags, see cudaEventRecordWithFlags. Does not accept cudaEventRecordExternal. deviceUpdatableKernelNode : anon_struct26 Value of launch attribute cudaLaunchAttributeDeviceUpdatableKernelNode with the following fields: - `int` deviceUpdatable - Whether or not the resulting kernel node should be device-updatable. - `cudaGraphDeviceNode_t` devNode - Returns a handle to pass to the various device-side update functions. sharedMemCarveout : unsigned int Value of launch attribute cudaLaunchAttributePreferredSharedMemoryCarveout. Methods ------- getPtr() Get memory address of class instance """ def __cinit__(self, void_ptr _ptr = 0): if _ptr == 0: self._pvt_ptr = &self._pvt_val else: self._pvt_ptr = _ptr def __init__(self, void_ptr _ptr = 0): pass self._accessPolicyWindow = cudaAccessPolicyWindow(_ptr=&self._pvt_ptr[0].accessPolicyWindow) self._clusterDim = anon_struct22(_ptr=self._pvt_ptr) self._programmaticEvent = anon_struct23(_ptr=self._pvt_ptr) self._memSyncDomainMap = cudaLaunchMemSyncDomainMap(_ptr=&self._pvt_ptr[0].memSyncDomainMap) self._preferredClusterDim = anon_struct24(_ptr=self._pvt_ptr) self._launchCompletionEvent = anon_struct25(_ptr=self._pvt_ptr) self._deviceUpdatableKernelNode = anon_struct26(_ptr=self._pvt_ptr) def __dealloc__(self): pass def getPtr(self): return self._pvt_ptr def __repr__(self): if self._pvt_ptr is not NULL: str_list = [] try: str_list += ['pad : ' + str(self.pad)] except ValueError: str_list += ['pad : '] try: str_list += ['accessPolicyWindow :\n' + '\n'.join([' ' + line for line in str(self.accessPolicyWindow).splitlines()])] except ValueError: str_list += ['accessPolicyWindow : '] try: str_list += ['cooperative : ' + str(self.cooperative)] except ValueError: str_list += ['cooperative : '] try: str_list += ['syncPolicy : ' + str(self.syncPolicy)] except ValueError: str_list += ['syncPolicy : '] try: str_list += ['clusterDim :\n' + '\n'.join([' ' + line for line in str(self.clusterDim).splitlines()])] except ValueError: str_list += ['clusterDim : '] try: str_list += ['clusterSchedulingPolicyPreference : ' + str(self.clusterSchedulingPolicyPreference)] except ValueError: str_list += ['clusterSchedulingPolicyPreference : '] try: str_list += ['programmaticStreamSerializationAllowed : ' + str(self.programmaticStreamSerializationAllowed)] except ValueError: str_list += ['programmaticStreamSerializationAllowed : '] try: str_list += ['programmaticEvent :\n' + '\n'.join([' ' + line for line in str(self.programmaticEvent).splitlines()])] except ValueError: str_list += ['programmaticEvent : '] try: str_list += ['priority : ' + str(self.priority)] except ValueError: str_list += ['priority : '] try: str_list += ['memSyncDomainMap :\n' + '\n'.join([' ' + line for line in str(self.memSyncDomainMap).splitlines()])] except ValueError: str_list += ['memSyncDomainMap : '] try: str_list += ['memSyncDomain : ' + str(self.memSyncDomain)] except ValueError: str_list += ['memSyncDomain : '] try: str_list += ['preferredClusterDim :\n' + '\n'.join([' ' + line for line in str(self.preferredClusterDim).splitlines()])] except ValueError: str_list += ['preferredClusterDim : '] try: str_list += ['launchCompletionEvent :\n' + '\n'.join([' ' + line for line in str(self.launchCompletionEvent).splitlines()])] except ValueError: str_list += ['launchCompletionEvent : '] try: str_list += ['deviceUpdatableKernelNode :\n' + '\n'.join([' ' + line for line in str(self.deviceUpdatableKernelNode).splitlines()])] except ValueError: str_list += ['deviceUpdatableKernelNode : '] try: str_list += ['sharedMemCarveout : ' + str(self.sharedMemCarveout)] except ValueError: str_list += ['sharedMemCarveout : '] return '\n'.join(str_list) else: return '' @property def pad(self): return PyBytes_FromStringAndSize(self._pvt_ptr[0].pad, 64) @pad.setter def pad(self, pad): if len(pad) != 64: raise ValueError("pad length must be 64, is " + str(len(pad))) if CHAR_MIN == 0: for i, b in enumerate(pad): if b < 0 and b > -129: b = b + 256 self._pvt_ptr[0].pad[i] = b else: for i, b in enumerate(pad): if b > 127 and b < 256: b = b - 256 self._pvt_ptr[0].pad[i] = b @property def accessPolicyWindow(self): return self._accessPolicyWindow @accessPolicyWindow.setter def accessPolicyWindow(self, accessPolicyWindow not None : cudaAccessPolicyWindow): string.memcpy(&self._pvt_ptr[0].accessPolicyWindow, accessPolicyWindow.getPtr(), sizeof(self._pvt_ptr[0].accessPolicyWindow)) @property def cooperative(self): return self._pvt_ptr[0].cooperative @cooperative.setter def cooperative(self, int cooperative): self._pvt_ptr[0].cooperative = cooperative @property def syncPolicy(self): if self._pvt_ptr[0].syncPolicy not in _dict_cudaSynchronizationPolicy: return None return _dict_cudaSynchronizationPolicy[self._pvt_ptr[0].syncPolicy] @syncPolicy.setter def syncPolicy(self, syncPolicy not None : cudaSynchronizationPolicy): self._pvt_ptr[0].syncPolicy = syncPolicy.value @property def clusterDim(self): return self._clusterDim @clusterDim.setter def clusterDim(self, clusterDim not None : anon_struct22): string.memcpy(&self._pvt_ptr[0].clusterDim, clusterDim.getPtr(), sizeof(self._pvt_ptr[0].clusterDim)) @property def clusterSchedulingPolicyPreference(self): if self._pvt_ptr[0].clusterSchedulingPolicyPreference not in _dict_cudaClusterSchedulingPolicy: return None return _dict_cudaClusterSchedulingPolicy[self._pvt_ptr[0].clusterSchedulingPolicyPreference] @clusterSchedulingPolicyPreference.setter def clusterSchedulingPolicyPreference(self, clusterSchedulingPolicyPreference not None : cudaClusterSchedulingPolicy): self._pvt_ptr[0].clusterSchedulingPolicyPreference = clusterSchedulingPolicyPreference.value @property def programmaticStreamSerializationAllowed(self): return self._pvt_ptr[0].programmaticStreamSerializationAllowed @programmaticStreamSerializationAllowed.setter def programmaticStreamSerializationAllowed(self, int programmaticStreamSerializationAllowed): self._pvt_ptr[0].programmaticStreamSerializationAllowed = programmaticStreamSerializationAllowed @property def programmaticEvent(self): return self._programmaticEvent @programmaticEvent.setter def programmaticEvent(self, programmaticEvent not None : anon_struct23): string.memcpy(&self._pvt_ptr[0].programmaticEvent, programmaticEvent.getPtr(), sizeof(self._pvt_ptr[0].programmaticEvent)) @property def priority(self): return self._pvt_ptr[0].priority @priority.setter def priority(self, int priority): self._pvt_ptr[0].priority = priority @property def memSyncDomainMap(self): return self._memSyncDomainMap @memSyncDomainMap.setter def memSyncDomainMap(self, memSyncDomainMap not None : cudaLaunchMemSyncDomainMap): string.memcpy(&self._pvt_ptr[0].memSyncDomainMap, memSyncDomainMap.getPtr(), sizeof(self._pvt_ptr[0].memSyncDomainMap)) @property def memSyncDomain(self): if self._pvt_ptr[0].memSyncDomain not in _dict_cudaLaunchMemSyncDomain: return None return _dict_cudaLaunchMemSyncDomain[self._pvt_ptr[0].memSyncDomain] @memSyncDomain.setter def memSyncDomain(self, memSyncDomain not None : cudaLaunchMemSyncDomain): self._pvt_ptr[0].memSyncDomain = memSyncDomain.value @property def preferredClusterDim(self): return self._preferredClusterDim @preferredClusterDim.setter def preferredClusterDim(self, preferredClusterDim not None : anon_struct24): string.memcpy(&self._pvt_ptr[0].preferredClusterDim, preferredClusterDim.getPtr(), sizeof(self._pvt_ptr[0].preferredClusterDim)) @property def launchCompletionEvent(self): return self._launchCompletionEvent @launchCompletionEvent.setter def launchCompletionEvent(self, launchCompletionEvent not None : anon_struct25): string.memcpy(&self._pvt_ptr[0].launchCompletionEvent, launchCompletionEvent.getPtr(), sizeof(self._pvt_ptr[0].launchCompletionEvent)) @property def deviceUpdatableKernelNode(self): return self._deviceUpdatableKernelNode @deviceUpdatableKernelNode.setter def deviceUpdatableKernelNode(self, deviceUpdatableKernelNode not None : anon_struct26): string.memcpy(&self._pvt_ptr[0].deviceUpdatableKernelNode, deviceUpdatableKernelNode.getPtr(), sizeof(self._pvt_ptr[0].deviceUpdatableKernelNode)) @property def sharedMemCarveout(self): return self._pvt_ptr[0].sharedMemCarveout @sharedMemCarveout.setter def sharedMemCarveout(self, unsigned int sharedMemCarveout): self._pvt_ptr[0].sharedMemCarveout = sharedMemCarveout cdef class cudaLaunchAttribute_st: """ Launch attribute Attributes ---------- id : cudaLaunchAttributeID Attribute to set val : cudaLaunchAttributeValue Value of the attribute Methods ------- getPtr() Get memory address of class instance """ def __cinit__(self, void_ptr _ptr = 0): if _ptr == 0: self._pvt_ptr = &self._pvt_val else: self._pvt_ptr = _ptr def __init__(self, void_ptr _ptr = 0): pass self._val = cudaLaunchAttributeValue(_ptr=&self._pvt_ptr[0].val) def __dealloc__(self): pass def getPtr(self): return self._pvt_ptr def __repr__(self): if self._pvt_ptr is not NULL: str_list = [] try: str_list += ['id : ' + str(self.id)] except ValueError: str_list += ['id : '] try: str_list += ['val :\n' + '\n'.join([' ' + line for line in str(self.val).splitlines()])] except ValueError: str_list += ['val : '] return '\n'.join(str_list) else: return '' @property def id(self): if self._pvt_ptr[0].id not in _dict_cudaLaunchAttributeID: return None return _dict_cudaLaunchAttributeID[self._pvt_ptr[0].id] @id.setter def id(self, id not None : cudaLaunchAttributeID): self._pvt_ptr[0].id = id.value @property def val(self): return self._val @val.setter def val(self, val not None : cudaLaunchAttributeValue): string.memcpy(&self._pvt_ptr[0].val, val.getPtr(), sizeof(self._pvt_ptr[0].val)) cdef class anon_struct27: """ Attributes ---------- bytesOverBudget : unsigned long long Methods ------- getPtr() Get memory address of class instance """ def __cinit__(self, void_ptr _ptr): self._pvt_ptr = _ptr def __init__(self, void_ptr _ptr): pass def __dealloc__(self): pass def getPtr(self): return &self._pvt_ptr[0].info.overBudget def __repr__(self): if self._pvt_ptr is not NULL: str_list = [] try: str_list += ['bytesOverBudget : ' + str(self.bytesOverBudget)] except ValueError: str_list += ['bytesOverBudget : '] return '\n'.join(str_list) else: return '' @property def bytesOverBudget(self): return self._pvt_ptr[0].info.overBudget.bytesOverBudget @bytesOverBudget.setter def bytesOverBudget(self, unsigned long long bytesOverBudget): self._pvt_ptr[0].info.overBudget.bytesOverBudget = bytesOverBudget cdef class anon_union10: """ Attributes ---------- overBudget : anon_struct27 Methods ------- getPtr() Get memory address of class instance """ def __cinit__(self, void_ptr _ptr): self._pvt_ptr = _ptr def __init__(self, void_ptr _ptr): pass self._overBudget = anon_struct27(_ptr=self._pvt_ptr) def __dealloc__(self): pass def getPtr(self): return &self._pvt_ptr[0].info def __repr__(self): if self._pvt_ptr is not NULL: str_list = [] try: str_list += ['overBudget :\n' + '\n'.join([' ' + line for line in str(self.overBudget).splitlines()])] except ValueError: str_list += ['overBudget : '] return '\n'.join(str_list) else: return '' @property def overBudget(self): return self._overBudget @overBudget.setter def overBudget(self, overBudget not None : anon_struct27): string.memcpy(&self._pvt_ptr[0].info.overBudget, overBudget.getPtr(), sizeof(self._pvt_ptr[0].info.overBudget)) cdef class cudaAsyncNotificationInfo: """ Information describing an async notification event Attributes ---------- type : cudaAsyncNotificationType The type of notification being sent info : anon_union10 Information about the notification. `typename` must be checked in order to interpret this field. Methods ------- getPtr() Get memory address of class instance """ def __cinit__(self, void_ptr _ptr = 0): if _ptr == 0: self._val_ptr = calloc(1, sizeof(cyruntime.cudaAsyncNotificationInfo)) self._pvt_ptr = self._val_ptr else: self._pvt_ptr = _ptr def __init__(self, void_ptr _ptr = 0): pass self._info = anon_union10(_ptr=self._pvt_ptr) def __dealloc__(self): if self._val_ptr is not NULL: free(self._val_ptr) def getPtr(self): return self._pvt_ptr def __repr__(self): if self._pvt_ptr is not NULL: str_list = [] try: str_list += ['type : ' + str(self.type)] except ValueError: str_list += ['type : '] try: str_list += ['info :\n' + '\n'.join([' ' + line for line in str(self.info).splitlines()])] except ValueError: str_list += ['info : '] return '\n'.join(str_list) else: return '' @property def type(self): if self._pvt_ptr[0].type not in _dict_cudaAsyncNotificationType: return None return _dict_cudaAsyncNotificationType[self._pvt_ptr[0].type] @type.setter def type(self, type not None : cudaAsyncNotificationType): self._pvt_ptr[0].type = type.value @property def info(self): return self._info @info.setter def info(self, info not None : anon_union10): string.memcpy(&self._pvt_ptr[0].info, info.getPtr(), sizeof(self._pvt_ptr[0].info)) cdef class cudaTextureDesc: """ CUDA texture descriptor Attributes ---------- addressMode : list[cudaTextureAddressMode] Texture address mode for up to 3 dimensions filterMode : cudaTextureFilterMode Texture filter mode readMode : cudaTextureReadMode Texture read mode sRGB : int Perform sRGB->linear conversion during texture read borderColor : list[float] Texture Border Color normalizedCoords : int Indicates whether texture reads are normalized or not maxAnisotropy : unsigned int Limit to the anisotropy ratio mipmapFilterMode : cudaTextureFilterMode Mipmap filter mode mipmapLevelBias : float Offset applied to the supplied mipmap level minMipmapLevelClamp : float Lower end of the mipmap level range to clamp access to maxMipmapLevelClamp : float Upper end of the mipmap level range to clamp access to disableTrilinearOptimization : int Disable any trilinear filtering optimizations. seamlessCubemap : int Enable seamless cube map filtering. Methods ------- getPtr() Get memory address of class instance """ def __cinit__(self, void_ptr _ptr = 0): if _ptr == 0: self._pvt_ptr = &self._pvt_val else: self._pvt_ptr = _ptr def __init__(self, void_ptr _ptr = 0): pass def __dealloc__(self): pass def getPtr(self): return self._pvt_ptr def __repr__(self): if self._pvt_ptr is not NULL: str_list = [] try: str_list += ['addressMode : ' + str(self.addressMode)] except ValueError: str_list += ['addressMode : '] try: str_list += ['filterMode : ' + str(self.filterMode)] except ValueError: str_list += ['filterMode : '] try: str_list += ['readMode : ' + str(self.readMode)] except ValueError: str_list += ['readMode : '] try: str_list += ['sRGB : ' + str(self.sRGB)] except ValueError: str_list += ['sRGB : '] try: str_list += ['borderColor : ' + str(self.borderColor)] except ValueError: str_list += ['borderColor : '] try: str_list += ['normalizedCoords : ' + str(self.normalizedCoords)] except ValueError: str_list += ['normalizedCoords : '] try: str_list += ['maxAnisotropy : ' + str(self.maxAnisotropy)] except ValueError: str_list += ['maxAnisotropy : '] try: str_list += ['mipmapFilterMode : ' + str(self.mipmapFilterMode)] except ValueError: str_list += ['mipmapFilterMode : '] try: str_list += ['mipmapLevelBias : ' + str(self.mipmapLevelBias)] except ValueError: str_list += ['mipmapLevelBias : '] try: str_list += ['minMipmapLevelClamp : ' + str(self.minMipmapLevelClamp)] except ValueError: str_list += ['minMipmapLevelClamp : '] try: str_list += ['maxMipmapLevelClamp : ' + str(self.maxMipmapLevelClamp)] except ValueError: str_list += ['maxMipmapLevelClamp : '] try: str_list += ['disableTrilinearOptimization : ' + str(self.disableTrilinearOptimization)] except ValueError: str_list += ['disableTrilinearOptimization : '] try: str_list += ['seamlessCubemap : ' + str(self.seamlessCubemap)] except ValueError: str_list += ['seamlessCubemap : '] return '\n'.join(str_list) else: return '' @property def addressMode(self): return [_dict_cudaTextureAddressMode[_x] if _x in _dict_cudaTextureAddressMode else None for _x in list(self._pvt_ptr[0].addressMode)] @addressMode.setter def addressMode(self, addressMode): self._pvt_ptr[0].addressMode = [_x.value for _x in addressMode] @property def filterMode(self): if self._pvt_ptr[0].filterMode not in _dict_cudaTextureFilterMode: return None return _dict_cudaTextureFilterMode[self._pvt_ptr[0].filterMode] @filterMode.setter def filterMode(self, filterMode not None : cudaTextureFilterMode): self._pvt_ptr[0].filterMode = filterMode.value @property def readMode(self): if self._pvt_ptr[0].readMode not in _dict_cudaTextureReadMode: return None return _dict_cudaTextureReadMode[self._pvt_ptr[0].readMode] @readMode.setter def readMode(self, readMode not None : cudaTextureReadMode): self._pvt_ptr[0].readMode = readMode.value @property def sRGB(self): return self._pvt_ptr[0].sRGB @sRGB.setter def sRGB(self, int sRGB): self._pvt_ptr[0].sRGB = sRGB @property def borderColor(self): return self._pvt_ptr[0].borderColor @borderColor.setter def borderColor(self, borderColor): self._pvt_ptr[0].borderColor = borderColor @property def normalizedCoords(self): return self._pvt_ptr[0].normalizedCoords @normalizedCoords.setter def normalizedCoords(self, int normalizedCoords): self._pvt_ptr[0].normalizedCoords = normalizedCoords @property def maxAnisotropy(self): return self._pvt_ptr[0].maxAnisotropy @maxAnisotropy.setter def maxAnisotropy(self, unsigned int maxAnisotropy): self._pvt_ptr[0].maxAnisotropy = maxAnisotropy @property def mipmapFilterMode(self): if self._pvt_ptr[0].mipmapFilterMode not in _dict_cudaTextureFilterMode: return None return _dict_cudaTextureFilterMode[self._pvt_ptr[0].mipmapFilterMode] @mipmapFilterMode.setter def mipmapFilterMode(self, mipmapFilterMode not None : cudaTextureFilterMode): self._pvt_ptr[0].mipmapFilterMode = mipmapFilterMode.value @property def mipmapLevelBias(self): return self._pvt_ptr[0].mipmapLevelBias @mipmapLevelBias.setter def mipmapLevelBias(self, float mipmapLevelBias): self._pvt_ptr[0].mipmapLevelBias = mipmapLevelBias @property def minMipmapLevelClamp(self): return self._pvt_ptr[0].minMipmapLevelClamp @minMipmapLevelClamp.setter def minMipmapLevelClamp(self, float minMipmapLevelClamp): self._pvt_ptr[0].minMipmapLevelClamp = minMipmapLevelClamp @property def maxMipmapLevelClamp(self): return self._pvt_ptr[0].maxMipmapLevelClamp @maxMipmapLevelClamp.setter def maxMipmapLevelClamp(self, float maxMipmapLevelClamp): self._pvt_ptr[0].maxMipmapLevelClamp = maxMipmapLevelClamp @property def disableTrilinearOptimization(self): return self._pvt_ptr[0].disableTrilinearOptimization @disableTrilinearOptimization.setter def disableTrilinearOptimization(self, int disableTrilinearOptimization): self._pvt_ptr[0].disableTrilinearOptimization = disableTrilinearOptimization @property def seamlessCubemap(self): return self._pvt_ptr[0].seamlessCubemap @seamlessCubemap.setter def seamlessCubemap(self, int seamlessCubemap): self._pvt_ptr[0].seamlessCubemap = seamlessCubemap cdef class cudaEglPlaneDesc_st: """ CUDA EGL Plane Descriptor - structure defining each plane of a CUDA EGLFrame Attributes ---------- width : unsigned int Width of plane height : unsigned int Height of plane depth : unsigned int Depth of plane pitch : unsigned int Pitch of plane numChannels : unsigned int Number of channels for the plane channelDesc : cudaChannelFormatDesc Channel Format Descriptor reserved : list[unsigned int] Reserved for future use Methods ------- getPtr() Get memory address of class instance """ def __cinit__(self, void_ptr _ptr = 0): if _ptr == 0: self._pvt_ptr = &self._pvt_val else: self._pvt_ptr = _ptr def __init__(self, void_ptr _ptr = 0): pass self._channelDesc = cudaChannelFormatDesc(_ptr=&self._pvt_ptr[0].channelDesc) def __dealloc__(self): pass def getPtr(self): return self._pvt_ptr def __repr__(self): if self._pvt_ptr is not NULL: str_list = [] try: str_list += ['width : ' + str(self.width)] except ValueError: str_list += ['width : '] try: str_list += ['height : ' + str(self.height)] except ValueError: str_list += ['height : '] try: str_list += ['depth : ' + str(self.depth)] except ValueError: str_list += ['depth : '] try: str_list += ['pitch : ' + str(self.pitch)] except ValueError: str_list += ['pitch : '] try: str_list += ['numChannels : ' + str(self.numChannels)] except ValueError: str_list += ['numChannels : '] try: str_list += ['channelDesc :\n' + '\n'.join([' ' + line for line in str(self.channelDesc).splitlines()])] except ValueError: str_list += ['channelDesc : '] try: str_list += ['reserved : ' + str(self.reserved)] except ValueError: str_list += ['reserved : '] return '\n'.join(str_list) else: return '' @property def width(self): return self._pvt_ptr[0].width @width.setter def width(self, unsigned int width): self._pvt_ptr[0].width = width @property def height(self): return self._pvt_ptr[0].height @height.setter def height(self, unsigned int height): self._pvt_ptr[0].height = height @property def depth(self): return self._pvt_ptr[0].depth @depth.setter def depth(self, unsigned int depth): self._pvt_ptr[0].depth = depth @property def pitch(self): return self._pvt_ptr[0].pitch @pitch.setter def pitch(self, unsigned int pitch): self._pvt_ptr[0].pitch = pitch @property def numChannels(self): return self._pvt_ptr[0].numChannels @numChannels.setter def numChannels(self, unsigned int numChannels): self._pvt_ptr[0].numChannels = numChannels @property def channelDesc(self): return self._channelDesc @channelDesc.setter def channelDesc(self, channelDesc not None : cudaChannelFormatDesc): string.memcpy(&self._pvt_ptr[0].channelDesc, channelDesc.getPtr(), sizeof(self._pvt_ptr[0].channelDesc)) @property def reserved(self): return self._pvt_ptr[0].reserved @reserved.setter def reserved(self, reserved): self._pvt_ptr[0].reserved = reserved cdef class anon_union11: """ Attributes ---------- pArray : list[cudaArray_t] pPitch : list[cudaPitchedPtr] Methods ------- getPtr() Get memory address of class instance """ def __cinit__(self, void_ptr _ptr): self._pvt_ptr = _ptr def __init__(self, void_ptr _ptr): pass def __dealloc__(self): pass def getPtr(self): return &self._pvt_ptr[0].frame def __repr__(self): if self._pvt_ptr is not NULL: str_list = [] try: str_list += ['pArray : ' + str(self.pArray)] except ValueError: str_list += ['pArray : '] try: str_list += ['pPitch :\n' + '\n'.join([' ' + line for line in str(self.pPitch).splitlines()])] except ValueError: str_list += ['pPitch : '] return '\n'.join(str_list) else: return '' @property def pArray(self): return [cudaArray_t(init_value=_pArray) for _pArray in self._pvt_ptr[0].frame.pArray] @pArray.setter def pArray(self, pArray : list[cudaArray_t]): if len(pArray) != 3: raise IndexError('not enough values found during array assignment, expected 3, got', len(pArray)) pArray = [int(_pArray) for _pArray in pArray] for _idx, _pArray in enumerate(pArray): self._pvt_ptr[0].frame.pArray[_idx] = _pArray @property def pPitch(self): out_pPitch = [cudaPitchedPtr() for _pPitch in self._pvt_ptr[0].frame.pPitch] for _idx in range(len(out_pPitch)): string.memcpy(out_pPitch[_idx].getPtr(), &self._pvt_ptr[0].frame.pPitch[_idx], sizeof(cyruntime.cudaPitchedPtr)) return out_pPitch @pPitch.setter def pPitch(self, pPitch : list[cudaPitchedPtr]): if len(pPitch) != 3: raise IndexError('not enough values found during array assignment, expected 3, got', len(pPitch)) for _idx in range(len(pPitch)): string.memcpy(&self._pvt_ptr[0].frame.pPitch[_idx], pPitch[_idx].getPtr(), sizeof(cyruntime.cudaPitchedPtr)) cdef class cudaEglFrame_st: """ CUDA EGLFrame Descriptor - structure defining one frame of EGL. Each frame may contain one or more planes depending on whether the surface is Multiplanar or not. Each plane of EGLFrame is represented by cudaEglPlaneDesc which is defined as: typedefstructcudaEglPlaneDesc_st unsignedintwidth; unsignedintheight; unsignedintdepth; unsignedintpitch; unsignedintnumChannels; structcudaChannelFormatDescchannelDesc; unsignedintreserved[4]; cudaEglPlaneDesc; Attributes ---------- frame : anon_union11 planeDesc : list[cudaEglPlaneDesc] CUDA EGL Plane Descriptor cudaEglPlaneDesc planeCount : unsigned int Number of planes frameType : cudaEglFrameType Array or Pitch eglColorFormat : cudaEglColorFormat CUDA EGL Color Format Methods ------- getPtr() Get memory address of class instance """ def __cinit__(self, void_ptr _ptr = 0): if _ptr == 0: self._val_ptr = calloc(1, sizeof(cyruntime.cudaEglFrame_st)) self._pvt_ptr = self._val_ptr else: self._pvt_ptr = _ptr def __init__(self, void_ptr _ptr = 0): pass self._frame = anon_union11(_ptr=self._pvt_ptr) def __dealloc__(self): if self._val_ptr is not NULL: free(self._val_ptr) def getPtr(self): return self._pvt_ptr def __repr__(self): if self._pvt_ptr is not NULL: str_list = [] try: str_list += ['frame :\n' + '\n'.join([' ' + line for line in str(self.frame).splitlines()])] except ValueError: str_list += ['frame : '] try: str_list += ['planeDesc :\n' + '\n'.join([' ' + line for line in str(self.planeDesc).splitlines()])] except ValueError: str_list += ['planeDesc : '] try: str_list += ['planeCount : ' + str(self.planeCount)] except ValueError: str_list += ['planeCount : '] try: str_list += ['frameType : ' + str(self.frameType)] except ValueError: str_list += ['frameType : '] try: str_list += ['eglColorFormat : ' + str(self.eglColorFormat)] except ValueError: str_list += ['eglColorFormat : '] return '\n'.join(str_list) else: return '' @property def frame(self): return self._frame @frame.setter def frame(self, frame not None : anon_union11): string.memcpy(&self._pvt_ptr[0].frame, frame.getPtr(), sizeof(self._pvt_ptr[0].frame)) @property def planeDesc(self): out_planeDesc = [cudaEglPlaneDesc() for _planeDesc in self._pvt_ptr[0].planeDesc] for _idx in range(len(out_planeDesc)): string.memcpy(out_planeDesc[_idx].getPtr(), &self._pvt_ptr[0].planeDesc[_idx], sizeof(cyruntime.cudaEglPlaneDesc)) return out_planeDesc @planeDesc.setter def planeDesc(self, planeDesc : list[cudaEglPlaneDesc]): if len(planeDesc) != 3: raise IndexError('not enough values found during array assignment, expected 3, got', len(planeDesc)) for _idx in range(len(planeDesc)): string.memcpy(&self._pvt_ptr[0].planeDesc[_idx], planeDesc[_idx].getPtr(), sizeof(cyruntime.cudaEglPlaneDesc)) @property def planeCount(self): return self._pvt_ptr[0].planeCount @planeCount.setter def planeCount(self, unsigned int planeCount): self._pvt_ptr[0].planeCount = planeCount @property def frameType(self): if self._pvt_ptr[0].frameType not in _dict_cudaEglFrameType: return None return _dict_cudaEglFrameType[self._pvt_ptr[0].frameType] @frameType.setter def frameType(self, frameType not None : cudaEglFrameType): self._pvt_ptr[0].frameType = frameType.value @property def eglColorFormat(self): if self._pvt_ptr[0].eglColorFormat not in _dict_cudaEglColorFormat: return None return _dict_cudaEglColorFormat[self._pvt_ptr[0].eglColorFormat] @eglColorFormat.setter def eglColorFormat(self, eglColorFormat not None : cudaEglColorFormat): self._pvt_ptr[0].eglColorFormat = eglColorFormat.value cdef class cudaGraphConditionalHandle: """ CUDA handle for conditional graph nodes Methods ------- getPtr() Get memory address of class instance """ def __cinit__(self, unsigned long long init_value = 0, void_ptr _ptr = 0): if _ptr == 0: self._pvt_ptr = &self._pvt_val else: self._pvt_ptr = _ptr if init_value: self._pvt_ptr[0] = init_value def __dealloc__(self): pass def __repr__(self): return '' def __int__(self): return self._pvt_ptr[0] def getPtr(self): return self._pvt_ptr cdef class cudaSurfaceObject_t: """ An opaque value that represents a CUDA Surface object Methods ------- getPtr() Get memory address of class instance """ def __cinit__(self, unsigned long long init_value = 0, void_ptr _ptr = 0): if _ptr == 0: self._pvt_ptr = &self._pvt_val else: self._pvt_ptr = _ptr if init_value: self._pvt_ptr[0] = init_value def __dealloc__(self): pass def __repr__(self): return '' def __int__(self): return self._pvt_ptr[0] def getPtr(self): return self._pvt_ptr cdef class cudaTextureObject_t: """ An opaque value that represents a CUDA texture object Methods ------- getPtr() Get memory address of class instance """ def __cinit__(self, unsigned long long init_value = 0, void_ptr _ptr = 0): if _ptr == 0: self._pvt_ptr = &self._pvt_val else: self._pvt_ptr = _ptr if init_value: self._pvt_ptr[0] = init_value def __dealloc__(self): pass def __repr__(self): return '' def __int__(self): return self._pvt_ptr[0] def getPtr(self): return self._pvt_ptr cdef class GLenum: """ Methods ------- getPtr() Get memory address of class instance """ def __cinit__(self, unsigned int init_value = 0, void_ptr _ptr = 0): if _ptr == 0: self._pvt_ptr = &self._pvt_val else: self._pvt_ptr = _ptr if init_value: self._pvt_ptr[0] = init_value def __dealloc__(self): pass def __repr__(self): return '' def __int__(self): return self._pvt_ptr[0] def getPtr(self): return self._pvt_ptr cdef class GLuint: """ Methods ------- getPtr() Get memory address of class instance """ def __cinit__(self, unsigned int init_value = 0, void_ptr _ptr = 0): if _ptr == 0: self._pvt_ptr = &self._pvt_val else: self._pvt_ptr = _ptr if init_value: self._pvt_ptr[0] = init_value def __dealloc__(self): pass def __repr__(self): return '' def __int__(self): return self._pvt_ptr[0] def getPtr(self): return self._pvt_ptr cdef class EGLint: """ Methods ------- getPtr() Get memory address of class instance """ def __cinit__(self, unsigned int init_value = 0, void_ptr _ptr = 0): if _ptr == 0: self._pvt_ptr = &self._pvt_val else: self._pvt_ptr = _ptr if init_value: self._pvt_ptr[0] = init_value def __dealloc__(self): pass def __repr__(self): return '' def __int__(self): return self._pvt_ptr[0] def getPtr(self): return self._pvt_ptr cdef class VdpDevice: """ Methods ------- getPtr() Get memory address of class instance """ def __cinit__(self, uint32_t init_value = 0, void_ptr _ptr = 0): if _ptr == 0: self._pvt_ptr = &self._pvt_val else: self._pvt_ptr = _ptr if init_value: self._pvt_ptr[0] = init_value def __dealloc__(self): pass def __repr__(self): return '' def __int__(self): return self._pvt_ptr[0] def getPtr(self): return self._pvt_ptr cdef class VdpGetProcAddress: """ Methods ------- getPtr() Get memory address of class instance """ def __cinit__(self, unsigned long long init_value = 0, void_ptr _ptr = 0): if _ptr == 0: self._pvt_ptr = &self._pvt_val else: self._pvt_ptr = _ptr if init_value: self._pvt_ptr[0] = init_value def __dealloc__(self): pass def __repr__(self): return '' def __int__(self): return self._pvt_ptr[0] def getPtr(self): return self._pvt_ptr cdef class VdpVideoSurface: """ Methods ------- getPtr() Get memory address of class instance """ def __cinit__(self, uint32_t init_value = 0, void_ptr _ptr = 0): if _ptr == 0: self._pvt_ptr = &self._pvt_val else: self._pvt_ptr = _ptr if init_value: self._pvt_ptr[0] = init_value def __dealloc__(self): pass def __repr__(self): return '' def __int__(self): return self._pvt_ptr[0] def getPtr(self): return self._pvt_ptr cdef class VdpOutputSurface: """ Methods ------- getPtr() Get memory address of class instance """ def __cinit__(self, uint32_t init_value = 0, void_ptr _ptr = 0): if _ptr == 0: self._pvt_ptr = &self._pvt_val else: self._pvt_ptr = _ptr if init_value: self._pvt_ptr[0] = init_value def __dealloc__(self): pass def __repr__(self): return '' def __int__(self): return self._pvt_ptr[0] def getPtr(self): return self._pvt_ptr @cython.embedsignature(True) def cudaDeviceReset(): """ Destroy all allocations and reset all state on the current device in the current process. Explicitly destroys and cleans up all resources associated with the current device in the current process. It is the caller's responsibility to ensure that the resources are not accessed or passed in subsequent API calls and doing so will result in undefined behavior. These resources include CUDA types :py:obj:`~.cudaStream_t`, :py:obj:`~.cudaEvent_t`, :py:obj:`~.cudaArray_t`, :py:obj:`~.cudaMipmappedArray_t`, :py:obj:`~.cudaPitchedPtr`, :py:obj:`~.cudaTextureObject_t`, :py:obj:`~.cudaSurfaceObject_t`, :py:obj:`~.textureReference`, :py:obj:`~.surfaceReference`, :py:obj:`~.cudaExternalMemory_t`, :py:obj:`~.cudaExternalSemaphore_t` and :py:obj:`~.cudaGraphicsResource_t`. These resources also include memory allocations by :py:obj:`~.cudaMalloc`, :py:obj:`~.cudaMallocHost`, :py:obj:`~.cudaMallocManaged` and :py:obj:`~.cudaMallocPitch`. Any subsequent API call to this device will reinitialize the device. Note that this function will reset the device immediately. It is the caller's responsibility to ensure that the device is not being accessed by any other host threads from the process when this function is called. Returns ------- cudaError_t :py:obj:`~.cudaSuccess` See Also -------- :py:obj:`~.cudaDeviceSynchronize` Notes ----- :py:obj:`~.cudaDeviceReset()` will not destroy memory allocations by :py:obj:`~.cudaMallocAsync()` and :py:obj:`~.cudaMallocFromPoolAsync()`. These memory allocations need to be destroyed explicitly. If a non-primary :py:obj:`~.CUcontext` is current to the thread, :py:obj:`~.cudaDeviceReset()` will destroy only the internal CUDA RT state for that :py:obj:`~.CUcontext`. """ with nogil: err = cyruntime.cudaDeviceReset() return (_dict_cudaError_t[err],) @cython.embedsignature(True) def cudaDeviceSynchronize(): """ Wait for compute device to finish. Blocks until the device has completed all preceding requested tasks. :py:obj:`~.cudaDeviceSynchronize()` returns an error if one of the preceding tasks has failed. If the :py:obj:`~.cudaDeviceScheduleBlockingSync` flag was set for this device, the host thread will block until the device has finished its work. Returns ------- cudaError_t :py:obj:`~.cudaSuccess` See Also -------- :py:obj:`~.cudaDeviceReset`, :py:obj:`~.cuCtxSynchronize` """ with nogil: err = cyruntime.cudaDeviceSynchronize() return (_dict_cudaError_t[err],) @cython.embedsignature(True) def cudaDeviceSetLimit(limit not None : cudaLimit, size_t value): """ Set resource limits. Setting `limit` to `value` is a request by the application to update the current limit maintained by the device. The driver is free to modify the requested value to meet h/w requirements (this could be clamping to minimum or maximum values, rounding up to nearest element size, etc). The application can use :py:obj:`~.cudaDeviceGetLimit()` to find out exactly what the limit has been set to. Setting each :py:obj:`~.cudaLimit` has its own specific restrictions, so each is discussed here. - :py:obj:`~.cudaLimitStackSize` controls the stack size in bytes of each GPU thread. - :py:obj:`~.cudaLimitPrintfFifoSize` controls the size in bytes of the shared FIFO used by the :py:obj:`~.printf()` device system call. Setting :py:obj:`~.cudaLimitPrintfFifoSize` must not be performed after launching any kernel that uses the :py:obj:`~.printf()` device system call - in such case :py:obj:`~.cudaErrorInvalidValue` will be returned. - :py:obj:`~.cudaLimitMallocHeapSize` controls the size in bytes of the heap used by the :py:obj:`~.malloc()` and :py:obj:`~.free()` device system calls. Setting :py:obj:`~.cudaLimitMallocHeapSize` must not be performed after launching any kernel that uses the :py:obj:`~.malloc()` or :py:obj:`~.free()` device system calls - in such case :py:obj:`~.cudaErrorInvalidValue` will be returned. - :py:obj:`~.cudaLimitDevRuntimeSyncDepth` controls the maximum nesting depth of a grid at which a thread can safely call :py:obj:`~.cudaDeviceSynchronize()`. Setting this limit must be performed before any launch of a kernel that uses the device runtime and calls :py:obj:`~.cudaDeviceSynchronize()` above the default sync depth, two levels of grids. Calls to :py:obj:`~.cudaDeviceSynchronize()` will fail with error code :py:obj:`~.cudaErrorSyncDepthExceeded` if the limitation is violated. This limit can be set smaller than the default or up the maximum launch depth of 24. When setting this limit, keep in mind that additional levels of sync depth require the runtime to reserve large amounts of device memory which can no longer be used for user allocations. If these reservations of device memory fail, :py:obj:`~.cudaDeviceSetLimit` will return :py:obj:`~.cudaErrorMemoryAllocation`, and the limit can be reset to a lower value. This limit is only applicable to devices of compute capability < 9.0. Attempting to set this limit on devices of other compute capability will results in error :py:obj:`~.cudaErrorUnsupportedLimit` being returned. - :py:obj:`~.cudaLimitDevRuntimePendingLaunchCount` controls the maximum number of outstanding device runtime launches that can be made from the current device. A grid is outstanding from the point of launch up until the grid is known to have been completed. Device runtime launches which violate this limitation fail and return :py:obj:`~.cudaErrorLaunchPendingCountExceeded` when :py:obj:`~.cudaGetLastError()` is called after launch. If more pending launches than the default (2048 launches) are needed for a module using the device runtime, this limit can be increased. Keep in mind that being able to sustain additional pending launches will require the runtime to reserve larger amounts of device memory upfront which can no longer be used for allocations. If these reservations fail, :py:obj:`~.cudaDeviceSetLimit` will return :py:obj:`~.cudaErrorMemoryAllocation`, and the limit can be reset to a lower value. This limit is only applicable to devices of compute capability 3.5 and higher. Attempting to set this limit on devices of compute capability less than 3.5 will result in the error :py:obj:`~.cudaErrorUnsupportedLimit` being returned. - :py:obj:`~.cudaLimitMaxL2FetchGranularity` controls the L2 cache fetch granularity. Values can range from 0B to 128B. This is purely a performance hint and it can be ignored or clamped depending on the platform. - :py:obj:`~.cudaLimitPersistingL2CacheSize` controls size in bytes available for persisting L2 cache. This is purely a performance hint and it can be ignored or clamped depending on the platform. Parameters ---------- limit : :py:obj:`~.cudaLimit` Limit to set value : size_t Size of limit Returns ------- cudaError_t :py:obj:`~.cudaSuccess`, :py:obj:`~.cudaErrorUnsupportedLimit`, :py:obj:`~.cudaErrorInvalidValue`, :py:obj:`~.cudaErrorMemoryAllocation` See Also -------- :py:obj:`~.cudaDeviceGetLimit`, :py:obj:`~.cuCtxSetLimit` """ cdef cyruntime.cudaLimit cylimit = limit.value with nogil: err = cyruntime.cudaDeviceSetLimit(cylimit, value) return (_dict_cudaError_t[err],) @cython.embedsignature(True) def cudaDeviceGetLimit(limit not None : cudaLimit): """ Return resource limits. Returns in `*pValue` the current size of `limit`. The following :py:obj:`~.cudaLimit` values are supported. - :py:obj:`~.cudaLimitStackSize` is the stack size in bytes of each GPU thread. - :py:obj:`~.cudaLimitPrintfFifoSize` is the size in bytes of the shared FIFO used by the :py:obj:`~.printf()` device system call. - :py:obj:`~.cudaLimitMallocHeapSize` is the size in bytes of the heap used by the :py:obj:`~.malloc()` and :py:obj:`~.free()` device system calls. - :py:obj:`~.cudaLimitDevRuntimeSyncDepth` is the maximum grid depth at which a thread can isssue the device runtime call :py:obj:`~.cudaDeviceSynchronize()` to wait on child grid launches to complete. This functionality is removed for devices of compute capability >= 9.0, and hence will return error :py:obj:`~.cudaErrorUnsupportedLimit` on such devices. - :py:obj:`~.cudaLimitDevRuntimePendingLaunchCount` is the maximum number of outstanding device runtime launches. - :py:obj:`~.cudaLimitMaxL2FetchGranularity` is the L2 cache fetch granularity. - :py:obj:`~.cudaLimitPersistingL2CacheSize` is the persisting L2 cache size in bytes. Parameters ---------- limit : :py:obj:`~.cudaLimit` Limit to query Returns ------- cudaError_t :py:obj:`~.cudaSuccess`, :py:obj:`~.cudaErrorUnsupportedLimit`, :py:obj:`~.cudaErrorInvalidValue` pValue : int Returned size of the limit See Also -------- :py:obj:`~.cudaDeviceSetLimit`, :py:obj:`~.cuCtxGetLimit` """ cdef size_t pValue = 0 cdef cyruntime.cudaLimit cylimit = limit.value with nogil: err = cyruntime.cudaDeviceGetLimit(&pValue, cylimit) if err != cyruntime.cudaSuccess: return (_dict_cudaError_t[err], None) return (_dict_cudaError_t[err], pValue) @cython.embedsignature(True) def cudaDeviceGetTexture1DLinearMaxWidth(fmtDesc : Optional[cudaChannelFormatDesc], int device): """ Returns the maximum number of elements allocatable in a 1D linear texture for a given element size. Returns in `maxWidthInElements` the maximum number of elements allocatable in a 1D linear texture for given format descriptor `fmtDesc`. Parameters ---------- fmtDesc : :py:obj:`~.cudaChannelFormatDesc` Texture format description. None : int None Returns ------- cudaError_t :py:obj:`~.cudaSuccess`, :py:obj:`~.cudaErrorUnsupportedLimit`, :py:obj:`~.cudaErrorInvalidValue` maxWidthInElements : int Returns maximum number of texture elements allocatable for given `fmtDesc`. See Also -------- :py:obj:`~.cuDeviceGetTexture1DLinearMaxWidth` """ cdef size_t maxWidthInElements = 0 cdef cyruntime.cudaChannelFormatDesc* cyfmtDesc_ptr = fmtDesc._pvt_ptr if fmtDesc is not None else NULL with nogil: err = cyruntime.cudaDeviceGetTexture1DLinearMaxWidth(&maxWidthInElements, cyfmtDesc_ptr, device) if err != cyruntime.cudaSuccess: return (_dict_cudaError_t[err], None) return (_dict_cudaError_t[err], maxWidthInElements) @cython.embedsignature(True) def cudaDeviceGetCacheConfig(): """ Returns the preferred cache configuration for the current device. On devices where the L1 cache and shared memory use the same hardware resources, this returns through `pCacheConfig` the preferred cache configuration for the current device. This is only a preference. The runtime will use the requested configuration if possible, but it is free to choose a different configuration if required to execute functions. This will return a `pCacheConfig` of :py:obj:`~.cudaFuncCachePreferNone` on devices where the size of the L1 cache and shared memory are fixed. The supported cache configurations are: - :py:obj:`~.cudaFuncCachePreferNone`: no preference for shared memory or L1 (default) - :py:obj:`~.cudaFuncCachePreferShared`: prefer larger shared memory and smaller L1 cache - :py:obj:`~.cudaFuncCachePreferL1`: prefer larger L1 cache and smaller shared memory - :py:obj:`~.cudaFuncCachePreferEqual`: prefer equal size L1 cache and shared memory Returns ------- cudaError_t :py:obj:`~.cudaSuccess` pCacheConfig : :py:obj:`~.cudaFuncCache` Returned cache configuration See Also -------- :py:obj:`~.cudaDeviceSetCacheConfig`, :py:obj:`~.cudaFuncSetCacheConfig (C API)`, cudaFuncSetCacheConfig (C++ API), :py:obj:`~.cuCtxGetCacheConfig` """ cdef cyruntime.cudaFuncCache pCacheConfig with nogil: err = cyruntime.cudaDeviceGetCacheConfig(&pCacheConfig) if err != cyruntime.cudaSuccess: return (_dict_cudaError_t[err], None) return (_dict_cudaError_t[err], cudaFuncCache(pCacheConfig)) @cython.embedsignature(True) def cudaDeviceGetStreamPriorityRange(): """ Returns numerical values that correspond to the least and greatest stream priorities. Returns in `*leastPriority` and `*greatestPriority` the numerical values that correspond to the least and greatest stream priorities respectively. Stream priorities follow a convention where lower numbers imply greater priorities. The range of meaningful stream priorities is given by [`*greatestPriority`, `*leastPriority`]. If the user attempts to create a stream with a priority value that is outside the the meaningful range as specified by this API, the priority is automatically clamped down or up to either `*leastPriority` or `*greatestPriority` respectively. See :py:obj:`~.cudaStreamCreateWithPriority` for details on creating a priority stream. A NULL may be passed in for `*leastPriority` or `*greatestPriority` if the value is not desired. This function will return '0' in both `*leastPriority` and `*greatestPriority` if the current context's device does not support stream priorities (see :py:obj:`~.cudaDeviceGetAttribute`). Returns ------- cudaError_t :py:obj:`~.cudaSuccess` leastPriority : int Pointer to an int in which the numerical value for least stream priority is returned greatestPriority : int Pointer to an int in which the numerical value for greatest stream priority is returned See Also -------- :py:obj:`~.cudaStreamCreateWithPriority`, :py:obj:`~.cudaStreamGetPriority`, :py:obj:`~.cuCtxGetStreamPriorityRange` """ cdef int leastPriority = 0 cdef int greatestPriority = 0 with nogil: err = cyruntime.cudaDeviceGetStreamPriorityRange(&leastPriority, &greatestPriority) if err != cyruntime.cudaSuccess: return (_dict_cudaError_t[err], None, None) return (_dict_cudaError_t[err], leastPriority, greatestPriority) @cython.embedsignature(True) def cudaDeviceSetCacheConfig(cacheConfig not None : cudaFuncCache): """ Sets the preferred cache configuration for the current device. On devices where the L1 cache and shared memory use the same hardware resources, this sets through `cacheConfig` the preferred cache configuration for the current device. This is only a preference. The runtime will use the requested configuration if possible, but it is free to choose a different configuration if required to execute the function. Any function preference set via :py:obj:`~.cudaFuncSetCacheConfig (C API)` or cudaFuncSetCacheConfig (C++ API) will be preferred over this device-wide setting. Setting the device-wide cache configuration to :py:obj:`~.cudaFuncCachePreferNone` will cause subsequent kernel launches to prefer to not change the cache configuration unless required to launch the kernel. This setting does nothing on devices where the size of the L1 cache and shared memory are fixed. Launching a kernel with a different preference than the most recent preference setting may insert a device-side synchronization point. The supported cache configurations are: - :py:obj:`~.cudaFuncCachePreferNone`: no preference for shared memory or L1 (default) - :py:obj:`~.cudaFuncCachePreferShared`: prefer larger shared memory and smaller L1 cache - :py:obj:`~.cudaFuncCachePreferL1`: prefer larger L1 cache and smaller shared memory - :py:obj:`~.cudaFuncCachePreferEqual`: prefer equal size L1 cache and shared memory Parameters ---------- cacheConfig : :py:obj:`~.cudaFuncCache` Requested cache configuration Returns ------- cudaError_t :py:obj:`~.cudaSuccess` See Also -------- :py:obj:`~.cudaDeviceGetCacheConfig`, :py:obj:`~.cudaFuncSetCacheConfig (C API)`, cudaFuncSetCacheConfig (C++ API), :py:obj:`~.cuCtxSetCacheConfig` """ cdef cyruntime.cudaFuncCache cycacheConfig = cacheConfig.value with nogil: err = cyruntime.cudaDeviceSetCacheConfig(cycacheConfig) return (_dict_cudaError_t[err],) @cython.embedsignature(True) def cudaDeviceGetByPCIBusId(char* pciBusId): """ Returns a handle to a compute device. Returns in `*device` a device ordinal given a PCI bus ID string. where `domain`, `bus`, `device`, and `function` are all hexadecimal values Parameters ---------- pciBusId : bytes String in one of the following forms: Returns ------- cudaError_t :py:obj:`~.cudaSuccess`, :py:obj:`~.cudaErrorInvalidValue`, :py:obj:`~.cudaErrorInvalidDevice` device : int Returned device ordinal See Also -------- :py:obj:`~.cudaDeviceGetPCIBusId`, :py:obj:`~.cuDeviceGetByPCIBusId` """ cdef int device = 0 with nogil: err = cyruntime.cudaDeviceGetByPCIBusId(&device, pciBusId) if err != cyruntime.cudaSuccess: return (_dict_cudaError_t[err], None) return (_dict_cudaError_t[err], device) @cython.embedsignature(True) def cudaDeviceGetPCIBusId(int length, int device): """ Returns a PCI Bus Id string for the device. Returns an ASCII string identifying the device `dev` in the NULL- terminated string pointed to by `pciBusId`. `length` specifies the maximum length of the string that may be returned. where `domain`, `bus`, `device`, and `function` are all hexadecimal values. pciBusId should be large enough to store 13 characters including the NULL-terminator. Parameters ---------- length : int Maximum length of string to store in `name` device : int Device to get identifier string for Returns ------- cudaError_t :py:obj:`~.cudaSuccess`, :py:obj:`~.cudaErrorInvalidValue`, :py:obj:`~.cudaErrorInvalidDevice` pciBusId : bytes Returned identifier string for the device in the following format See Also -------- :py:obj:`~.cudaDeviceGetByPCIBusId`, :py:obj:`~.cuDeviceGetPCIBusId` """ pypciBusId = b" " * length cdef char* pciBusId = pypciBusId with nogil: err = cyruntime.cudaDeviceGetPCIBusId(pciBusId, length, device) if err != cyruntime.cudaSuccess: return (_dict_cudaError_t[err], None) return (_dict_cudaError_t[err], pypciBusId) @cython.embedsignature(True) def cudaIpcGetEventHandle(event): """ Gets an interprocess handle for a previously allocated event. Takes as input a previously allocated event. This event must have been created with the :py:obj:`~.cudaEventInterprocess` and :py:obj:`~.cudaEventDisableTiming` flags set. This opaque handle may be copied into other processes and opened with :py:obj:`~.cudaIpcOpenEventHandle` to allow efficient hardware synchronization between GPU work in different processes. After the event has been been opened in the importing process, :py:obj:`~.cudaEventRecord`, :py:obj:`~.cudaEventSynchronize`, :py:obj:`~.cudaStreamWaitEvent` and :py:obj:`~.cudaEventQuery` may be used in either process. Performing operations on the imported event after the exported event has been freed with :py:obj:`~.cudaEventDestroy` will result in undefined behavior. IPC functionality is restricted to devices with support for unified addressing on Linux and Windows operating systems. IPC functionality on Windows is supported for compatibility purposes but not recommended as it comes with performance cost. Users can test their device for IPC functionality by calling :py:obj:`~.cudaDeviceGetAttribute` with :py:obj:`~.cudaDevAttrIpcEventSupport` Parameters ---------- event : :py:obj:`~.CUevent` or :py:obj:`~.cudaEvent_t` Event allocated with :py:obj:`~.cudaEventInterprocess` and :py:obj:`~.cudaEventDisableTiming` flags. Returns ------- cudaError_t :py:obj:`~.cudaSuccess`, :py:obj:`~.cudaErrorInvalidResourceHandle`, :py:obj:`~.cudaErrorMemoryAllocation`, :py:obj:`~.cudaErrorMapBufferObjectFailed`, :py:obj:`~.cudaErrorNotSupported`, :py:obj:`~.cudaErrorInvalidValue` handle : :py:obj:`~.cudaIpcEventHandle_t` Pointer to a user allocated cudaIpcEventHandle in which to return the opaque event handle See Also -------- :py:obj:`~.cudaEventCreate`, :py:obj:`~.cudaEventDestroy`, :py:obj:`~.cudaEventSynchronize`, :py:obj:`~.cudaEventQuery`, :py:obj:`~.cudaStreamWaitEvent`, :py:obj:`~.cudaIpcOpenEventHandle`, :py:obj:`~.cudaIpcGetMemHandle`, :py:obj:`~.cudaIpcOpenMemHandle`, :py:obj:`~.cudaIpcCloseMemHandle`, :py:obj:`~.cuIpcGetEventHandle` """ cdef cyruntime.cudaEvent_t cyevent if event is None: pevent = 0 elif isinstance(event, (cudaEvent_t,driver.CUevent)): pevent = int(event) else: pevent = int(cudaEvent_t(event)) cyevent = pevent cdef cudaIpcEventHandle_t handle = cudaIpcEventHandle_t() with nogil: err = cyruntime.cudaIpcGetEventHandle(handle._pvt_ptr, cyevent) if err != cyruntime.cudaSuccess: return (_dict_cudaError_t[err], None) return (_dict_cudaError_t[err], handle) @cython.embedsignature(True) def cudaIpcOpenEventHandle(handle not None : cudaIpcEventHandle_t): """ Opens an interprocess event handle for use in the current process. Opens an interprocess event handle exported from another process with :py:obj:`~.cudaIpcGetEventHandle`. This function returns a :py:obj:`~.cudaEvent_t` that behaves like a locally created event with the :py:obj:`~.cudaEventDisableTiming` flag specified. This event must be freed with :py:obj:`~.cudaEventDestroy`. Performing operations on the imported event after the exported event has been freed with :py:obj:`~.cudaEventDestroy` will result in undefined behavior. IPC functionality is restricted to devices with support for unified addressing on Linux and Windows operating systems. IPC functionality on Windows is supported for compatibility purposes but not recommended as it comes with performance cost. Users can test their device for IPC functionality by calling :py:obj:`~.cudaDeviceGetAttribute` with :py:obj:`~.cudaDevAttrIpcEventSupport` Parameters ---------- handle : :py:obj:`~.cudaIpcEventHandle_t` Interprocess handle to open Returns ------- cudaError_t :py:obj:`~.cudaSuccess`, :py:obj:`~.cudaErrorMapBufferObjectFailed`, :py:obj:`~.cudaErrorNotSupported`, :py:obj:`~.cudaErrorInvalidValue`, :py:obj:`~.cudaErrorDeviceUninitialized` event : :py:obj:`~.cudaEvent_t` Returns the imported event See Also -------- :py:obj:`~.cudaEventCreate`, :py:obj:`~.cudaEventDestroy`, :py:obj:`~.cudaEventSynchronize`, :py:obj:`~.cudaEventQuery`, :py:obj:`~.cudaStreamWaitEvent`, :py:obj:`~.cudaIpcGetEventHandle`, :py:obj:`~.cudaIpcGetMemHandle`, :py:obj:`~.cudaIpcOpenMemHandle`, :py:obj:`~.cudaIpcCloseMemHandle`, :py:obj:`~.cuIpcOpenEventHandle` """ cdef cudaEvent_t event = cudaEvent_t() with nogil: err = cyruntime.cudaIpcOpenEventHandle(event._pvt_ptr, handle._pvt_ptr[0]) if err != cyruntime.cudaSuccess: return (_dict_cudaError_t[err], None) return (_dict_cudaError_t[err], event) @cython.embedsignature(True) def cudaIpcGetMemHandle(devPtr): """ Gets an interprocess memory handle for an existing device memory allocation. Takes a pointer to the base of an existing device memory allocation created with :py:obj:`~.cudaMalloc` and exports it for use in another process. This is a lightweight operation and may be called multiple times on an allocation without adverse effects. If a region of memory is freed with :py:obj:`~.cudaFree` and a subsequent call to :py:obj:`~.cudaMalloc` returns memory with the same device address, :py:obj:`~.cudaIpcGetMemHandle` will return a unique handle for the new memory. IPC functionality is restricted to devices with support for unified addressing on Linux and Windows operating systems. IPC functionality on Windows is supported for compatibility purposes but not recommended as it comes with performance cost. Users can test their device for IPC functionality by calling :py:obj:`~.cudaDeviceGetAttribute` with :py:obj:`~.cudaDevAttrIpcEventSupport` Parameters ---------- devPtr : Any Base pointer to previously allocated device memory Returns ------- cudaError_t :py:obj:`~.cudaSuccess`, :py:obj:`~.cudaErrorMemoryAllocation`, :py:obj:`~.cudaErrorMapBufferObjectFailed`, :py:obj:`~.cudaErrorNotSupported`, :py:obj:`~.cudaErrorInvalidValue` handle : :py:obj:`~.cudaIpcMemHandle_t` Pointer to user allocated :py:obj:`~.cudaIpcMemHandle` to return the handle in. See Also -------- :py:obj:`~.cudaMalloc`, :py:obj:`~.cudaFree`, :py:obj:`~.cudaIpcGetEventHandle`, :py:obj:`~.cudaIpcOpenEventHandle`, :py:obj:`~.cudaIpcOpenMemHandle`, :py:obj:`~.cudaIpcCloseMemHandle`, :py:obj:`~.cuIpcGetMemHandle` """ cdef cudaIpcMemHandle_t handle = cudaIpcMemHandle_t() cydevPtr = _HelperInputVoidPtr(devPtr) cdef void* cydevPtr_ptr = cydevPtr.cptr with nogil: err = cyruntime.cudaIpcGetMemHandle(handle._pvt_ptr, cydevPtr_ptr) if err != cyruntime.cudaSuccess: return (_dict_cudaError_t[err], None) return (_dict_cudaError_t[err], handle) @cython.embedsignature(True) def cudaIpcOpenMemHandle(handle not None : cudaIpcMemHandle_t, unsigned int flags): """ Opens an interprocess memory handle exported from another process and returns a device pointer usable in the local process. Maps memory exported from another process with :py:obj:`~.cudaIpcGetMemHandle` into the current device address space. For contexts on different devices :py:obj:`~.cudaIpcOpenMemHandle` can attempt to enable peer access between the devices as if the user called :py:obj:`~.cudaDeviceEnablePeerAccess`. This behavior is controlled by the :py:obj:`~.cudaIpcMemLazyEnablePeerAccess` flag. :py:obj:`~.cudaDeviceCanAccessPeer` can determine if a mapping is possible. :py:obj:`~.cudaIpcOpenMemHandle` can open handles to devices that may not be visible in the process calling the API. Contexts that may open :py:obj:`~.cudaIpcMemHandles` are restricted in the following way. :py:obj:`~.cudaIpcMemHandles` from each device in a given process may only be opened by one context per device per other process. If the memory handle has already been opened by the current context, the reference count on the handle is incremented by 1 and the existing device pointer is returned. Memory returned from :py:obj:`~.cudaIpcOpenMemHandle` must be freed with :py:obj:`~.cudaIpcCloseMemHandle`. Calling :py:obj:`~.cudaFree` on an exported memory region before calling :py:obj:`~.cudaIpcCloseMemHandle` in the importing context will result in undefined behavior. IPC functionality is restricted to devices with support for unified addressing on Linux and Windows operating systems. IPC functionality on Windows is supported for compatibility purposes but not recommended as it comes with performance cost. Users can test their device for IPC functionality by calling :py:obj:`~.cudaDeviceGetAttribute` with :py:obj:`~.cudaDevAttrIpcEventSupport` Parameters ---------- handle : :py:obj:`~.cudaIpcMemHandle_t` :py:obj:`~.cudaIpcMemHandle` to open flags : unsigned int Flags for this operation. Must be specified as :py:obj:`~.cudaIpcMemLazyEnablePeerAccess` Returns ------- cudaError_t :py:obj:`~.cudaSuccess`, :py:obj:`~.cudaErrorMapBufferObjectFailed`, :py:obj:`~.cudaErrorInvalidResourceHandle`, :py:obj:`~.cudaErrorDeviceUninitialized`, :py:obj:`~.cudaErrorTooManyPeers`, :py:obj:`~.cudaErrorNotSupported`, :py:obj:`~.cudaErrorInvalidValue` devPtr : Any Returned device pointer See Also -------- :py:obj:`~.cudaMalloc`, :py:obj:`~.cudaFree`, :py:obj:`~.cudaIpcGetEventHandle`, :py:obj:`~.cudaIpcOpenEventHandle`, :py:obj:`~.cudaIpcGetMemHandle`, :py:obj:`~.cudaIpcCloseMemHandle`, :py:obj:`~.cudaDeviceEnablePeerAccess`, :py:obj:`~.cudaDeviceCanAccessPeer`, :py:obj:`~.cuIpcOpenMemHandle` Notes ----- No guarantees are made about the address returned in `*devPtr`. In particular, multiple processes may not receive the same address for the same `handle`. """ cdef void_ptr devPtr = 0 with nogil: err = cyruntime.cudaIpcOpenMemHandle(&devPtr, handle._pvt_ptr[0], flags) if err != cyruntime.cudaSuccess: return (_dict_cudaError_t[err], None) return (_dict_cudaError_t[err], devPtr) @cython.embedsignature(True) def cudaIpcCloseMemHandle(devPtr): """ Attempts to close memory mapped with cudaIpcOpenMemHandle. Decrements the reference count of the memory returnd by :py:obj:`~.cudaIpcOpenMemHandle` by 1. When the reference count reaches 0, this API unmaps the memory. The original allocation in the exporting process as well as imported mappings in other processes will be unaffected. Any resources used to enable peer access will be freed if this is the last mapping using them. IPC functionality is restricted to devices with support for unified addressing on Linux and Windows operating systems. IPC functionality on Windows is supported for compatibility purposes but not recommended as it comes with performance cost. Users can test their device for IPC functionality by calling :py:obj:`~.cudaDeviceGetAttribute` with :py:obj:`~.cudaDevAttrIpcEventSupport` Parameters ---------- devPtr : Any Device pointer returned by :py:obj:`~.cudaIpcOpenMemHandle` Returns ------- cudaError_t :py:obj:`~.cudaSuccess`, :py:obj:`~.cudaErrorMapBufferObjectFailed`, :py:obj:`~.cudaErrorNotSupported`, :py:obj:`~.cudaErrorInvalidValue` See Also -------- :py:obj:`~.cudaMalloc`, :py:obj:`~.cudaFree`, :py:obj:`~.cudaIpcGetEventHandle`, :py:obj:`~.cudaIpcOpenEventHandle`, :py:obj:`~.cudaIpcGetMemHandle`, :py:obj:`~.cudaIpcOpenMemHandle`, :py:obj:`~.cuIpcCloseMemHandle` """ cydevPtr = _HelperInputVoidPtr(devPtr) cdef void* cydevPtr_ptr = cydevPtr.cptr with nogil: err = cyruntime.cudaIpcCloseMemHandle(cydevPtr_ptr) return (_dict_cudaError_t[err],) @cython.embedsignature(True) def cudaDeviceFlushGPUDirectRDMAWrites(target not None : cudaFlushGPUDirectRDMAWritesTarget, scope not None : cudaFlushGPUDirectRDMAWritesScope): """ Blocks until remote writes are visible to the specified scope. Blocks until remote writes to the target context via mappings created through GPUDirect RDMA APIs, like nvidia_p2p_get_pages (see https://docs.nvidia.com/cuda/gpudirect-rdma for more information), are visible to the specified scope. If the scope equals or lies within the scope indicated by :py:obj:`~.cudaDevAttrGPUDirectRDMAWritesOrdering`, the call will be a no-op and can be safely omitted for performance. This can be determined by comparing the numerical values between the two enums, with smaller scopes having smaller values. Users may query support for this API via :py:obj:`~.cudaDevAttrGPUDirectRDMAFlushWritesOptions`. Parameters ---------- target : :py:obj:`~.cudaFlushGPUDirectRDMAWritesTarget` The target of the operation, see cudaFlushGPUDirectRDMAWritesTarget scope : :py:obj:`~.cudaFlushGPUDirectRDMAWritesScope` The scope of the operation, see cudaFlushGPUDirectRDMAWritesScope Returns ------- cudaError_t :py:obj:`~.cudaSuccess`, :py:obj:`~.cudaErrorNotSupported`, See Also -------- :py:obj:`~.cuFlushGPUDirectRDMAWrites` """ cdef cyruntime.cudaFlushGPUDirectRDMAWritesTarget cytarget = target.value cdef cyruntime.cudaFlushGPUDirectRDMAWritesScope cyscope = scope.value with nogil: err = cyruntime.cudaDeviceFlushGPUDirectRDMAWrites(cytarget, cyscope) return (_dict_cudaError_t[err],) ctypedef struct cudaAsyncCallbackData_st: cyruntime.cudaAsyncCallback callback void *userData ctypedef cudaAsyncCallbackData_st cudaAsyncCallbackData @cython.show_performance_hints(False) cdef void cudaAsyncNotificationCallbackWrapper(cyruntime.cudaAsyncNotificationInfo_t *info, void *data, cyruntime.cudaAsyncCallbackHandle_t handle) nogil: cdef cudaAsyncCallbackData *cbData = data with gil: cbData.callback(info, cbData.userData, handle) @cython.embedsignature(True) def cudaDeviceRegisterAsyncNotification(int device, callbackFunc, userData): """ Registers a callback function to receive async notifications. Registers `callbackFunc` to receive async notifications. The `userData` parameter is passed to the callback function at async notification time. Likewise, `callback` is also passed to the callback function to distinguish between multiple registered callbacks. The callback function being registered should be designed to return quickly (~10ms). Any long running tasks should be queued for execution on an application thread. Callbacks may not call cudaDeviceRegisterAsyncNotification or cudaDeviceUnregisterAsyncNotification. Doing so will result in :py:obj:`~.cudaErrorNotPermitted`. Async notification callbacks execute in an undefined order and may be serialized. Returns in `*callback` a handle representing the registered callback instance. Parameters ---------- device : int The device on which to register the callback callbackFunc : :py:obj:`~.cudaAsyncCallback` The function to register as a callback userData : Any A generic pointer to user data. This is passed into the callback function. Returns ------- cudaError_t :py:obj:`~.cudaSuccess` :py:obj:`~.cudaErrorNotSupported` :py:obj:`~.cudaErrorInvalidDevice` :py:obj:`~.cudaErrorInvalidValue` :py:obj:`~.cudaErrorNotPermitted` :py:obj:`~.cudaErrorUnknown` callback : :py:obj:`~.cudaAsyncCallbackHandle_t` A handle representing the registered callback instance See Also -------- :py:obj:`~.cudaDeviceUnregisterAsyncNotification` """ cdef cyruntime.cudaAsyncCallback cycallbackFunc if callbackFunc is None: pcallbackFunc = 0 elif isinstance(callbackFunc, (cudaAsyncCallback,)): pcallbackFunc = int(callbackFunc) else: pcallbackFunc = int(cudaAsyncCallback(callbackFunc)) cycallbackFunc = pcallbackFunc cyuserData = _HelperInputVoidPtr(userData) cdef void* cyuserData_ptr = cyuserData.cptr cdef cudaAsyncCallbackData *cbData = NULL cbData = malloc(sizeof(cbData[0])) if cbData == NULL: return (cudaError_t.cudaErrorMemoryAllocation, None) cbData.callback = cycallbackFunc cbData.userData = cyuserData_ptr cdef cudaAsyncCallbackHandle_t callback = cudaAsyncCallbackHandle_t() with nogil: err = cyruntime.cudaDeviceRegisterAsyncNotification(device, cudaAsyncNotificationCallbackWrapper, cbData, callback._pvt_ptr) if err != cyruntime.cudaSuccess: free(cbData) else: m_global._allocated[int(callback)] = cbData if err != cyruntime.cudaSuccess: return (_dict_cudaError_t[err], None) return (_dict_cudaError_t[err], callback) @cython.embedsignature(True) def cudaDeviceUnregisterAsyncNotification(int device, callback): """ Unregisters an async notification callback. Unregisters `callback` so that the corresponding callback function will stop receiving async notifications. Parameters ---------- device : int The device from which to remove `callback`. callback : :py:obj:`~.cudaAsyncCallbackHandle_t` The callback instance to unregister from receiving async notifications. Returns ------- cudaError_t :py:obj:`~.cudaSuccess` :py:obj:`~.cudaErrorNotSupported` :py:obj:`~.cudaErrorInvalidDevice` :py:obj:`~.cudaErrorInvalidValue` :py:obj:`~.cudaErrorNotPermitted` :py:obj:`~.cudaErrorUnknown` See Also -------- :py:obj:`~.cudaDeviceRegisterAsyncNotification` """ cdef cyruntime.cudaAsyncCallbackHandle_t cycallback if callback is None: pcallback = 0 elif isinstance(callback, (cudaAsyncCallbackHandle_t,)): pcallback = int(callback) else: pcallback = int(cudaAsyncCallbackHandle_t(callback)) cycallback = pcallback with nogil: err = cyruntime.cudaDeviceUnregisterAsyncNotification(device, cycallback) if err == cyruntime.cudaSuccess: free(m_global._allocated[pcallback]) m_global._allocated.erase(pcallback) return (_dict_cudaError_t[err],) @cython.embedsignature(True) def cudaDeviceGetSharedMemConfig(): """ Returns the shared memory configuration for the current device. [Deprecated] This function will return in `pConfig` the current size of shared memory banks on the current device. On devices with configurable shared memory banks, :py:obj:`~.cudaDeviceSetSharedMemConfig` can be used to change this setting, so that all subsequent kernel launches will by default use the new bank size. When :py:obj:`~.cudaDeviceGetSharedMemConfig` is called on devices without configurable shared memory, it will return the fixed bank size of the hardware. The returned bank configurations can be either: - :py:obj:`~.cudaSharedMemBankSizeFourByte` - shared memory bank width is four bytes. - :py:obj:`~.cudaSharedMemBankSizeEightByte` - shared memory bank width is eight bytes. Returns ------- cudaError_t :py:obj:`~.cudaSuccess`, :py:obj:`~.cudaErrorInvalidValue` pConfig : :py:obj:`~.cudaSharedMemConfig` Returned cache configuration See Also -------- :py:obj:`~.cudaDeviceSetCacheConfig`, :py:obj:`~.cudaDeviceGetCacheConfig`, :py:obj:`~.cudaDeviceSetSharedMemConfig`, :py:obj:`~.cudaFuncSetCacheConfig`, :py:obj:`~.cuCtxGetSharedMemConfig` """ cdef cyruntime.cudaSharedMemConfig pConfig with nogil: err = cyruntime.cudaDeviceGetSharedMemConfig(&pConfig) if err != cyruntime.cudaSuccess: return (_dict_cudaError_t[err], None) return (_dict_cudaError_t[err], cudaSharedMemConfig(pConfig)) @cython.embedsignature(True) def cudaDeviceSetSharedMemConfig(config not None : cudaSharedMemConfig): """ Sets the shared memory configuration for the current device. [Deprecated] On devices with configurable shared memory banks, this function will set the shared memory bank size which is used for all subsequent kernel launches. Any per-function setting of shared memory set via :py:obj:`~.cudaFuncSetSharedMemConfig` will override the device wide setting. Changing the shared memory configuration between launches may introduce a device side synchronization point. Changing the shared memory bank size will not increase shared memory usage or affect occupancy of kernels, but may have major effects on performance. Larger bank sizes will allow for greater potential bandwidth to shared memory, but will change what kinds of accesses to shared memory will result in bank conflicts. This function will do nothing on devices with fixed shared memory bank size. The supported bank configurations are: - :py:obj:`~.cudaSharedMemBankSizeDefault`: set bank width the device default (currently, four bytes) - :py:obj:`~.cudaSharedMemBankSizeFourByte`: set shared memory bank width to be four bytes natively. - :py:obj:`~.cudaSharedMemBankSizeEightByte`: set shared memory bank width to be eight bytes natively. Parameters ---------- config : :py:obj:`~.cudaSharedMemConfig` Requested cache configuration Returns ------- cudaError_t :py:obj:`~.cudaSuccess`, :py:obj:`~.cudaErrorInvalidValue` See Also -------- :py:obj:`~.cudaDeviceSetCacheConfig`, :py:obj:`~.cudaDeviceGetCacheConfig`, :py:obj:`~.cudaDeviceGetSharedMemConfig`, :py:obj:`~.cudaFuncSetCacheConfig`, :py:obj:`~.cuCtxSetSharedMemConfig` """ cdef cyruntime.cudaSharedMemConfig cyconfig = config.value with nogil: err = cyruntime.cudaDeviceSetSharedMemConfig(cyconfig) return (_dict_cudaError_t[err],) @cython.embedsignature(True) def cudaGetLastError(): """ Returns the last error from a runtime call. Returns the last error that has been produced by any of the runtime calls in the same instance of the CUDA Runtime library in the host thread and resets it to :py:obj:`~.cudaSuccess`. Note: Multiple instances of the CUDA Runtime library can be present in an application when using a library that statically links the CUDA Runtime. Returns ------- cudaError_t :py:obj:`~.cudaSuccess`, :py:obj:`~.cudaErrorMissingConfiguration`, :py:obj:`~.cudaErrorMemoryAllocation`, :py:obj:`~.cudaErrorInitializationError`, :py:obj:`~.cudaErrorLaunchFailure`, :py:obj:`~.cudaErrorLaunchTimeout`, :py:obj:`~.cudaErrorLaunchOutOfResources`, :py:obj:`~.cudaErrorInvalidDeviceFunction`, :py:obj:`~.cudaErrorInvalidConfiguration`, :py:obj:`~.cudaErrorInvalidDevice`, :py:obj:`~.cudaErrorInvalidValue`, :py:obj:`~.cudaErrorInvalidPitchValue`, :py:obj:`~.cudaErrorInvalidSymbol`, :py:obj:`~.cudaErrorUnmapBufferObjectFailed`, :py:obj:`~.cudaErrorInvalidDevicePointer`, :py:obj:`~.cudaErrorInvalidTexture`, :py:obj:`~.cudaErrorInvalidTextureBinding`, :py:obj:`~.cudaErrorInvalidChannelDescriptor`, :py:obj:`~.cudaErrorInvalidMemcpyDirection`, :py:obj:`~.cudaErrorInvalidFilterSetting`, :py:obj:`~.cudaErrorInvalidNormSetting`, :py:obj:`~.cudaErrorUnknown`, :py:obj:`~.cudaErrorInvalidResourceHandle`, :py:obj:`~.cudaErrorInsufficientDriver`, :py:obj:`~.cudaErrorNoDevice`, :py:obj:`~.cudaErrorSetOnActiveProcess`, :py:obj:`~.cudaErrorStartupFailure`, :py:obj:`~.cudaErrorInvalidPtx`, :py:obj:`~.cudaErrorUnsupportedPtxVersion`, :py:obj:`~.cudaErrorNoKernelImageForDevice`, :py:obj:`~.cudaErrorJitCompilerNotFound`, :py:obj:`~.cudaErrorJitCompilationDisabled` See Also -------- :py:obj:`~.cudaPeekAtLastError`, :py:obj:`~.cudaGetErrorName`, :py:obj:`~.cudaGetErrorString`, :py:obj:`~.cudaError` """ with nogil: err = cyruntime.cudaGetLastError() return (_dict_cudaError_t[err],) @cython.embedsignature(True) def cudaPeekAtLastError(): """ Returns the last error from a runtime call. Returns the last error that has been produced by any of the runtime calls in the same instance of the CUDA Runtime library in the host thread. This call does not reset the error to :py:obj:`~.cudaSuccess` like :py:obj:`~.cudaGetLastError()`. Note: Multiple instances of the CUDA Runtime library can be present in an application when using a library that statically links the CUDA Runtime. Returns ------- cudaError_t :py:obj:`~.cudaSuccess`, :py:obj:`~.cudaErrorMissingConfiguration`, :py:obj:`~.cudaErrorMemoryAllocation`, :py:obj:`~.cudaErrorInitializationError`, :py:obj:`~.cudaErrorLaunchFailure`, :py:obj:`~.cudaErrorLaunchTimeout`, :py:obj:`~.cudaErrorLaunchOutOfResources`, :py:obj:`~.cudaErrorInvalidDeviceFunction`, :py:obj:`~.cudaErrorInvalidConfiguration`, :py:obj:`~.cudaErrorInvalidDevice`, :py:obj:`~.cudaErrorInvalidValue`, :py:obj:`~.cudaErrorInvalidPitchValue`, :py:obj:`~.cudaErrorInvalidSymbol`, :py:obj:`~.cudaErrorUnmapBufferObjectFailed`, :py:obj:`~.cudaErrorInvalidDevicePointer`, :py:obj:`~.cudaErrorInvalidTexture`, :py:obj:`~.cudaErrorInvalidTextureBinding`, :py:obj:`~.cudaErrorInvalidChannelDescriptor`, :py:obj:`~.cudaErrorInvalidMemcpyDirection`, :py:obj:`~.cudaErrorInvalidFilterSetting`, :py:obj:`~.cudaErrorInvalidNormSetting`, :py:obj:`~.cudaErrorUnknown`, :py:obj:`~.cudaErrorInvalidResourceHandle`, :py:obj:`~.cudaErrorInsufficientDriver`, :py:obj:`~.cudaErrorNoDevice`, :py:obj:`~.cudaErrorSetOnActiveProcess`, :py:obj:`~.cudaErrorStartupFailure`, :py:obj:`~.cudaErrorInvalidPtx`, :py:obj:`~.cudaErrorUnsupportedPtxVersion`, :py:obj:`~.cudaErrorNoKernelImageForDevice`, :py:obj:`~.cudaErrorJitCompilerNotFound`, :py:obj:`~.cudaErrorJitCompilationDisabled` See Also -------- :py:obj:`~.cudaGetLastError`, :py:obj:`~.cudaGetErrorName`, :py:obj:`~.cudaGetErrorString`, :py:obj:`~.cudaError` """ with nogil: err = cyruntime.cudaPeekAtLastError() return (_dict_cudaError_t[err],) @cython.embedsignature(True) def cudaGetErrorName(error not None : cudaError_t): """ Returns the string representation of an error code enum name. Returns a string containing the name of an error code in the enum. If the error code is not recognized, "unrecognized error code" is returned. Parameters ---------- error : :py:obj:`~.cudaError_t` Error code to convert to string Returns ------- cudaError_t.cudaSuccess cudaError_t.cudaSuccess bytes `char*` pointer to a NULL-terminated string See Also -------- :py:obj:`~.cudaGetErrorString`, :py:obj:`~.cudaGetLastError`, :py:obj:`~.cudaPeekAtLastError`, :py:obj:`~.cudaError`, :py:obj:`~.cuGetErrorName` """ cdef cyruntime.cudaError_t cyerror = error.value with nogil: err = cyruntime.cudaGetErrorName(cyerror) return (cudaError_t.cudaSuccess, err) @cython.embedsignature(True) def cudaGetErrorString(error not None : cudaError_t): """ Returns the description string for an error code. Returns the description string for an error code. If the error code is not recognized, "unrecognized error code" is returned. Parameters ---------- error : :py:obj:`~.cudaError_t` Error code to convert to string Returns ------- cudaError_t.cudaSuccess cudaError_t.cudaSuccess bytes `char*` pointer to a NULL-terminated string See Also -------- :py:obj:`~.cudaGetErrorName`, :py:obj:`~.cudaGetLastError`, :py:obj:`~.cudaPeekAtLastError`, :py:obj:`~.cudaError`, :py:obj:`~.cuGetErrorString` """ cdef cyruntime.cudaError_t cyerror = error.value with nogil: err = cyruntime.cudaGetErrorString(cyerror) return (cudaError_t.cudaSuccess, err) @cython.embedsignature(True) def cudaGetDeviceCount(): """ Returns the number of compute-capable devices. Returns in `*count` the number of devices with compute capability greater or equal to 2.0 that are available for execution. Returns ------- cudaError_t :py:obj:`~.cudaSuccess` count : int Returns the number of devices with compute capability greater or equal to 2.0 See Also -------- :py:obj:`~.cudaGetDevice`, :py:obj:`~.cudaSetDevice`, :py:obj:`~.cudaGetDeviceProperties`, :py:obj:`~.cudaChooseDevice`, :py:obj:`~.cudaInitDevice`, :py:obj:`~.cuDeviceGetCount` """ cdef int count = 0 with nogil: err = cyruntime.cudaGetDeviceCount(&count) if err != cyruntime.cudaSuccess: return (_dict_cudaError_t[err], None) return (_dict_cudaError_t[err], count) @cython.embedsignature(True) def cudaGetDeviceProperties(int device): """ Returns information about the compute-device. Returns in `*prop` the properties of device `dev`. The :py:obj:`~.cudaDeviceProp` structure is defined as: **View CUDA Toolkit Documentation for a C++ code example** where: - :py:obj:`~.name[256]` is an ASCII string identifying the device. - :py:obj:`~.uuid` is a 16-byte unique identifier. - :py:obj:`~.totalGlobalMem` is the total amount of global memory available on the device in bytes. - :py:obj:`~.sharedMemPerBlock` is the maximum amount of shared memory available to a thread block in bytes. - :py:obj:`~.regsPerBlock` is the maximum number of 32-bit registers available to a thread block. - :py:obj:`~.warpSize` is the warp size in threads. - :py:obj:`~.memPitch` is the maximum pitch in bytes allowed by the memory copy functions that involve memory regions allocated through :py:obj:`~.cudaMallocPitch()`. - :py:obj:`~.maxThreadsPerBlock` is the maximum number of threads per block. - :py:obj:`~.maxThreadsDim[3]` contains the maximum size of each dimension of a block. - :py:obj:`~.maxGridSize[3]` contains the maximum size of each dimension of a grid. - :py:obj:`~.clockRate` is the clock frequency in kilohertz. - :py:obj:`~.totalConstMem` is the total amount of constant memory available on the device in bytes. - :py:obj:`~.major`, :py:obj:`~.minor` are the major and minor revision numbers defining the device's compute capability. - :py:obj:`~.textureAlignment` is the alignment requirement; texture base addresses that are aligned to :py:obj:`~.textureAlignment` bytes do not need an offset applied to texture fetches. - :py:obj:`~.texturePitchAlignment` is the pitch alignment requirement for 2D texture references that are bound to pitched memory. - :py:obj:`~.deviceOverlap` is 1 if the device can concurrently copy memory between host and device while executing a kernel, or 0 if not. Deprecated, use instead asyncEngineCount. - :py:obj:`~.multiProcessorCount` is the number of multiprocessors on the device. - :py:obj:`~.kernelExecTimeoutEnabled` is 1 if there is a run time limit for kernels executed on the device, or 0 if not. - :py:obj:`~.integrated` is 1 if the device is an integrated (motherboard) GPU and 0 if it is a discrete (card) component. - :py:obj:`~.canMapHostMemory` is 1 if the device can map host memory into the CUDA address space for use with :py:obj:`~.cudaHostAlloc()`/:py:obj:`~.cudaHostGetDevicePointer()`, or 0 if not. - :py:obj:`~.computeMode` is the compute mode that the device is currently in. Available modes are as follows: - cudaComputeModeDefault: Default mode - Device is not restricted and multiple threads can use :py:obj:`~.cudaSetDevice()` with this device. - cudaComputeModeProhibited: Compute-prohibited mode - No threads can use :py:obj:`~.cudaSetDevice()` with this device. - cudaComputeModeExclusiveProcess: Compute-exclusive-process mode - Many threads in one process will be able to use :py:obj:`~.cudaSetDevice()` with this device. When an occupied exclusive mode device is chosen with :py:obj:`~.cudaSetDevice`, all subsequent non-device management runtime functions will return :py:obj:`~.cudaErrorDevicesUnavailable`. - :py:obj:`~.maxTexture1D` is the maximum 1D texture size. - :py:obj:`~.maxTexture1DMipmap` is the maximum 1D mipmapped texture texture size. - :py:obj:`~.maxTexture1DLinear` is the maximum 1D texture size for textures bound to linear memory. - :py:obj:`~.maxTexture2D[2]` contains the maximum 2D texture dimensions. - :py:obj:`~.maxTexture2DMipmap[2]` contains the maximum 2D mipmapped texture dimensions. - :py:obj:`~.maxTexture2DLinear[3]` contains the maximum 2D texture dimensions for 2D textures bound to pitch linear memory. - :py:obj:`~.maxTexture2DGather[2]` contains the maximum 2D texture dimensions if texture gather operations have to be performed. - :py:obj:`~.maxTexture3D[3]` contains the maximum 3D texture dimensions. - :py:obj:`~.maxTexture3DAlt[3]` contains the maximum alternate 3D texture dimensions. - :py:obj:`~.maxTextureCubemap` is the maximum cubemap texture width or height. - :py:obj:`~.maxTexture1DLayered[2]` contains the maximum 1D layered texture dimensions. - :py:obj:`~.maxTexture2DLayered[3]` contains the maximum 2D layered texture dimensions. - :py:obj:`~.maxTextureCubemapLayered[2]` contains the maximum cubemap layered texture dimensions. - :py:obj:`~.maxSurface1D` is the maximum 1D surface size. - :py:obj:`~.maxSurface2D[2]` contains the maximum 2D surface dimensions. - :py:obj:`~.maxSurface3D[3]` contains the maximum 3D surface dimensions. - :py:obj:`~.maxSurface1DLayered[2]` contains the maximum 1D layered surface dimensions. - :py:obj:`~.maxSurface2DLayered[3]` contains the maximum 2D layered surface dimensions. - :py:obj:`~.maxSurfaceCubemap` is the maximum cubemap surface width or height. - :py:obj:`~.maxSurfaceCubemapLayered[2]` contains the maximum cubemap layered surface dimensions. - :py:obj:`~.surfaceAlignment` specifies the alignment requirements for surfaces. - :py:obj:`~.concurrentKernels` is 1 if the device supports executing multiple kernels within the same context simultaneously, or 0 if not. It is not guaranteed that multiple kernels will be resident on the device concurrently so this feature should not be relied upon for correctness. - :py:obj:`~.ECCEnabled` is 1 if the device has ECC support turned on, or 0 if not. - :py:obj:`~.pciBusID` is the PCI bus identifier of the device. - :py:obj:`~.pciDeviceID` is the PCI device (sometimes called slot) identifier of the device. - :py:obj:`~.pciDomainID` is the PCI domain identifier of the device. - :py:obj:`~.tccDriver` is 1 if the device is using a TCC driver or 0 if not. - :py:obj:`~.asyncEngineCount` is 1 when the device can concurrently copy memory between host and device while executing a kernel. It is 2 when the device can concurrently copy memory between host and device in both directions and execute a kernel at the same time. It is 0 if neither of these is supported. - :py:obj:`~.unifiedAddressing` is 1 if the device shares a unified address space with the host and 0 otherwise. - :py:obj:`~.memoryClockRate` is the peak memory clock frequency in kilohertz. - :py:obj:`~.memoryBusWidth` is the memory bus width in bits. - :py:obj:`~.l2CacheSize` is L2 cache size in bytes. - :py:obj:`~.persistingL2CacheMaxSize` is L2 cache's maximum persisting lines size in bytes. - :py:obj:`~.maxThreadsPerMultiProcessor` is the number of maximum resident threads per multiprocessor. - :py:obj:`~.streamPrioritiesSupported` is 1 if the device supports stream priorities, or 0 if it is not supported. - :py:obj:`~.globalL1CacheSupported` is 1 if the device supports caching of globals in L1 cache, or 0 if it is not supported. - :py:obj:`~.localL1CacheSupported` is 1 if the device supports caching of locals in L1 cache, or 0 if it is not supported. - :py:obj:`~.sharedMemPerMultiprocessor` is the maximum amount of shared memory available to a multiprocessor in bytes; this amount is shared by all thread blocks simultaneously resident on a multiprocessor. - :py:obj:`~.regsPerMultiprocessor` is the maximum number of 32-bit registers available to a multiprocessor; this number is shared by all thread blocks simultaneously resident on a multiprocessor. - :py:obj:`~.managedMemory` is 1 if the device supports allocating managed memory on this system, or 0 if it is not supported. - :py:obj:`~.isMultiGpuBoard` is 1 if the device is on a multi-GPU board (e.g. Gemini cards), and 0 if not; - :py:obj:`~.multiGpuBoardGroupID` is a unique identifier for a group of devices associated with the same board. Devices on the same multi- GPU board will share the same identifier. - :py:obj:`~.hostNativeAtomicSupported` is 1 if the link between the device and the host supports native atomic operations, or 0 if it is not supported. - :py:obj:`~.singleToDoublePrecisionPerfRatio` is the ratio of single precision performance (in floating-point operations per second) to double precision performance. - :py:obj:`~.pageableMemoryAccess` is 1 if the device supports coherently accessing pageable memory without calling cudaHostRegister on it, and 0 otherwise. - :py:obj:`~.concurrentManagedAccess` is 1 if the device can coherently access managed memory concurrently with the CPU, and 0 otherwise. - :py:obj:`~.computePreemptionSupported` is 1 if the device supports Compute Preemption, and 0 otherwise. - :py:obj:`~.canUseHostPointerForRegisteredMem` is 1 if the device can access host registered memory at the same virtual address as the CPU, and 0 otherwise. - :py:obj:`~.cooperativeLaunch` is 1 if the device supports launching cooperative kernels via :py:obj:`~.cudaLaunchCooperativeKernel`, and 0 otherwise. - :py:obj:`~.cooperativeMultiDeviceLaunch` is 1 if the device supports launching cooperative kernels via :py:obj:`~.cudaLaunchCooperativeKernelMultiDevice`, and 0 otherwise. - :py:obj:`~.sharedMemPerBlockOptin` is the per device maximum shared memory per block usable by special opt in - :py:obj:`~.pageableMemoryAccessUsesHostPageTables` is 1 if the device accesses pageable memory via the host's page tables, and 0 otherwise. - :py:obj:`~.directManagedMemAccessFromHost` is 1 if the host can directly access managed memory on the device without migration, and 0 otherwise. - :py:obj:`~.maxBlocksPerMultiProcessor` is the maximum number of thread blocks that can reside on a multiprocessor. - :py:obj:`~.accessPolicyMaxWindowSize` is the maximum value of :py:obj:`~.cudaAccessPolicyWindow.num_bytes`. - :py:obj:`~.reservedSharedMemPerBlock` is the shared memory reserved by CUDA driver per block in bytes - :py:obj:`~.hostRegisterSupported` is 1 if the device supports host memory registration via :py:obj:`~.cudaHostRegister`, and 0 otherwise. - :py:obj:`~.sparseCudaArraySupported` is 1 if the device supports sparse CUDA arrays and sparse CUDA mipmapped arrays, 0 otherwise - :py:obj:`~.hostRegisterReadOnlySupported` is 1 if the device supports using the :py:obj:`~.cudaHostRegister` flag cudaHostRegisterReadOnly to register memory that must be mapped as read-only to the GPU - :py:obj:`~.timelineSemaphoreInteropSupported` is 1 if external timeline semaphore interop is supported on the device, 0 otherwise - :py:obj:`~.memoryPoolsSupported` is 1 if the device supports using the cudaMallocAsync and cudaMemPool family of APIs, 0 otherwise - :py:obj:`~.gpuDirectRDMASupported` is 1 if the device supports GPUDirect RDMA APIs, 0 otherwise - :py:obj:`~.gpuDirectRDMAFlushWritesOptions` is a bitmask to be interpreted according to the :py:obj:`~.cudaFlushGPUDirectRDMAWritesOptions` enum - :py:obj:`~.gpuDirectRDMAWritesOrdering` See the :py:obj:`~.cudaGPUDirectRDMAWritesOrdering` enum for numerical values - :py:obj:`~.memoryPoolSupportedHandleTypes` is a bitmask of handle types supported with mempool-based IPC - :py:obj:`~.deferredMappingCudaArraySupported` is 1 if the device supports deferred mapping CUDA arrays and CUDA mipmapped arrays - :py:obj:`~.ipcEventSupported` is 1 if the device supports IPC Events, and 0 otherwise - :py:obj:`~.unifiedFunctionPointers` is 1 if the device support unified pointers, and 0 otherwise Parameters ---------- device : int None Returns ------- cudaError_t prop : :py:obj:`~.cudaDeviceProp` None """ cdef cudaDeviceProp prop = cudaDeviceProp() with nogil: err = cyruntime.cudaGetDeviceProperties(prop._pvt_ptr, device) if err != cyruntime.cudaSuccess: return (_dict_cudaError_t[err], None) return (_dict_cudaError_t[err], prop) @cython.embedsignature(True) def cudaDeviceGetAttribute(attr not None : cudaDeviceAttr, int device): """ Returns information about the device. Returns in `*value` the integer value of the attribute `attr` on device `device`. The supported attributes are: - :py:obj:`~.cudaDevAttrMaxThreadsPerBlock`: Maximum number of threads per block - :py:obj:`~.cudaDevAttrMaxBlockDimX`: Maximum x-dimension of a block - :py:obj:`~.cudaDevAttrMaxBlockDimY`: Maximum y-dimension of a block - :py:obj:`~.cudaDevAttrMaxBlockDimZ`: Maximum z-dimension of a block - :py:obj:`~.cudaDevAttrMaxGridDimX`: Maximum x-dimension of a grid - :py:obj:`~.cudaDevAttrMaxGridDimY`: Maximum y-dimension of a grid - :py:obj:`~.cudaDevAttrMaxGridDimZ`: Maximum z-dimension of a grid - :py:obj:`~.cudaDevAttrMaxSharedMemoryPerBlock`: Maximum amount of shared memory available to a thread block in bytes - :py:obj:`~.cudaDevAttrTotalConstantMemory`: Memory available on device for constant variables in a CUDA C kernel in bytes - :py:obj:`~.cudaDevAttrWarpSize`: Warp size in threads - :py:obj:`~.cudaDevAttrMaxPitch`: Maximum pitch in bytes allowed by the memory copy functions that involve memory regions allocated through :py:obj:`~.cudaMallocPitch()` - :py:obj:`~.cudaDevAttrMaxTexture1DWidth`: Maximum 1D texture width - :py:obj:`~.cudaDevAttrMaxTexture1DLinearWidth`: Maximum width for a 1D texture bound to linear memory - :py:obj:`~.cudaDevAttrMaxTexture1DMipmappedWidth`: Maximum mipmapped 1D texture width - :py:obj:`~.cudaDevAttrMaxTexture2DWidth`: Maximum 2D texture width - :py:obj:`~.cudaDevAttrMaxTexture2DHeight`: Maximum 2D texture height - :py:obj:`~.cudaDevAttrMaxTexture2DLinearWidth`: Maximum width for a 2D texture bound to linear memory - :py:obj:`~.cudaDevAttrMaxTexture2DLinearHeight`: Maximum height for a 2D texture bound to linear memory - :py:obj:`~.cudaDevAttrMaxTexture2DLinearPitch`: Maximum pitch in bytes for a 2D texture bound to linear memory - :py:obj:`~.cudaDevAttrMaxTexture2DMipmappedWidth`: Maximum mipmapped 2D texture width - :py:obj:`~.cudaDevAttrMaxTexture2DMipmappedHeight`: Maximum mipmapped 2D texture height - :py:obj:`~.cudaDevAttrMaxTexture3DWidth`: Maximum 3D texture width - :py:obj:`~.cudaDevAttrMaxTexture3DHeight`: Maximum 3D texture height - :py:obj:`~.cudaDevAttrMaxTexture3DDepth`: Maximum 3D texture depth - :py:obj:`~.cudaDevAttrMaxTexture3DWidthAlt`: Alternate maximum 3D texture width, 0 if no alternate maximum 3D texture size is supported - :py:obj:`~.cudaDevAttrMaxTexture3DHeightAlt`: Alternate maximum 3D texture height, 0 if no alternate maximum 3D texture size is supported - :py:obj:`~.cudaDevAttrMaxTexture3DDepthAlt`: Alternate maximum 3D texture depth, 0 if no alternate maximum 3D texture size is supported - :py:obj:`~.cudaDevAttrMaxTextureCubemapWidth`: Maximum cubemap texture width or height - :py:obj:`~.cudaDevAttrMaxTexture1DLayeredWidth`: Maximum 1D layered texture width - :py:obj:`~.cudaDevAttrMaxTexture1DLayeredLayers`: Maximum layers in a 1D layered texture - :py:obj:`~.cudaDevAttrMaxTexture2DLayeredWidth`: Maximum 2D layered texture width - :py:obj:`~.cudaDevAttrMaxTexture2DLayeredHeight`: Maximum 2D layered texture height - :py:obj:`~.cudaDevAttrMaxTexture2DLayeredLayers`: Maximum layers in a 2D layered texture - :py:obj:`~.cudaDevAttrMaxTextureCubemapLayeredWidth`: Maximum cubemap layered texture width or height - :py:obj:`~.cudaDevAttrMaxTextureCubemapLayeredLayers`: Maximum layers in a cubemap layered texture - :py:obj:`~.cudaDevAttrMaxSurface1DWidth`: Maximum 1D surface width - :py:obj:`~.cudaDevAttrMaxSurface2DWidth`: Maximum 2D surface width - :py:obj:`~.cudaDevAttrMaxSurface2DHeight`: Maximum 2D surface height - :py:obj:`~.cudaDevAttrMaxSurface3DWidth`: Maximum 3D surface width - :py:obj:`~.cudaDevAttrMaxSurface3DHeight`: Maximum 3D surface height - :py:obj:`~.cudaDevAttrMaxSurface3DDepth`: Maximum 3D surface depth - :py:obj:`~.cudaDevAttrMaxSurface1DLayeredWidth`: Maximum 1D layered surface width - :py:obj:`~.cudaDevAttrMaxSurface1DLayeredLayers`: Maximum layers in a 1D layered surface - :py:obj:`~.cudaDevAttrMaxSurface2DLayeredWidth`: Maximum 2D layered surface width - :py:obj:`~.cudaDevAttrMaxSurface2DLayeredHeight`: Maximum 2D layered surface height - :py:obj:`~.cudaDevAttrMaxSurface2DLayeredLayers`: Maximum layers in a 2D layered surface - :py:obj:`~.cudaDevAttrMaxSurfaceCubemapWidth`: Maximum cubemap surface width - :py:obj:`~.cudaDevAttrMaxSurfaceCubemapLayeredWidth`: Maximum cubemap layered surface width - :py:obj:`~.cudaDevAttrMaxSurfaceCubemapLayeredLayers`: Maximum layers in a cubemap layered surface - :py:obj:`~.cudaDevAttrMaxRegistersPerBlock`: Maximum number of 32-bit registers available to a thread block - :py:obj:`~.cudaDevAttrClockRate`: Peak clock frequency in kilohertz - :py:obj:`~.cudaDevAttrTextureAlignment`: Alignment requirement; texture base addresses aligned to :py:obj:`~.textureAlign` bytes do not need an offset applied to texture fetches - :py:obj:`~.cudaDevAttrTexturePitchAlignment`: Pitch alignment requirement for 2D texture references bound to pitched memory - :py:obj:`~.cudaDevAttrGpuOverlap`: 1 if the device can concurrently copy memory between host and device while executing a kernel, or 0 if not - :py:obj:`~.cudaDevAttrMultiProcessorCount`: Number of multiprocessors on the device - :py:obj:`~.cudaDevAttrKernelExecTimeout`: 1 if there is a run time limit for kernels executed on the device, or 0 if not - :py:obj:`~.cudaDevAttrIntegrated`: 1 if the device is integrated with the memory subsystem, or 0 if not - :py:obj:`~.cudaDevAttrCanMapHostMemory`: 1 if the device can map host memory into the CUDA address space, or 0 if not - :py:obj:`~.cudaDevAttrComputeMode`: Compute mode is the compute mode that the device is currently in. Available modes are as follows: - :py:obj:`~.cudaComputeModeDefault`: Default mode - Device is not restricted and multiple threads can use :py:obj:`~.cudaSetDevice()` with this device. - :py:obj:`~.cudaComputeModeProhibited`: Compute-prohibited mode - No threads can use :py:obj:`~.cudaSetDevice()` with this device. - :py:obj:`~.cudaComputeModeExclusiveProcess`: Compute-exclusive- process mode - Many threads in one process will be able to use :py:obj:`~.cudaSetDevice()` with this device. - :py:obj:`~.cudaDevAttrConcurrentKernels`: 1 if the device supports executing multiple kernels within the same context simultaneously, or 0 if not. It is not guaranteed that multiple kernels will be resident on the device concurrently so this feature should not be relied upon for correctness. - :py:obj:`~.cudaDevAttrEccEnabled`: 1 if error correction is enabled on the device, 0 if error correction is disabled or not supported by the device - :py:obj:`~.cudaDevAttrPciBusId`: PCI bus identifier of the device - :py:obj:`~.cudaDevAttrPciDeviceId`: PCI device (also known as slot) identifier of the device - :py:obj:`~.cudaDevAttrTccDriver`: 1 if the device is using a TCC driver. TCC is only available on Tesla hardware running Windows Vista or later. - :py:obj:`~.cudaDevAttrMemoryClockRate`: Peak memory clock frequency in kilohertz - :py:obj:`~.cudaDevAttrGlobalMemoryBusWidth`: Global memory bus width in bits - :py:obj:`~.cudaDevAttrL2CacheSize`: Size of L2 cache in bytes. 0 if the device doesn't have L2 cache. - :py:obj:`~.cudaDevAttrMaxThreadsPerMultiProcessor`: Maximum resident threads per multiprocessor - :py:obj:`~.cudaDevAttrUnifiedAddressing`: 1 if the device shares a unified address space with the host, or 0 if not - :py:obj:`~.cudaDevAttrComputeCapabilityMajor`: Major compute capability version number - :py:obj:`~.cudaDevAttrComputeCapabilityMinor`: Minor compute capability version number - :py:obj:`~.cudaDevAttrStreamPrioritiesSupported`: 1 if the device supports stream priorities, or 0 if not - :py:obj:`~.cudaDevAttrGlobalL1CacheSupported`: 1 if device supports caching globals in L1 cache, 0 if not - :py:obj:`~.cudaDevAttrLocalL1CacheSupported`: 1 if device supports caching locals in L1 cache, 0 if not - :py:obj:`~.cudaDevAttrMaxSharedMemoryPerMultiprocessor`: Maximum amount of shared memory available to a multiprocessor in bytes; this amount is shared by all thread blocks simultaneously resident on a multiprocessor - :py:obj:`~.cudaDevAttrMaxRegistersPerMultiprocessor`: Maximum number of 32-bit registers available to a multiprocessor; this number is shared by all thread blocks simultaneously resident on a multiprocessor - :py:obj:`~.cudaDevAttrManagedMemory`: 1 if device supports allocating managed memory, 0 if not - :py:obj:`~.cudaDevAttrIsMultiGpuBoard`: 1 if device is on a multi-GPU board, 0 if not - :py:obj:`~.cudaDevAttrMultiGpuBoardGroupID`: Unique identifier for a group of devices on the same multi-GPU board - :py:obj:`~.cudaDevAttrHostNativeAtomicSupported`: 1 if the link between the device and the host supports native atomic operations - :py:obj:`~.cudaDevAttrSingleToDoublePrecisionPerfRatio`: Ratio of single precision performance (in floating-point operations per second) to double precision performance - :py:obj:`~.cudaDevAttrPageableMemoryAccess`: 1 if the device supports coherently accessing pageable memory without calling cudaHostRegister on it, and 0 otherwise - :py:obj:`~.cudaDevAttrConcurrentManagedAccess`: 1 if the device can coherently access managed memory concurrently with the CPU, and 0 otherwise - :py:obj:`~.cudaDevAttrComputePreemptionSupported`: 1 if the device supports Compute Preemption, 0 if not - :py:obj:`~.cudaDevAttrCanUseHostPointerForRegisteredMem`: 1 if the device can access host registered memory at the same virtual address as the CPU, and 0 otherwise - :py:obj:`~.cudaDevAttrCooperativeLaunch`: 1 if the device supports launching cooperative kernels via :py:obj:`~.cudaLaunchCooperativeKernel`, and 0 otherwise - :py:obj:`~.cudaDevAttrCooperativeMultiDeviceLaunch`: 1 if the device supports launching cooperative kernels via :py:obj:`~.cudaLaunchCooperativeKernelMultiDevice`, and 0 otherwise - :py:obj:`~.cudaDevAttrCanFlushRemoteWrites`: 1 if the device supports flushing of outstanding remote writes, and 0 otherwise - :py:obj:`~.cudaDevAttrHostRegisterSupported`: 1 if the device supports host memory registration via :py:obj:`~.cudaHostRegister`, and 0 otherwise - :py:obj:`~.cudaDevAttrPageableMemoryAccessUsesHostPageTables`: 1 if the device accesses pageable memory via the host's page tables, and 0 otherwise - :py:obj:`~.cudaDevAttrDirectManagedMemAccessFromHost`: 1 if the host can directly access managed memory on the device without migration, and 0 otherwise - :py:obj:`~.cudaDevAttrMaxSharedMemoryPerBlockOptin`: Maximum per block shared memory size on the device. This value can be opted into when using :py:obj:`~.cudaFuncSetAttribute` - :py:obj:`~.cudaDevAttrMaxBlocksPerMultiprocessor`: Maximum number of thread blocks that can reside on a multiprocessor - :py:obj:`~.cudaDevAttrMaxPersistingL2CacheSize`: Maximum L2 persisting lines capacity setting in bytes - :py:obj:`~.cudaDevAttrMaxAccessPolicyWindowSize`: Maximum value of :py:obj:`~.cudaAccessPolicyWindow.num_bytes` - :py:obj:`~.cudaDevAttrReservedSharedMemoryPerBlock`: Shared memory reserved by CUDA driver per block in bytes - :py:obj:`~.cudaDevAttrSparseCudaArraySupported`: 1 if the device supports sparse CUDA arrays and sparse CUDA mipmapped arrays. - :py:obj:`~.cudaDevAttrHostRegisterReadOnlySupported`: Device supports using the :py:obj:`~.cudaHostRegister` flag cudaHostRegisterReadOnly to register memory that must be mapped as read-only to the GPU - :py:obj:`~.cudaDevAttrMemoryPoolsSupported`: 1 if the device supports using the cudaMallocAsync and cudaMemPool family of APIs, and 0 otherwise - :py:obj:`~.cudaDevAttrGPUDirectRDMASupported`: 1 if the device supports GPUDirect RDMA APIs, and 0 otherwise - :py:obj:`~.cudaDevAttrGPUDirectRDMAFlushWritesOptions`: bitmask to be interpreted according to the :py:obj:`~.cudaFlushGPUDirectRDMAWritesOptions` enum - :py:obj:`~.cudaDevAttrGPUDirectRDMAWritesOrdering`: see the :py:obj:`~.cudaGPUDirectRDMAWritesOrdering` enum for numerical values - :py:obj:`~.cudaDevAttrMemoryPoolSupportedHandleTypes`: Bitmask of handle types supported with mempool based IPC - :py:obj:`~.cudaDevAttrDeferredMappingCudaArraySupported` : 1 if the device supports deferred mapping CUDA arrays and CUDA mipmapped arrays. - :py:obj:`~.cudaDevAttrIpcEventSupport`: 1 if the device supports IPC Events. - :py:obj:`~.cudaDevAttrNumaConfig`: NUMA configuration of a device: value is of type :py:obj:`~.cudaDeviceNumaConfig` enum - :py:obj:`~.cudaDevAttrNumaId`: NUMA node ID of the GPU memory - :py:obj:`~.cudaDevAttrGpuPciDeviceId`: The combined 16-bit PCI device ID and 16-bit PCI vendor ID. - :py:obj:`~.cudaDevAttrGpuPciSubsystemId`: The combined 16-bit PCI subsystem ID and 16-bit PCI vendor subsystem ID. Parameters ---------- attr : :py:obj:`~.cudaDeviceAttr` Device attribute to query device : int Device number to query Returns ------- cudaError_t :py:obj:`~.cudaSuccess`, :py:obj:`~.cudaErrorInvalidDevice`, :py:obj:`~.cudaErrorInvalidValue` value : int Returned device attribute value See Also -------- :py:obj:`~.cudaGetDeviceCount`, :py:obj:`~.cudaGetDevice`, :py:obj:`~.cudaSetDevice`, :py:obj:`~.cudaChooseDevice`, :py:obj:`~.cudaGetDeviceProperties`, :py:obj:`~.cudaInitDevice`, :py:obj:`~.cuDeviceGetAttribute` """ cdef int value = 0 cdef cyruntime.cudaDeviceAttr cyattr = attr.value with nogil: err = cyruntime.cudaDeviceGetAttribute(&value, cyattr, device) if err != cyruntime.cudaSuccess: return (_dict_cudaError_t[err], None) return (_dict_cudaError_t[err], value) @cython.embedsignature(True) def cudaDeviceGetDefaultMemPool(int device): """ Returns the default mempool of a device. The default mempool of a device contains device memory from that device. Parameters ---------- device : int None Returns ------- cudaError_t :py:obj:`~.cudaSuccess`, :py:obj:`~.cudaErrorInvalidDevice`, :py:obj:`~.cudaErrorInvalidValue` :py:obj:`~.cudaErrorNotSupported` memPool : :py:obj:`~.cudaMemPool_t` None See Also -------- :py:obj:`~.cuDeviceGetDefaultMemPool`, :py:obj:`~.cudaMallocAsync`, :py:obj:`~.cudaMemPoolTrimTo`, :py:obj:`~.cudaMemPoolGetAttribute`, :py:obj:`~.cudaDeviceSetMemPool`, :py:obj:`~.cudaMemPoolSetAttribute`, :py:obj:`~.cudaMemPoolSetAccess` """ cdef cudaMemPool_t memPool = cudaMemPool_t() with nogil: err = cyruntime.cudaDeviceGetDefaultMemPool(memPool._pvt_ptr, device) if err != cyruntime.cudaSuccess: return (_dict_cudaError_t[err], None) return (_dict_cudaError_t[err], memPool) @cython.embedsignature(True) def cudaDeviceSetMemPool(int device, memPool): """ Sets the current memory pool of a device. The memory pool must be local to the specified device. Unless a mempool is specified in the :py:obj:`~.cudaMallocAsync` call, :py:obj:`~.cudaMallocAsync` allocates from the current mempool of the provided stream's device. By default, a device's current memory pool is its default memory pool. Parameters ---------- device : int None memPool : :py:obj:`~.CUmemoryPool` or :py:obj:`~.cudaMemPool_t` None Returns ------- cudaError_t :py:obj:`~.cudaSuccess`, :py:obj:`~.cudaErrorInvalidValue` :py:obj:`~.cudaErrorInvalidDevice` :py:obj:`~.cudaErrorNotSupported` See Also -------- :py:obj:`~.cuDeviceSetMemPool`, :py:obj:`~.cudaDeviceGetMemPool`, :py:obj:`~.cudaDeviceGetDefaultMemPool`, :py:obj:`~.cudaMemPoolCreate`, :py:obj:`~.cudaMemPoolDestroy`, :py:obj:`~.cudaMallocFromPoolAsync` Notes ----- Use :py:obj:`~.cudaMallocFromPoolAsync` to specify asynchronous allocations from a device different than the one the stream runs on. """ cdef cyruntime.cudaMemPool_t cymemPool if memPool is None: pmemPool = 0 elif isinstance(memPool, (cudaMemPool_t,driver.CUmemoryPool)): pmemPool = int(memPool) else: pmemPool = int(cudaMemPool_t(memPool)) cymemPool = pmemPool with nogil: err = cyruntime.cudaDeviceSetMemPool(device, cymemPool) return (_dict_cudaError_t[err],) @cython.embedsignature(True) def cudaDeviceGetMemPool(int device): """ Gets the current mempool for a device. Returns the last pool provided to :py:obj:`~.cudaDeviceSetMemPool` for this device or the device's default memory pool if :py:obj:`~.cudaDeviceSetMemPool` has never been called. By default the current mempool is the default mempool for a device, otherwise the returned pool must have been set with :py:obj:`~.cuDeviceSetMemPool` or :py:obj:`~.cudaDeviceSetMemPool`. Parameters ---------- device : int None Returns ------- cudaError_t :py:obj:`~.cudaSuccess`, :py:obj:`~.cudaErrorInvalidValue` :py:obj:`~.cudaErrorNotSupported` memPool : :py:obj:`~.cudaMemPool_t` None See Also -------- :py:obj:`~.cuDeviceGetMemPool`, :py:obj:`~.cudaDeviceGetDefaultMemPool`, :py:obj:`~.cudaDeviceSetMemPool` """ cdef cudaMemPool_t memPool = cudaMemPool_t() with nogil: err = cyruntime.cudaDeviceGetMemPool(memPool._pvt_ptr, device) if err != cyruntime.cudaSuccess: return (_dict_cudaError_t[err], None) return (_dict_cudaError_t[err], memPool) @cython.embedsignature(True) def cudaDeviceGetNvSciSyncAttributes(nvSciSyncAttrList, int device, int flags): """ Return NvSciSync attributes that this device can support. Returns in `nvSciSyncAttrList`, the properties of NvSciSync that this CUDA device, `dev` can support. The returned `nvSciSyncAttrList` can be used to create an NvSciSync that matches this device's capabilities. If NvSciSyncAttrKey_RequiredPerm field in `nvSciSyncAttrList` is already set this API will return :py:obj:`~.cudaErrorInvalidValue`. The applications should set `nvSciSyncAttrList` to a valid NvSciSyncAttrList failing which this API will return :py:obj:`~.cudaErrorInvalidHandle`. The `flags` controls how applications intends to use the NvSciSync created from the `nvSciSyncAttrList`. The valid flags are: - :py:obj:`~.cudaNvSciSyncAttrSignal`, specifies that the applications intends to signal an NvSciSync on this CUDA device. - :py:obj:`~.cudaNvSciSyncAttrWait`, specifies that the applications intends to wait on an NvSciSync on this CUDA device. At least one of these flags must be set, failing which the API returns :py:obj:`~.cudaErrorInvalidValue`. Both the flags are orthogonal to one another: a developer may set both these flags that allows to set both wait and signal specific attributes in the same `nvSciSyncAttrList`. Note that this API updates the input `nvSciSyncAttrList` with values equivalent to the following public attribute key-values: NvSciSyncAttrKey_RequiredPerm is set to - NvSciSyncAccessPerm_SignalOnly if :py:obj:`~.cudaNvSciSyncAttrSignal` is set in `flags`. - NvSciSyncAccessPerm_WaitOnly if :py:obj:`~.cudaNvSciSyncAttrWait` is set in `flags`. - NvSciSyncAccessPerm_WaitSignal if both :py:obj:`~.cudaNvSciSyncAttrWait` and :py:obj:`~.cudaNvSciSyncAttrSignal` are set in `flags`. NvSciSyncAttrKey_PrimitiveInfo is set to - NvSciSyncAttrValPrimitiveType_SysmemSemaphore on any valid `device`. - NvSciSyncAttrValPrimitiveType_Syncpoint if `device` is a Tegra device. - NvSciSyncAttrValPrimitiveType_SysmemSemaphorePayload64b if `device` is GA10X+. NvSciSyncAttrKey_GpuId is set to the same UUID that is returned in `None` from :py:obj:`~.cudaDeviceGetProperties` for this `device`. :py:obj:`~.cudaSuccess`, :py:obj:`~.cudaErrorDeviceUninitialized`, :py:obj:`~.cudaErrorInvalidValue`, :py:obj:`~.cudaErrorInvalidHandle`, :py:obj:`~.cudaErrorInvalidDevice`, :py:obj:`~.cudaErrorNotSupported`, :py:obj:`~.cudaErrorMemoryAllocation` Parameters ---------- nvSciSyncAttrList : Any Return NvSciSync attributes supported. device : int Valid Cuda Device to get NvSciSync attributes for. flags : int flags describing NvSciSync usage. Returns ------- cudaError_t See Also -------- :py:obj:`~.cudaImportExternalSemaphore`, :py:obj:`~.cudaDestroyExternalSemaphore`, :py:obj:`~.cudaSignalExternalSemaphoresAsync`, :py:obj:`~.cudaWaitExternalSemaphoresAsync` """ cynvSciSyncAttrList = _HelperInputVoidPtr(nvSciSyncAttrList) cdef void* cynvSciSyncAttrList_ptr = cynvSciSyncAttrList.cptr with nogil: err = cyruntime.cudaDeviceGetNvSciSyncAttributes(cynvSciSyncAttrList_ptr, device, flags) return (_dict_cudaError_t[err],) @cython.embedsignature(True) def cudaDeviceGetP2PAttribute(attr not None : cudaDeviceP2PAttr, int srcDevice, int dstDevice): """ Queries attributes of the link between two devices. Returns in `*value` the value of the requested attribute `attrib` of the link between `srcDevice` and `dstDevice`. The supported attributes are: - :py:obj:`~.cudaDevP2PAttrPerformanceRank`: A relative value indicating the performance of the link between two devices. Lower value means better performance (0 being the value used for most performant link). - :py:obj:`~.cudaDevP2PAttrAccessSupported`: 1 if peer access is enabled. - :py:obj:`~.cudaDevP2PAttrNativeAtomicSupported`: 1 if native atomic operations over the link are supported. - :py:obj:`~.cudaDevP2PAttrCudaArrayAccessSupported`: 1 if accessing CUDA arrays over the link is supported. Returns :py:obj:`~.cudaErrorInvalidDevice` if `srcDevice` or `dstDevice` are not valid or if they represent the same device. Returns :py:obj:`~.cudaErrorInvalidValue` if `attrib` is not valid or if `value` is a null pointer. Parameters ---------- attrib : :py:obj:`~.cudaDeviceP2PAttr` The requested attribute of the link between `srcDevice` and `dstDevice`. srcDevice : int The source device of the target link. dstDevice : int The destination device of the target link. Returns ------- cudaError_t :py:obj:`~.cudaSuccess`, :py:obj:`~.cudaErrorInvalidDevice`, :py:obj:`~.cudaErrorInvalidValue` value : int Returned value of the requested attribute See Also -------- :py:obj:`~.cudaDeviceEnablePeerAccess`, :py:obj:`~.cudaDeviceDisablePeerAccess`, :py:obj:`~.cudaDeviceCanAccessPeer`, :py:obj:`~.cuDeviceGetP2PAttribute` """ cdef int value = 0 cdef cyruntime.cudaDeviceP2PAttr cyattr = attr.value with nogil: err = cyruntime.cudaDeviceGetP2PAttribute(&value, cyattr, srcDevice, dstDevice) if err != cyruntime.cudaSuccess: return (_dict_cudaError_t[err], None) return (_dict_cudaError_t[err], value) @cython.embedsignature(True) def cudaChooseDevice(prop : Optional[cudaDeviceProp]): """ Select compute-device which best matches criteria. Returns in `*device` the device which has properties that best match `*prop`. Parameters ---------- prop : :py:obj:`~.cudaDeviceProp` Desired device properties Returns ------- cudaError_t :py:obj:`~.cudaSuccess`, :py:obj:`~.cudaErrorInvalidValue` device : int Device with best match See Also -------- :py:obj:`~.cudaGetDeviceCount`, :py:obj:`~.cudaGetDevice`, :py:obj:`~.cudaSetDevice`, :py:obj:`~.cudaGetDeviceProperties`, :py:obj:`~.cudaInitDevice` """ cdef int device = 0 cdef cyruntime.cudaDeviceProp* cyprop_ptr = prop._pvt_ptr if prop is not None else NULL with nogil: err = cyruntime.cudaChooseDevice(&device, cyprop_ptr) if err != cyruntime.cudaSuccess: return (_dict_cudaError_t[err], None) return (_dict_cudaError_t[err], device) @cython.embedsignature(True) def cudaInitDevice(int device, unsigned int deviceFlags, unsigned int flags): """ Initialize device to be used for GPU executions. This function will initialize the CUDA Runtime structures and primary context on `device` when called, but the context will not be made current to `device`. When :py:obj:`~.cudaInitDeviceFlagsAreValid` is set in `flags`, deviceFlags are applied to the requested device. The values of deviceFlags match those of the flags parameters in :py:obj:`~.cudaSetDeviceFlags`. The effect may be verified by :py:obj:`~.cudaGetDeviceFlags`. This function will return an error if the device is in :py:obj:`~.cudaComputeModeExclusiveProcess` and is occupied by another process or if the device is in :py:obj:`~.cudaComputeModeProhibited`. Parameters ---------- device : int Device on which the runtime will initialize itself. deviceFlags : unsigned int Parameters for device operation. flags : unsigned int Flags for controlling the device initialization. Returns ------- cudaError_t :py:obj:`~.cudaSuccess`, :py:obj:`~.cudaErrorInvalidDevice`, See Also -------- :py:obj:`~.cudaGetDeviceCount`, :py:obj:`~.cudaGetDevice`, :py:obj:`~.cudaGetDeviceProperties`, :py:obj:`~.cudaChooseDevice`, :py:obj:`~.cudaSetDevice` :py:obj:`~.cuCtxSetCurrent` """ with nogil: err = cyruntime.cudaInitDevice(device, deviceFlags, flags) return (_dict_cudaError_t[err],) @cython.embedsignature(True) def cudaSetDevice(int device): """ Set device to be used for GPU executions. Sets `device` as the current device for the calling host thread. Valid device id's are 0 to (:py:obj:`~.cudaGetDeviceCount()` - 1). Any device memory subsequently allocated from this host thread using :py:obj:`~.cudaMalloc()`, :py:obj:`~.cudaMallocPitch()` or :py:obj:`~.cudaMallocArray()` will be physically resident on `device`. Any host memory allocated from this host thread using :py:obj:`~.cudaMallocHost()` or :py:obj:`~.cudaHostAlloc()` or :py:obj:`~.cudaHostRegister()` will have its lifetime associated with `device`. Any streams or events created from this host thread will be associated with `device`. Any kernels launched from this host thread using the <<<>>> operator or :py:obj:`~.cudaLaunchKernel()` will be executed on `device`. This call may be made from any host thread, to any device, and at any time. This function will do no synchronization with the previous or new device, and should only take significant time when it initializes the runtime's context state. This call will bind the primary context of the specified device to the calling thread and all the subsequent memory allocations, stream and event creations, and kernel launches will be associated with the primary context. This function will also immediately initialize the runtime state on the primary context, and the context will be current on `device` immediately. This function will return an error if the device is in :py:obj:`~.cudaComputeModeExclusiveProcess` and is occupied by another process or if the device is in :py:obj:`~.cudaComputeModeProhibited`. It is not required to call :py:obj:`~.cudaInitDevice` before using this function. Parameters ---------- device : int Device on which the active host thread should execute the device code. Returns ------- cudaError_t :py:obj:`~.cudaSuccess`, :py:obj:`~.cudaErrorInvalidDevice`, :py:obj:`~.cudaErrorDeviceUnavailable`, See Also -------- :py:obj:`~.cudaGetDeviceCount`, :py:obj:`~.cudaGetDevice`, :py:obj:`~.cudaGetDeviceProperties`, :py:obj:`~.cudaChooseDevice`, :py:obj:`~.cudaInitDevice`, :py:obj:`~.cuCtxSetCurrent` """ with nogil: err = cyruntime.cudaSetDevice(device) return (_dict_cudaError_t[err],) @cython.embedsignature(True) def cudaGetDevice(): """ Returns which device is currently being used. Returns in `*device` the current device for the calling host thread. Returns ------- cudaError_t :py:obj:`~.cudaSuccess`, :py:obj:`~.cudaErrorInvalidValue`, :py:obj:`~.cudaErrorDeviceUnavailable`, device : int Returns the device on which the active host thread executes the device code. See Also -------- :py:obj:`~.cudaGetDeviceCount`, :py:obj:`~.cudaSetDevice`, :py:obj:`~.cudaGetDeviceProperties`, :py:obj:`~.cudaChooseDevice`, :py:obj:`~.cuCtxGetCurrent` """ cdef int device = 0 with nogil: err = cyruntime.cudaGetDevice(&device) if err != cyruntime.cudaSuccess: return (_dict_cudaError_t[err], None) return (_dict_cudaError_t[err], device) @cython.embedsignature(True) def cudaSetDeviceFlags(unsigned int flags): """ Sets flags to be used for device executions. Records `flags` as the flags for the current device. If the current device has been set and that device has already been initialized, the previous flags are overwritten. If the current device has not been initialized, it is initialized with the provided flags. If no device has been made current to the calling thread, a default device is selected and initialized with the provided flags. The three LSBs of the `flags` parameter can be used to control how the CPU thread interacts with the OS scheduler when waiting for results from the device. - :py:obj:`~.cudaDeviceScheduleAuto`: The default value if the `flags` parameter is zero, uses a heuristic based on the number of active CUDA contexts in the process `C` and the number of logical processors in the system `P`. If `C` > `P`, then CUDA will yield to other OS threads when waiting for the device, otherwise CUDA will not yield while waiting for results and actively spin on the processor. Additionally, on Tegra devices, :py:obj:`~.cudaDeviceScheduleAuto` uses a heuristic based on the power profile of the platform and may choose :py:obj:`~.cudaDeviceScheduleBlockingSync` for low-powered devices. - :py:obj:`~.cudaDeviceScheduleSpin`: Instruct CUDA to actively spin when waiting for results from the device. This can decrease latency when waiting for the device, but may lower the performance of CPU threads if they are performing work in parallel with the CUDA thread. - :py:obj:`~.cudaDeviceScheduleYield`: Instruct CUDA to yield its thread when waiting for results from the device. This can increase latency when waiting for the device, but can increase the performance of CPU threads performing work in parallel with the device. - :py:obj:`~.cudaDeviceScheduleBlockingSync`: Instruct CUDA to block the CPU thread on a synchronization primitive when waiting for the device to finish work. - :py:obj:`~.cudaDeviceBlockingSync`: Instruct CUDA to block the CPU thread on a synchronization primitive when waiting for the device to finish work. :py:obj:`~.Deprecated:` This flag was deprecated as of CUDA 4.0 and replaced with :py:obj:`~.cudaDeviceScheduleBlockingSync`. - :py:obj:`~.cudaDeviceMapHost`: This flag enables allocating pinned host memory that is accessible to the device. It is implicit for the runtime but may be absent if a context is created using the driver API. If this flag is not set, :py:obj:`~.cudaHostGetDevicePointer()` will always return a failure code. - :py:obj:`~.cudaDeviceLmemResizeToMax`: Instruct CUDA to not reduce local memory after resizing local memory for a kernel. This can prevent thrashing by local memory allocations when launching many kernels with high local memory usage at the cost of potentially increased memory usage. :py:obj:`~.Deprecated:` This flag is deprecated and the behavior enabled by this flag is now the default and cannot be disabled. - :py:obj:`~.cudaDeviceSyncMemops`: Ensures that synchronous memory operations initiated on this context will always synchronize. See further documentation in the section titled "API Synchronization behavior" to learn more about cases when synchronous memory operations can exhibit asynchronous behavior. Parameters ---------- flags : unsigned int Parameters for device operation Returns ------- cudaError_t :py:obj:`~.cudaSuccess`, :py:obj:`~.cudaErrorInvalidValue` See Also -------- :py:obj:`~.cudaGetDeviceFlags`, :py:obj:`~.cudaGetDeviceCount`, :py:obj:`~.cudaGetDevice`, :py:obj:`~.cudaGetDeviceProperties`, :py:obj:`~.cudaSetDevice`, :py:obj:`~.cudaSetValidDevices`, :py:obj:`~.cudaInitDevice`, :py:obj:`~.cudaChooseDevice`, :py:obj:`~.cuDevicePrimaryCtxSetFlags` """ with nogil: err = cyruntime.cudaSetDeviceFlags(flags) return (_dict_cudaError_t[err],) @cython.embedsignature(True) def cudaGetDeviceFlags(): """ Gets the flags for the current device. Returns in `flags` the flags for the current device. If there is a current device for the calling thread, the flags for the device are returned. If there is no current device, the flags for the first device are returned, which may be the default flags. Compare to the behavior of :py:obj:`~.cudaSetDeviceFlags`. Typically, the flags returned should match the behavior that will be seen if the calling thread uses a device after this call, without any change to the flags or current device inbetween by this or another thread. Note that if the device is not initialized, it is possible for another thread to change the flags for the current device before it is initialized. Additionally, when using exclusive mode, if this thread has not requested a specific device, it may use a device other than the first device, contrary to the assumption made by this function. If a context has been created via the driver API and is current to the calling thread, the flags for that context are always returned. Flags returned by this function may specifically include :py:obj:`~.cudaDeviceMapHost` even though it is not accepted by :py:obj:`~.cudaSetDeviceFlags` because it is implicit in runtime API flags. The reason for this is that the current context may have been created via the driver API in which case the flag is not implicit and may be unset. Returns ------- cudaError_t :py:obj:`~.cudaSuccess`, :py:obj:`~.cudaErrorInvalidDevice` flags : unsigned int Pointer to store the device flags See Also -------- :py:obj:`~.cudaGetDevice`, :py:obj:`~.cudaGetDeviceProperties`, :py:obj:`~.cudaSetDevice`, :py:obj:`~.cudaSetDeviceFlags`, :py:obj:`~.cudaInitDevice`, :py:obj:`~.cuCtxGetFlags`, :py:obj:`~.cuDevicePrimaryCtxGetState` """ cdef unsigned int flags = 0 with nogil: err = cyruntime.cudaGetDeviceFlags(&flags) if err != cyruntime.cudaSuccess: return (_dict_cudaError_t[err], None) return (_dict_cudaError_t[err], flags) @cython.embedsignature(True) def cudaStreamCreate(): """ Create an asynchronous stream. Creates a new asynchronous stream on the context that is current to the calling host thread. If no context is current to the calling host thread, then the primary context for a device is selected, made current to the calling thread, and initialized before creating a stream on it. Returns ------- cudaError_t :py:obj:`~.cudaSuccess`, :py:obj:`~.cudaErrorInvalidValue` pStream : :py:obj:`~.cudaStream_t` Pointer to new stream identifier See Also -------- :py:obj:`~.cudaStreamCreateWithPriority`, :py:obj:`~.cudaStreamCreateWithFlags`, :py:obj:`~.cudaStreamGetPriority`, :py:obj:`~.cudaStreamGetFlags`, :py:obj:`~.cudaStreamGetDevice`, :py:obj:`~.cudaStreamQuery`, :py:obj:`~.cudaStreamSynchronize`, :py:obj:`~.cudaStreamWaitEvent`, :py:obj:`~.cudaStreamAddCallback`, :py:obj:`~.cudaSetDevice`, :py:obj:`~.cudaStreamDestroy`, :py:obj:`~.cuStreamCreate` """ cdef cudaStream_t pStream = cudaStream_t() with nogil: err = cyruntime.cudaStreamCreate(pStream._pvt_ptr) if err != cyruntime.cudaSuccess: return (_dict_cudaError_t[err], None) return (_dict_cudaError_t[err], pStream) @cython.embedsignature(True) def cudaStreamCreateWithFlags(unsigned int flags): """ Create an asynchronous stream. Creates a new asynchronous stream on the context that is current to the calling host thread. If no context is current to the calling host thread, then the primary context for a device is selected, made current to the calling thread, and initialized before creating a stream on it. The `flags` argument determines the behaviors of the stream. Valid values for `flags` are - :py:obj:`~.cudaStreamDefault`: Default stream creation flag. - :py:obj:`~.cudaStreamNonBlocking`: Specifies that work running in the created stream may run concurrently with work in stream 0 (the NULL stream), and that the created stream should perform no implicit synchronization with stream 0. Parameters ---------- flags : unsigned int Parameters for stream creation Returns ------- cudaError_t :py:obj:`~.cudaSuccess`, :py:obj:`~.cudaErrorInvalidValue` pStream : :py:obj:`~.cudaStream_t` Pointer to new stream identifier See Also -------- :py:obj:`~.cudaStreamCreate`, :py:obj:`~.cudaStreamCreateWithPriority`, :py:obj:`~.cudaStreamGetFlags`, :py:obj:`~.cudaStreamGetDevice`, :py:obj:`~.cudaStreamQuery`, :py:obj:`~.cudaStreamSynchronize`, :py:obj:`~.cudaStreamWaitEvent`, :py:obj:`~.cudaStreamAddCallback`, :py:obj:`~.cudaSetDevice`, :py:obj:`~.cudaStreamDestroy`, :py:obj:`~.cuStreamCreate` """ cdef cudaStream_t pStream = cudaStream_t() with nogil: err = cyruntime.cudaStreamCreateWithFlags(pStream._pvt_ptr, flags) if err != cyruntime.cudaSuccess: return (_dict_cudaError_t[err], None) return (_dict_cudaError_t[err], pStream) @cython.embedsignature(True) def cudaStreamCreateWithPriority(unsigned int flags, int priority): """ Create an asynchronous stream with the specified priority. Creates a stream with the specified priority and returns a handle in `pStream`. The stream is created on the context that is current to the calling host thread. If no context is current to the calling host thread, then the primary context for a device is selected, made current to the calling thread, and initialized before creating a stream on it. This affects the scheduling priority of work in the stream. Priorities provide a hint to preferentially run work with higher priority when possible, but do not preempt already-running work or provide any other functional guarantee on execution order. `priority` follows a convention where lower numbers represent higher priorities. '0' represents default priority. The range of meaningful numerical priorities can be queried using :py:obj:`~.cudaDeviceGetStreamPriorityRange`. If the specified priority is outside the numerical range returned by :py:obj:`~.cudaDeviceGetStreamPriorityRange`, it will automatically be clamped to the lowest or the highest number in the range. Parameters ---------- flags : unsigned int Flags for stream creation. See :py:obj:`~.cudaStreamCreateWithFlags` for a list of valid flags that can be passed priority : int Priority of the stream. Lower numbers represent higher priorities. See :py:obj:`~.cudaDeviceGetStreamPriorityRange` for more information about the meaningful stream priorities that can be passed. Returns ------- cudaError_t :py:obj:`~.cudaSuccess`, :py:obj:`~.cudaErrorInvalidValue` pStream : :py:obj:`~.cudaStream_t` Pointer to new stream identifier See Also -------- :py:obj:`~.cudaStreamCreate`, :py:obj:`~.cudaStreamCreateWithFlags`, :py:obj:`~.cudaDeviceGetStreamPriorityRange`, :py:obj:`~.cudaStreamGetPriority`, :py:obj:`~.cudaStreamQuery`, :py:obj:`~.cudaStreamWaitEvent`, :py:obj:`~.cudaStreamAddCallback`, :py:obj:`~.cudaStreamSynchronize`, :py:obj:`~.cudaSetDevice`, :py:obj:`~.cudaStreamDestroy`, :py:obj:`~.cuStreamCreateWithPriority` Notes ----- Stream priorities are supported only on GPUs with compute capability 3.5 or higher. In the current implementation, only compute kernels launched in priority streams are affected by the stream's priority. Stream priorities have no effect on host-to-device and device-to-host memory operations. """ cdef cudaStream_t pStream = cudaStream_t() with nogil: err = cyruntime.cudaStreamCreateWithPriority(pStream._pvt_ptr, flags, priority) if err != cyruntime.cudaSuccess: return (_dict_cudaError_t[err], None) return (_dict_cudaError_t[err], pStream) @cython.embedsignature(True) def cudaStreamGetPriority(hStream): """ Query the priority of a stream. Query the priority of a stream. The priority is returned in in `priority`. Note that if the stream was created with a priority outside the meaningful numerical range returned by :py:obj:`~.cudaDeviceGetStreamPriorityRange`, this function returns the clamped priority. See :py:obj:`~.cudaStreamCreateWithPriority` for details about priority clamping. Parameters ---------- hStream : :py:obj:`~.CUstream` or :py:obj:`~.cudaStream_t` Handle to the stream to be queried Returns ------- cudaError_t :py:obj:`~.cudaSuccess`, :py:obj:`~.cudaErrorInvalidValue`, :py:obj:`~.cudaErrorInvalidResourceHandle` priority : int Pointer to a signed integer in which the stream's priority is returned See Also -------- :py:obj:`~.cudaStreamCreateWithPriority`, :py:obj:`~.cudaDeviceGetStreamPriorityRange`, :py:obj:`~.cudaStreamGetFlags`, :py:obj:`~.cudaStreamGetDevice`, :py:obj:`~.cuStreamGetPriority` """ cdef cyruntime.cudaStream_t cyhStream if hStream is None: phStream = 0 elif isinstance(hStream, (cudaStream_t,driver.CUstream)): phStream = int(hStream) else: phStream = int(cudaStream_t(hStream)) cyhStream = phStream cdef int priority = 0 with nogil: err = cyruntime.cudaStreamGetPriority(cyhStream, &priority) if err != cyruntime.cudaSuccess: return (_dict_cudaError_t[err], None) return (_dict_cudaError_t[err], priority) @cython.embedsignature(True) def cudaStreamGetFlags(hStream): """ Query the flags of a stream. Query the flags of a stream. The flags are returned in `flags`. See :py:obj:`~.cudaStreamCreateWithFlags` for a list of valid flags. Parameters ---------- hStream : :py:obj:`~.CUstream` or :py:obj:`~.cudaStream_t` Handle to the stream to be queried Returns ------- cudaError_t :py:obj:`~.cudaSuccess`, :py:obj:`~.cudaErrorInvalidValue`, :py:obj:`~.cudaErrorInvalidResourceHandle` flags : unsigned int Pointer to an unsigned integer in which the stream's flags are returned See Also -------- :py:obj:`~.cudaStreamCreateWithPriority`, :py:obj:`~.cudaStreamCreateWithFlags`, :py:obj:`~.cudaStreamGetPriority`, :py:obj:`~.cudaStreamGetDevice`, :py:obj:`~.cuStreamGetFlags` """ cdef cyruntime.cudaStream_t cyhStream if hStream is None: phStream = 0 elif isinstance(hStream, (cudaStream_t,driver.CUstream)): phStream = int(hStream) else: phStream = int(cudaStream_t(hStream)) cyhStream = phStream cdef unsigned int flags = 0 with nogil: err = cyruntime.cudaStreamGetFlags(cyhStream, &flags) if err != cyruntime.cudaSuccess: return (_dict_cudaError_t[err], None) return (_dict_cudaError_t[err], flags) @cython.embedsignature(True) def cudaStreamGetId(hStream): """ Query the Id of a stream. Query the Id of a stream. The Id is returned in `streamId`. The Id is unique for the life of the program. The stream handle `hStream` can refer to any of the following: - a stream created via any of the CUDA runtime APIs such as :py:obj:`~.cudaStreamCreate`, :py:obj:`~.cudaStreamCreateWithFlags` and :py:obj:`~.cudaStreamCreateWithPriority`, or their driver API equivalents such as :py:obj:`~.cuStreamCreate` or :py:obj:`~.cuStreamCreateWithPriority`. Passing an invalid handle will result in undefined behavior. - any of the special streams such as the NULL stream, :py:obj:`~.cudaStreamLegacy` and :py:obj:`~.cudaStreamPerThread` respectively. The driver API equivalents of these are also accepted which are NULL, :py:obj:`~.CU_STREAM_LEGACY` and :py:obj:`~.CU_STREAM_PER_THREAD`. Parameters ---------- hStream : :py:obj:`~.CUstream` or :py:obj:`~.cudaStream_t` Handle to the stream to be queried Returns ------- cudaError_t :py:obj:`~.cudaSuccess`, :py:obj:`~.cudaErrorInvalidValue`, :py:obj:`~.cudaErrorInvalidResourceHandle` streamId : unsigned long long Pointer to an unsigned long long in which the stream Id is returned See Also -------- :py:obj:`~.cudaStreamCreateWithPriority`, :py:obj:`~.cudaStreamCreateWithFlags`, :py:obj:`~.cudaStreamGetPriority`, :py:obj:`~.cudaStreamGetFlags`, :py:obj:`~.cuStreamGetId` """ cdef cyruntime.cudaStream_t cyhStream if hStream is None: phStream = 0 elif isinstance(hStream, (cudaStream_t,driver.CUstream)): phStream = int(hStream) else: phStream = int(cudaStream_t(hStream)) cyhStream = phStream cdef unsigned long long streamId = 0 with nogil: err = cyruntime.cudaStreamGetId(cyhStream, &streamId) if err != cyruntime.cudaSuccess: return (_dict_cudaError_t[err], None) return (_dict_cudaError_t[err], streamId) @cython.embedsignature(True) def cudaStreamGetDevice(hStream): """ Query the device of a stream. Returns in `*device` the device of the stream. Parameters ---------- hStream : :py:obj:`~.CUstream` or :py:obj:`~.cudaStream_t` Handle to the stream to be queried Returns ------- cudaError_t :py:obj:`~.cudaSuccess`, :py:obj:`~.cudaErrorInvalidValue`, :py:obj:`~.cudaErrorDeviceUnavailable`, device : int Returns the device to which the stream belongs See Also -------- :py:obj:`~.cudaSetDevice`, :py:obj:`~.cudaGetDevice`, :py:obj:`~.cudaStreamCreate`, :py:obj:`~.cudaStreamGetPriority`, :py:obj:`~.cudaStreamGetFlags`, :py:obj:`~.cuStreamGetId` """ cdef cyruntime.cudaStream_t cyhStream if hStream is None: phStream = 0 elif isinstance(hStream, (cudaStream_t,driver.CUstream)): phStream = int(hStream) else: phStream = int(cudaStream_t(hStream)) cyhStream = phStream cdef int device = 0 with nogil: err = cyruntime.cudaStreamGetDevice(cyhStream, &device) if err != cyruntime.cudaSuccess: return (_dict_cudaError_t[err], None) return (_dict_cudaError_t[err], device) @cython.embedsignature(True) def cudaCtxResetPersistingL2Cache(): """ Resets all persisting lines in cache to normal status. Resets all persisting lines in cache to normal status. Takes effect on function return. Returns ------- cudaError_t :py:obj:`~.cudaSuccess`, See Also -------- :py:obj:`~.cudaAccessPolicyWindow` """ with nogil: err = cyruntime.cudaCtxResetPersistingL2Cache() return (_dict_cudaError_t[err],) @cython.embedsignature(True) def cudaStreamCopyAttributes(dst, src): """ Copies attributes from source stream to destination stream. Copies attributes from source stream `src` to destination stream `dst`. Both streams must have the same context. Parameters ---------- dst : :py:obj:`~.CUstream` or :py:obj:`~.cudaStream_t` Destination stream src : :py:obj:`~.CUstream` or :py:obj:`~.cudaStream_t` Source stream For attributes see :py:obj:`~.cudaStreamAttrID` Returns ------- cudaError_t :py:obj:`~.cudaSuccess`, :py:obj:`~.cudaErrorNotSupported` See Also -------- :py:obj:`~.cudaAccessPolicyWindow` """ cdef cyruntime.cudaStream_t cysrc if src is None: psrc = 0 elif isinstance(src, (cudaStream_t,driver.CUstream)): psrc = int(src) else: psrc = int(cudaStream_t(src)) cysrc = psrc cdef cyruntime.cudaStream_t cydst if dst is None: pdst = 0 elif isinstance(dst, (cudaStream_t,driver.CUstream)): pdst = int(dst) else: pdst = int(cudaStream_t(dst)) cydst = pdst with nogil: err = cyruntime.cudaStreamCopyAttributes(cydst, cysrc) return (_dict_cudaError_t[err],) @cython.embedsignature(True) def cudaStreamGetAttribute(hStream, attr not None : cudaStreamAttrID): """ Queries stream attribute. Queries attribute `attr` from `hStream` and stores it in corresponding member of `value_out`. Parameters ---------- hStream : :py:obj:`~.CUstream` or :py:obj:`~.cudaStream_t` attr : :py:obj:`~.cudaStreamAttrID` Returns ------- cudaError_t :py:obj:`~.cudaSuccess`, :py:obj:`~.cudaErrorInvalidValue`, :py:obj:`~.cudaErrorInvalidResourceHandle` value_out : :py:obj:`~.cudaStreamAttrValue` See Also -------- :py:obj:`~.cudaAccessPolicyWindow` """ cdef cyruntime.cudaStream_t cyhStream if hStream is None: phStream = 0 elif isinstance(hStream, (cudaStream_t,driver.CUstream)): phStream = int(hStream) else: phStream = int(cudaStream_t(hStream)) cyhStream = phStream cdef cyruntime.cudaStreamAttrID cyattr = attr.value cdef cudaStreamAttrValue value_out = cudaStreamAttrValue() with nogil: err = cyruntime.cudaStreamGetAttribute(cyhStream, cyattr, value_out._pvt_ptr) if err != cyruntime.cudaSuccess: return (_dict_cudaError_t[err], None) return (_dict_cudaError_t[err], value_out) @cython.embedsignature(True) def cudaStreamSetAttribute(hStream, attr not None : cudaStreamAttrID, value : Optional[cudaStreamAttrValue]): """ Sets stream attribute. Sets attribute `attr` on `hStream` from corresponding attribute of `value`. The updated attribute will be applied to subsequent work submitted to the stream. It will not affect previously submitted work. Parameters ---------- hStream : :py:obj:`~.CUstream` or :py:obj:`~.cudaStream_t` attr : :py:obj:`~.cudaStreamAttrID` value : :py:obj:`~.cudaStreamAttrValue` Returns ------- cudaError_t :py:obj:`~.cudaSuccess`, :py:obj:`~.cudaErrorInvalidValue`, :py:obj:`~.cudaErrorInvalidResourceHandle` See Also -------- :py:obj:`~.cudaAccessPolicyWindow` """ cdef cyruntime.cudaStream_t cyhStream if hStream is None: phStream = 0 elif isinstance(hStream, (cudaStream_t,driver.CUstream)): phStream = int(hStream) else: phStream = int(cudaStream_t(hStream)) cyhStream = phStream cdef cyruntime.cudaStreamAttrID cyattr = attr.value cdef cyruntime.cudaStreamAttrValue* cyvalue_ptr = value._pvt_ptr if value is not None else NULL with nogil: err = cyruntime.cudaStreamSetAttribute(cyhStream, cyattr, cyvalue_ptr) return (_dict_cudaError_t[err],) @cython.embedsignature(True) def cudaStreamDestroy(stream): """ Destroys and cleans up an asynchronous stream. Destroys and cleans up the asynchronous stream specified by `stream`. In case the device is still doing work in the stream `stream` when :py:obj:`~.cudaStreamDestroy()` is called, the function will return immediately and the resources associated with `stream` will be released automatically once the device has completed all work in `stream`. Parameters ---------- stream : :py:obj:`~.CUstream` or :py:obj:`~.cudaStream_t` Stream identifier Returns ------- cudaError_t :py:obj:`~.cudaSuccess`, :py:obj:`~.cudaErrorInvalidValue`, :py:obj:`~.cudaErrorInvalidResourceHandle` See Also -------- :py:obj:`~.cudaStreamCreate`, :py:obj:`~.cudaStreamCreateWithFlags`, :py:obj:`~.cudaStreamQuery`, :py:obj:`~.cudaStreamWaitEvent`, :py:obj:`~.cudaStreamSynchronize`, :py:obj:`~.cudaStreamAddCallback`, :py:obj:`~.cuStreamDestroy` """ cdef cyruntime.cudaStream_t cystream if stream is None: pstream = 0 elif isinstance(stream, (cudaStream_t,driver.CUstream)): pstream = int(stream) else: pstream = int(cudaStream_t(stream)) cystream = pstream with nogil: err = cyruntime.cudaStreamDestroy(cystream) return (_dict_cudaError_t[err],) @cython.embedsignature(True) def cudaStreamWaitEvent(stream, event, unsigned int flags): """ Make a compute stream wait on an event. Makes all future work submitted to `stream` wait for all work captured in `event`. See :py:obj:`~.cudaEventRecord()` for details on what is captured by an event. The synchronization will be performed efficiently on the device when applicable. `event` may be from a different device than `stream`. flags include: - :py:obj:`~.cudaEventWaitDefault`: Default event creation flag. - :py:obj:`~.cudaEventWaitExternal`: Event is captured in the graph as an external event node when performing stream capture. Parameters ---------- stream : :py:obj:`~.CUstream` or :py:obj:`~.cudaStream_t` Stream to wait event : :py:obj:`~.CUevent` or :py:obj:`~.cudaEvent_t` Event to wait on flags : unsigned int Parameters for the operation(See above) Returns ------- cudaError_t :py:obj:`~.cudaSuccess`, :py:obj:`~.cudaErrorInvalidValue`, :py:obj:`~.cudaErrorInvalidResourceHandle` See Also -------- :py:obj:`~.cudaStreamCreate`, :py:obj:`~.cudaStreamCreateWithFlags`, :py:obj:`~.cudaStreamQuery`, :py:obj:`~.cudaStreamSynchronize`, :py:obj:`~.cudaStreamAddCallback`, :py:obj:`~.cudaStreamDestroy`, :py:obj:`~.cuStreamWaitEvent` """ cdef cyruntime.cudaEvent_t cyevent if event is None: pevent = 0 elif isinstance(event, (cudaEvent_t,driver.CUevent)): pevent = int(event) else: pevent = int(cudaEvent_t(event)) cyevent = pevent cdef cyruntime.cudaStream_t cystream if stream is None: pstream = 0 elif isinstance(stream, (cudaStream_t,driver.CUstream)): pstream = int(stream) else: pstream = int(cudaStream_t(stream)) cystream = pstream with nogil: err = cyruntime.cudaStreamWaitEvent(cystream, cyevent, flags) return (_dict_cudaError_t[err],) ctypedef struct cudaStreamCallbackData_st: cyruntime.cudaStreamCallback_t callback void *userData ctypedef cudaStreamCallbackData_st cudaStreamCallbackData @cython.show_performance_hints(False) cdef void cudaStreamRtCallbackWrapper(cyruntime.cudaStream_t stream, cyruntime.cudaError_t status, void *data) nogil: cdef cudaStreamCallbackData *cbData = data with gil: cbData.callback(stream, status, cbData.userData) free(cbData) @cython.embedsignature(True) def cudaStreamAddCallback(stream, callback, userData, unsigned int flags): """ Add a callback to a compute stream. Adds a callback to be called on the host after all currently enqueued items in the stream have completed. For each cudaStreamAddCallback call, a callback will be executed exactly once. The callback will block later work in the stream until it is finished. The callback may be passed :py:obj:`~.cudaSuccess` or an error code. In the event of a device error, all subsequently executed callbacks will receive an appropriate :py:obj:`~.cudaError_t`. Callbacks must not make any CUDA API calls. Attempting to use CUDA APIs may result in :py:obj:`~.cudaErrorNotPermitted`. Callbacks must not perform any synchronization that may depend on outstanding device work or other callbacks that are not mandated to run earlier. Callbacks without a mandated order (in independent streams) execute in undefined order and may be serialized. For the purposes of Unified Memory, callback execution makes a number of guarantees: - The callback stream is considered idle for the duration of the callback. Thus, for example, a callback may always use memory attached to the callback stream. - The start of execution of a callback has the same effect as synchronizing an event recorded in the same stream immediately prior to the callback. It thus synchronizes streams which have been "joined" prior to the callback. - Adding device work to any stream does not have the effect of making the stream active until all preceding callbacks have executed. Thus, for example, a callback might use global attached memory even if work has been added to another stream, if it has been properly ordered with an event. - Completion of a callback does not cause a stream to become active except as described above. The callback stream will remain idle if no device work follows the callback, and will remain idle across consecutive callbacks without device work in between. Thus, for example, stream synchronization can be done by signaling from a callback at the end of the stream. Parameters ---------- stream : :py:obj:`~.CUstream` or :py:obj:`~.cudaStream_t` Stream to add callback to callback : :py:obj:`~.cudaStreamCallback_t` The function to call once preceding stream operations are complete userData : Any User specified data to be passed to the callback function flags : unsigned int Reserved for future use, must be 0 Returns ------- cudaError_t :py:obj:`~.cudaSuccess`, :py:obj:`~.cudaErrorInvalidResourceHandle`, :py:obj:`~.cudaErrorInvalidValue`, :py:obj:`~.cudaErrorNotSupported` See Also -------- :py:obj:`~.cudaStreamCreate`, :py:obj:`~.cudaStreamCreateWithFlags`, :py:obj:`~.cudaStreamQuery`, :py:obj:`~.cudaStreamSynchronize`, :py:obj:`~.cudaStreamWaitEvent`, :py:obj:`~.cudaStreamDestroy`, :py:obj:`~.cudaMallocManaged`, :py:obj:`~.cudaStreamAttachMemAsync`, :py:obj:`~.cudaLaunchHostFunc`, :py:obj:`~.cuStreamAddCallback` Notes ----- This function is slated for eventual deprecation and removal. If you do not require the callback to execute in case of a device error, consider using :py:obj:`~.cudaLaunchHostFunc`. Additionally, this function is not supported with :py:obj:`~.cudaStreamBeginCapture` and :py:obj:`~.cudaStreamEndCapture`, unlike :py:obj:`~.cudaLaunchHostFunc`. """ cdef cyruntime.cudaStreamCallback_t cycallback if callback is None: pcallback = 0 elif isinstance(callback, (cudaStreamCallback_t,)): pcallback = int(callback) else: pcallback = int(cudaStreamCallback_t(callback)) cycallback = pcallback cdef cyruntime.cudaStream_t cystream if stream is None: pstream = 0 elif isinstance(stream, (cudaStream_t,driver.CUstream)): pstream = int(stream) else: pstream = int(cudaStream_t(stream)) cystream = pstream cyuserData = _HelperInputVoidPtr(userData) cdef void* cyuserData_ptr = cyuserData.cptr cdef cudaStreamCallbackData *cbData = NULL cbData = malloc(sizeof(cbData[0])) if cbData == NULL: return (cudaError_t.cudaErrorMemoryAllocation,) cbData.callback = cycallback cbData.userData = cyuserData_ptr with nogil: err = cyruntime.cudaStreamAddCallback(cystream, cudaStreamRtCallbackWrapper, cbData, flags) if err != cyruntime.cudaSuccess: free(cbData) return (_dict_cudaError_t[err],) @cython.embedsignature(True) def cudaStreamSynchronize(stream): """ Waits for stream tasks to complete. Blocks until `stream` has completed all operations. If the :py:obj:`~.cudaDeviceScheduleBlockingSync` flag was set for this device, the host thread will block until the stream is finished with all of its tasks. Parameters ---------- stream : :py:obj:`~.CUstream` or :py:obj:`~.cudaStream_t` Stream identifier Returns ------- cudaError_t :py:obj:`~.cudaSuccess`, :py:obj:`~.cudaErrorInvalidResourceHandle` See Also -------- :py:obj:`~.cudaStreamCreate`, :py:obj:`~.cudaStreamCreateWithFlags`, :py:obj:`~.cudaStreamQuery`, :py:obj:`~.cudaStreamWaitEvent`, :py:obj:`~.cudaStreamAddCallback`, :py:obj:`~.cudaStreamDestroy`, :py:obj:`~.cuStreamSynchronize` """ cdef cyruntime.cudaStream_t cystream if stream is None: pstream = 0 elif isinstance(stream, (cudaStream_t,driver.CUstream)): pstream = int(stream) else: pstream = int(cudaStream_t(stream)) cystream = pstream with nogil: err = cyruntime.cudaStreamSynchronize(cystream) return (_dict_cudaError_t[err],) @cython.embedsignature(True) def cudaStreamQuery(stream): """ Queries an asynchronous stream for completion status. Returns :py:obj:`~.cudaSuccess` if all operations in `stream` have completed, or :py:obj:`~.cudaErrorNotReady` if not. For the purposes of Unified Memory, a return value of :py:obj:`~.cudaSuccess` is equivalent to having called :py:obj:`~.cudaStreamSynchronize()`. Parameters ---------- stream : :py:obj:`~.CUstream` or :py:obj:`~.cudaStream_t` Stream identifier Returns ------- cudaError_t :py:obj:`~.cudaSuccess`, :py:obj:`~.cudaErrorNotReady`, :py:obj:`~.cudaErrorInvalidResourceHandle` See Also -------- :py:obj:`~.cudaStreamCreate`, :py:obj:`~.cudaStreamCreateWithFlags`, :py:obj:`~.cudaStreamWaitEvent`, :py:obj:`~.cudaStreamSynchronize`, :py:obj:`~.cudaStreamAddCallback`, :py:obj:`~.cudaStreamDestroy`, :py:obj:`~.cuStreamQuery` """ cdef cyruntime.cudaStream_t cystream if stream is None: pstream = 0 elif isinstance(stream, (cudaStream_t,driver.CUstream)): pstream = int(stream) else: pstream = int(cudaStream_t(stream)) cystream = pstream with nogil: err = cyruntime.cudaStreamQuery(cystream) return (_dict_cudaError_t[err],) @cython.embedsignature(True) def cudaStreamAttachMemAsync(stream, devPtr, size_t length, unsigned int flags): """ Attach memory to a stream asynchronously. Enqueues an operation in `stream` to specify stream association of `length` bytes of memory starting from `devPtr`. This function is a stream-ordered operation, meaning that it is dependent on, and will only take effect when, previous work in stream has completed. Any previous association is automatically replaced. `devPtr` must point to an one of the following types of memories: - managed memory declared using the managed keyword or allocated with :py:obj:`~.cudaMallocManaged`. - a valid host-accessible region of system-allocated pageable memory. This type of memory may only be specified if the device associated with the stream reports a non-zero value for the device attribute :py:obj:`~.cudaDevAttrPageableMemoryAccess`. For managed allocations, `length` must be either zero or the entire allocation's size. Both indicate that the entire allocation's stream association is being changed. Currently, it is not possible to change stream association for a portion of a managed allocation. For pageable allocations, `length` must be non-zero. The stream association is specified using `flags` which must be one of :py:obj:`~.cudaMemAttachGlobal`, :py:obj:`~.cudaMemAttachHost` or :py:obj:`~.cudaMemAttachSingle`. The default value for `flags` is :py:obj:`~.cudaMemAttachSingle` If the :py:obj:`~.cudaMemAttachGlobal` flag is specified, the memory can be accessed by any stream on any device. If the :py:obj:`~.cudaMemAttachHost` flag is specified, the program makes a guarantee that it won't access the memory on the device from any stream on a device that has a zero value for the device attribute :py:obj:`~.cudaDevAttrConcurrentManagedAccess`. If the :py:obj:`~.cudaMemAttachSingle` flag is specified and `stream` is associated with a device that has a zero value for the device attribute :py:obj:`~.cudaDevAttrConcurrentManagedAccess`, the program makes a guarantee that it will only access the memory on the device from `stream`. It is illegal to attach singly to the NULL stream, because the NULL stream is a virtual global stream and not a specific stream. An error will be returned in this case. When memory is associated with a single stream, the Unified Memory system will allow CPU access to this memory region so long as all operations in `stream` have completed, regardless of whether other streams are active. In effect, this constrains exclusive ownership of the managed memory region by an active GPU to per-stream activity instead of whole-GPU activity. Accessing memory on the device from streams that are not associated with it will produce undefined results. No error checking is performed by the Unified Memory system to ensure that kernels launched into other streams do not access this region. It is a program's responsibility to order calls to :py:obj:`~.cudaStreamAttachMemAsync` via events, synchronization or other means to ensure legal access to memory at all times. Data visibility and coherency will be changed appropriately for all kernels which follow a stream-association change. If `stream` is destroyed while data is associated with it, the association is removed and the association reverts to the default visibility of the allocation as specified at :py:obj:`~.cudaMallocManaged`. For managed variables, the default association is always :py:obj:`~.cudaMemAttachGlobal`. Note that destroying a stream is an asynchronous operation, and as a result, the change to default association won't happen until all work in the stream has completed. Parameters ---------- stream : :py:obj:`~.CUstream` or :py:obj:`~.cudaStream_t` Stream in which to enqueue the attach operation devPtr : Any Pointer to memory (must be a pointer to managed memory or to a valid host-accessible region of system-allocated memory) length : size_t Length of memory (defaults to zero) flags : unsigned int Must be one of :py:obj:`~.cudaMemAttachGlobal`, :py:obj:`~.cudaMemAttachHost` or :py:obj:`~.cudaMemAttachSingle` (defaults to :py:obj:`~.cudaMemAttachSingle`) Returns ------- cudaError_t :py:obj:`~.cudaSuccess`, :py:obj:`~.cudaErrorNotReady`, :py:obj:`~.cudaErrorInvalidValue`, :py:obj:`~.cudaErrorInvalidResourceHandle` See Also -------- :py:obj:`~.cudaStreamCreate`, :py:obj:`~.cudaStreamCreateWithFlags`, :py:obj:`~.cudaStreamWaitEvent`, :py:obj:`~.cudaStreamSynchronize`, :py:obj:`~.cudaStreamAddCallback`, :py:obj:`~.cudaStreamDestroy`, :py:obj:`~.cudaMallocManaged`, :py:obj:`~.cuStreamAttachMemAsync` """ cdef cyruntime.cudaStream_t cystream if stream is None: pstream = 0 elif isinstance(stream, (cudaStream_t,driver.CUstream)): pstream = int(stream) else: pstream = int(cudaStream_t(stream)) cystream = pstream cydevPtr = _HelperInputVoidPtr(devPtr) cdef void* cydevPtr_ptr = cydevPtr.cptr with nogil: err = cyruntime.cudaStreamAttachMemAsync(cystream, cydevPtr_ptr, length, flags) return (_dict_cudaError_t[err],) @cython.embedsignature(True) def cudaStreamBeginCapture(stream, mode not None : cudaStreamCaptureMode): """ Begins graph capture on a stream. Begin graph capture on `stream`. When a stream is in capture mode, all operations pushed into the stream will not be executed, but will instead be captured into a graph, which will be returned via :py:obj:`~.cudaStreamEndCapture`. Capture may not be initiated if `stream` is :py:obj:`~.cudaStreamLegacy`. Capture must be ended on the same stream in which it was initiated, and it may only be initiated if the stream is not already in capture mode. The capture mode may be queried via :py:obj:`~.cudaStreamIsCapturing`. A unique id representing the capture sequence may be queried via :py:obj:`~.cudaStreamGetCaptureInfo`. If `mode` is not :py:obj:`~.cudaStreamCaptureModeRelaxed`, :py:obj:`~.cudaStreamEndCapture` must be called on this stream from the same thread. Parameters ---------- stream : :py:obj:`~.CUstream` or :py:obj:`~.cudaStream_t` Stream in which to initiate capture mode : :py:obj:`~.cudaStreamCaptureMode` Controls the interaction of this capture sequence with other API calls that are potentially unsafe. For more details see :py:obj:`~.cudaThreadExchangeStreamCaptureMode`. Returns ------- cudaError_t :py:obj:`~.cudaSuccess`, :py:obj:`~.cudaErrorInvalidValue` See Also -------- :py:obj:`~.cudaStreamCreate`, :py:obj:`~.cudaStreamIsCapturing`, :py:obj:`~.cudaStreamEndCapture`, :py:obj:`~.cudaThreadExchangeStreamCaptureMode` Notes ----- Kernels captured using this API must not use texture and surface references. Reading or writing through any texture or surface reference is undefined behavior. This restriction does not apply to texture and surface objects. """ cdef cyruntime.cudaStream_t cystream if stream is None: pstream = 0 elif isinstance(stream, (cudaStream_t,driver.CUstream)): pstream = int(stream) else: pstream = int(cudaStream_t(stream)) cystream = pstream cdef cyruntime.cudaStreamCaptureMode cymode = mode.value with nogil: err = cyruntime.cudaStreamBeginCapture(cystream, cymode) return (_dict_cudaError_t[err],) @cython.embedsignature(True) def cudaStreamBeginCaptureToGraph(stream, graph, dependencies : Optional[tuple[cudaGraphNode_t] | list[cudaGraphNode_t]], dependencyData : Optional[tuple[cudaGraphEdgeData] | list[cudaGraphEdgeData]], size_t numDependencies, mode not None : cudaStreamCaptureMode): """ Begins graph capture on a stream to an existing graph. Begin graph capture on `stream`. When a stream is in capture mode, all operations pushed into the stream will not be executed, but will instead be captured into `graph`, which will be returned via :py:obj:`~.cudaStreamEndCapture`. Capture may not be initiated if `stream` is :py:obj:`~.cudaStreamLegacy`. Capture must be ended on the same stream in which it was initiated, and it may only be initiated if the stream is not already in capture mode. The capture mode may be queried via :py:obj:`~.cudaStreamIsCapturing`. A unique id representing the capture sequence may be queried via :py:obj:`~.cudaStreamGetCaptureInfo`. If `mode` is not :py:obj:`~.cudaStreamCaptureModeRelaxed`, :py:obj:`~.cudaStreamEndCapture` must be called on this stream from the same thread. Parameters ---------- stream : :py:obj:`~.CUstream` or :py:obj:`~.cudaStream_t` Stream in which to initiate capture. graph : :py:obj:`~.CUgraph` or :py:obj:`~.cudaGraph_t` Graph to capture into. dependencies : list[:py:obj:`~.cudaGraphNode_t`] Dependencies of the first node captured in the stream. Can be NULL if numDependencies is 0. dependencyData : list[:py:obj:`~.cudaGraphEdgeData`] Optional array of data associated with each dependency. numDependencies : size_t Number of dependencies. mode : :py:obj:`~.cudaStreamCaptureMode` Controls the interaction of this capture sequence with other API calls that are potentially unsafe. For more details see :py:obj:`~.cudaThreadExchangeStreamCaptureMode`. Returns ------- cudaError_t :py:obj:`~.cudaSuccess`, :py:obj:`~.cudaErrorInvalidValue` See Also -------- :py:obj:`~.cudaStreamCreate`, :py:obj:`~.cudaStreamIsCapturing`, :py:obj:`~.cudaStreamEndCapture`, :py:obj:`~.cudaThreadExchangeStreamCaptureMode` Notes ----- Kernels captured using this API must not use texture and surface references. Reading or writing through any texture or surface reference is undefined behavior. This restriction does not apply to texture and surface objects. """ dependencyData = [] if dependencyData is None else dependencyData if not all(isinstance(_x, (cudaGraphEdgeData,)) for _x in dependencyData): raise TypeError("Argument 'dependencyData' is not instance of type (expected tuple[cyruntime.cudaGraphEdgeData,] or list[cyruntime.cudaGraphEdgeData,]") dependencies = [] if dependencies is None else dependencies if not all(isinstance(_x, (cudaGraphNode_t,driver.CUgraphNode)) for _x in dependencies): raise TypeError("Argument 'dependencies' is not instance of type (expected tuple[cyruntime.cudaGraphNode_t,driver.CUgraphNode] or list[cyruntime.cudaGraphNode_t,driver.CUgraphNode]") cdef cyruntime.cudaGraph_t cygraph if graph is None: pgraph = 0 elif isinstance(graph, (cudaGraph_t,driver.CUgraph)): pgraph = int(graph) else: pgraph = int(cudaGraph_t(graph)) cygraph = pgraph cdef cyruntime.cudaStream_t cystream if stream is None: pstream = 0 elif isinstance(stream, (cudaStream_t,driver.CUstream)): pstream = int(stream) else: pstream = int(cudaStream_t(stream)) cystream = pstream cdef cyruntime.cudaGraphNode_t* cydependencies = NULL if len(dependencies) > 1: cydependencies = calloc(len(dependencies), sizeof(cyruntime.cudaGraphNode_t)) if cydependencies is NULL: raise MemoryError('Failed to allocate length x size memory: ' + str(len(dependencies)) + 'x' + str(sizeof(cyruntime.cudaGraphNode_t))) else: for idx in range(len(dependencies)): cydependencies[idx] = (dependencies[idx])._pvt_ptr[0] elif len(dependencies) == 1: cydependencies = (dependencies[0])._pvt_ptr cdef cyruntime.cudaGraphEdgeData* cydependencyData = NULL if len(dependencyData) > 1: cydependencyData = calloc(len(dependencyData), sizeof(cyruntime.cudaGraphEdgeData)) if cydependencyData is NULL: raise MemoryError('Failed to allocate length x size memory: ' + str(len(dependencyData)) + 'x' + str(sizeof(cyruntime.cudaGraphEdgeData))) for idx in range(len(dependencyData)): string.memcpy(&cydependencyData[idx], (dependencyData[idx])._pvt_ptr, sizeof(cyruntime.cudaGraphEdgeData)) elif len(dependencyData) == 1: cydependencyData = (dependencyData[0])._pvt_ptr if numDependencies > len(dependencies): raise RuntimeError("List is too small: " + str(len(dependencies)) + " < " + str(numDependencies)) cdef cyruntime.cudaStreamCaptureMode cymode = mode.value with nogil: err = cyruntime.cudaStreamBeginCaptureToGraph(cystream, cygraph, cydependencies, cydependencyData, numDependencies, cymode) if len(dependencies) > 1 and cydependencies is not NULL: free(cydependencies) if len(dependencyData) > 1 and cydependencyData is not NULL: free(cydependencyData) return (_dict_cudaError_t[err],) @cython.embedsignature(True) def cudaThreadExchangeStreamCaptureMode(mode not None : cudaStreamCaptureMode): """ Swaps the stream capture interaction mode for a thread. Sets the calling thread's stream capture interaction mode to the value contained in `*mode`, and overwrites `*mode` with the previous mode for the thread. To facilitate deterministic behavior across function or module boundaries, callers are encouraged to use this API in a push-pop fashion: **View CUDA Toolkit Documentation for a C++ code example** During stream capture (see :py:obj:`~.cudaStreamBeginCapture`), some actions, such as a call to :py:obj:`~.cudaMalloc`, may be unsafe. In the case of :py:obj:`~.cudaMalloc`, the operation is not enqueued asynchronously to a stream, and is not observed by stream capture. Therefore, if the sequence of operations captured via :py:obj:`~.cudaStreamBeginCapture` depended on the allocation being replayed whenever the graph is launched, the captured graph would be invalid. Therefore, stream capture places restrictions on API calls that can be made within or concurrently to a :py:obj:`~.cudaStreamBeginCapture`-:py:obj:`~.cudaStreamEndCapture` sequence. This behavior can be controlled via this API and flags to :py:obj:`~.cudaStreamBeginCapture`. A thread's mode is one of the following: - `cudaStreamCaptureModeGlobal:` This is the default mode. If the local thread has an ongoing capture sequence that was not initiated with `cudaStreamCaptureModeRelaxed` at `cuStreamBeginCapture`, or if any other thread has a concurrent capture sequence initiated with `cudaStreamCaptureModeGlobal`, this thread is prohibited from potentially unsafe API calls. - `cudaStreamCaptureModeThreadLocal:` If the local thread has an ongoing capture sequence not initiated with `cudaStreamCaptureModeRelaxed`, it is prohibited from potentially unsafe API calls. Concurrent capture sequences in other threads are ignored. - `cudaStreamCaptureModeRelaxed:` The local thread is not prohibited from potentially unsafe API calls. Note that the thread is still prohibited from API calls which necessarily conflict with stream capture, for example, attempting :py:obj:`~.cudaEventQuery` on an event that was last recorded inside a capture sequence. Parameters ---------- mode : :py:obj:`~.cudaStreamCaptureMode` Pointer to mode value to swap with the current mode Returns ------- cudaError_t :py:obj:`~.cudaSuccess`, :py:obj:`~.cudaErrorInvalidValue` mode : :py:obj:`~.cudaStreamCaptureMode` Pointer to mode value to swap with the current mode See Also -------- :py:obj:`~.cudaStreamBeginCapture` """ cdef cyruntime.cudaStreamCaptureMode cymode = mode.value with nogil: err = cyruntime.cudaThreadExchangeStreamCaptureMode(&cymode) if err != cyruntime.cudaSuccess: return (_dict_cudaError_t[err], None) return (_dict_cudaError_t[err], cudaStreamCaptureMode(cymode)) @cython.embedsignature(True) def cudaStreamEndCapture(stream): """ Ends capture on a stream, returning the captured graph. End capture on `stream`, returning the captured graph via `pGraph`. Capture must have been initiated on `stream` via a call to :py:obj:`~.cudaStreamBeginCapture`. If capture was invalidated, due to a violation of the rules of stream capture, then a NULL graph will be returned. If the `mode` argument to :py:obj:`~.cudaStreamBeginCapture` was not :py:obj:`~.cudaStreamCaptureModeRelaxed`, this call must be from the same thread as :py:obj:`~.cudaStreamBeginCapture`. Parameters ---------- stream : :py:obj:`~.CUstream` or :py:obj:`~.cudaStream_t` Stream to query Returns ------- cudaError_t :py:obj:`~.cudaSuccess`, :py:obj:`~.cudaErrorInvalidValue`, :py:obj:`~.cudaErrorStreamCaptureWrongThread` pGraph : :py:obj:`~.cudaGraph_t` The captured graph See Also -------- :py:obj:`~.cudaStreamCreate`, :py:obj:`~.cudaStreamBeginCapture`, :py:obj:`~.cudaStreamIsCapturing`, :py:obj:`~.cudaGraphDestroy` """ cdef cyruntime.cudaStream_t cystream if stream is None: pstream = 0 elif isinstance(stream, (cudaStream_t,driver.CUstream)): pstream = int(stream) else: pstream = int(cudaStream_t(stream)) cystream = pstream cdef cudaGraph_t pGraph = cudaGraph_t() with nogil: err = cyruntime.cudaStreamEndCapture(cystream, pGraph._pvt_ptr) if err != cyruntime.cudaSuccess: return (_dict_cudaError_t[err], None) return (_dict_cudaError_t[err], pGraph) @cython.embedsignature(True) def cudaStreamIsCapturing(stream): """ Returns a stream's capture status. Return the capture status of `stream` via `pCaptureStatus`. After a successful call, `*pCaptureStatus` will contain one of the following: - :py:obj:`~.cudaStreamCaptureStatusNone`: The stream is not capturing. - :py:obj:`~.cudaStreamCaptureStatusActive`: The stream is capturing. - :py:obj:`~.cudaStreamCaptureStatusInvalidated`: The stream was capturing but an error has invalidated the capture sequence. The capture sequence must be terminated with :py:obj:`~.cudaStreamEndCapture` on the stream where it was initiated in order to continue using `stream`. Note that, if this is called on :py:obj:`~.cudaStreamLegacy` (the "null stream") while a blocking stream on the same device is capturing, it will return :py:obj:`~.cudaErrorStreamCaptureImplicit` and `*pCaptureStatus` is unspecified after the call. The blocking stream capture is not invalidated. When a blocking stream is capturing, the legacy stream is in an unusable state until the blocking stream capture is terminated. The legacy stream is not supported for stream capture, but attempted use would have an implicit dependency on the capturing stream(s). Parameters ---------- stream : :py:obj:`~.CUstream` or :py:obj:`~.cudaStream_t` Stream to query Returns ------- cudaError_t :py:obj:`~.cudaSuccess`, :py:obj:`~.cudaErrorInvalidValue`, :py:obj:`~.cudaErrorStreamCaptureImplicit` pCaptureStatus : :py:obj:`~.cudaStreamCaptureStatus` Returns the stream's capture status See Also -------- :py:obj:`~.cudaStreamCreate`, :py:obj:`~.cudaStreamBeginCapture`, :py:obj:`~.cudaStreamEndCapture` """ cdef cyruntime.cudaStream_t cystream if stream is None: pstream = 0 elif isinstance(stream, (cudaStream_t,driver.CUstream)): pstream = int(stream) else: pstream = int(cudaStream_t(stream)) cystream = pstream cdef cyruntime.cudaStreamCaptureStatus pCaptureStatus with nogil: err = cyruntime.cudaStreamIsCapturing(cystream, &pCaptureStatus) if err != cyruntime.cudaSuccess: return (_dict_cudaError_t[err], None) return (_dict_cudaError_t[err], cudaStreamCaptureStatus(pCaptureStatus)) @cython.embedsignature(True) def cudaStreamGetCaptureInfo(stream): """ Query a stream's capture state. Query stream state related to stream capture. If called on :py:obj:`~.cudaStreamLegacy` (the "null stream") while a stream not created with :py:obj:`~.cudaStreamNonBlocking` is capturing, returns :py:obj:`~.cudaErrorStreamCaptureImplicit`. Valid data (other than capture status) is returned only if both of the following are true: - the call returns cudaSuccess - the returned capture status is :py:obj:`~.cudaStreamCaptureStatusActive` Parameters ---------- stream : :py:obj:`~.CUstream` or :py:obj:`~.cudaStream_t` The stream to query Returns ------- cudaError_t :py:obj:`~.cudaSuccess`, :py:obj:`~.cudaErrorInvalidValue`, :py:obj:`~.cudaErrorStreamCaptureImplicit` captureStatus_out : :py:obj:`~.cudaStreamCaptureStatus` Location to return the capture status of the stream; required id_out : unsigned long long Optional location to return an id for the capture sequence, which is unique over the lifetime of the process graph_out : :py:obj:`~.cudaGraph_t` Optional location to return the graph being captured into. All operations other than destroy and node removal are permitted on the graph while the capture sequence is in progress. This API does not transfer ownership of the graph, which is transferred or destroyed at :py:obj:`~.cudaStreamEndCapture`. Note that the graph handle may be invalidated before end of capture for certain errors. Nodes that are or become unreachable from the original stream at :py:obj:`~.cudaStreamEndCapture` due to direct actions on the graph do not trigger :py:obj:`~.cudaErrorStreamCaptureUnjoined`. dependencies_out : list[:py:obj:`~.cudaGraphNode_t`] Optional location to store a pointer to an array of nodes. The next node to be captured in the stream will depend on this set of nodes, absent operations such as event wait which modify this set. The array pointer is valid until the next API call which operates on the stream or until the capture is terminated. The node handles may be copied out and are valid until they or the graph is destroyed. The driver-owned array may also be passed directly to APIs that operate on the graph (not the stream) without copying. numDependencies_out : int Optional location to store the size of the array returned in dependencies_out. See Also -------- :py:obj:`~.cudaStreamGetCaptureInfo_v3`, :py:obj:`~.cudaStreamBeginCapture`, :py:obj:`~.cudaStreamIsCapturing`, :py:obj:`~.cudaStreamUpdateCaptureDependencies` """ cdef cyruntime.cudaStream_t cystream if stream is None: pstream = 0 elif isinstance(stream, (cudaStream_t,driver.CUstream)): pstream = int(stream) else: pstream = int(cudaStream_t(stream)) cystream = pstream cdef cyruntime.cudaStreamCaptureStatus captureStatus_out cdef unsigned long long id_out = 0 cdef cudaGraph_t graph_out = cudaGraph_t() cdef const cyruntime.cudaGraphNode_t* cydependencies_out = NULL pydependencies_out = [] cdef size_t numDependencies_out = 0 with nogil: err = cyruntime.cudaStreamGetCaptureInfo(cystream, &captureStatus_out, &id_out, graph_out._pvt_ptr, &cydependencies_out, &numDependencies_out) if cudaError_t(err) == cudaError_t(0): pydependencies_out = [cudaGraphNode_t(init_value=cydependencies_out[idx]) for idx in range(numDependencies_out)] if err != cyruntime.cudaSuccess: return (_dict_cudaError_t[err], None, None, None, None, None) return (_dict_cudaError_t[err], cudaStreamCaptureStatus(captureStatus_out), id_out, graph_out, pydependencies_out, numDependencies_out) @cython.embedsignature(True) def cudaStreamGetCaptureInfo_v3(stream): """ Query a stream's capture state (12.3+) Query stream state related to stream capture. If called on :py:obj:`~.cudaStreamLegacy` (the "null stream") while a stream not created with :py:obj:`~.cudaStreamNonBlocking` is capturing, returns :py:obj:`~.cudaErrorStreamCaptureImplicit`. Valid data (other than capture status) is returned only if both of the following are true: - the call returns cudaSuccess - the returned capture status is :py:obj:`~.cudaStreamCaptureStatusActive` If `edgeData_out` is non-NULL then `dependencies_out` must be as well. If `dependencies_out` is non-NULL and `edgeData_out` is NULL, but there is non-zero edge data for one or more of the current stream dependencies, the call will return :py:obj:`~.cudaErrorLossyQuery`. Parameters ---------- stream : :py:obj:`~.CUstream` or :py:obj:`~.cudaStream_t` The stream to query Returns ------- cudaError_t :py:obj:`~.cudaSuccess`, :py:obj:`~.cudaErrorInvalidValue`, :py:obj:`~.cudaErrorStreamCaptureImplicit`, :py:obj:`~.cudaErrorLossyQuery` captureStatus_out : :py:obj:`~.cudaStreamCaptureStatus` Location to return the capture status of the stream; required id_out : unsigned long long Optional location to return an id for the capture sequence, which is unique over the lifetime of the process graph_out : :py:obj:`~.cudaGraph_t` Optional location to return the graph being captured into. All operations other than destroy and node removal are permitted on the graph while the capture sequence is in progress. This API does not transfer ownership of the graph, which is transferred or destroyed at :py:obj:`~.cudaStreamEndCapture`. Note that the graph handle may be invalidated before end of capture for certain errors. Nodes that are or become unreachable from the original stream at :py:obj:`~.cudaStreamEndCapture` due to direct actions on the graph do not trigger :py:obj:`~.cudaErrorStreamCaptureUnjoined`. dependencies_out : list[:py:obj:`~.cudaGraphNode_t`] Optional location to store a pointer to an array of nodes. The next node to be captured in the stream will depend on this set of nodes, absent operations such as event wait which modify this set. The array pointer is valid until the next API call which operates on the stream or until the capture is terminated. The node handles may be copied out and are valid until they or the graph is destroyed. The driver-owned array may also be passed directly to APIs that operate on the graph (not the stream) without copying. edgeData_out : list[:py:obj:`~.cudaGraphEdgeData`] Optional location to store a pointer to an array of graph edge data. This array parallels `dependencies_out`; the next node to be added has an edge to `dependencies_out`[i] with annotation `edgeData_out`[i] for each `i`. The array pointer is valid until the next API call which operates on the stream or until the capture is terminated. numDependencies_out : int Optional location to store the size of the array returned in dependencies_out. See Also -------- :py:obj:`~.cudaStreamBeginCapture`, :py:obj:`~.cudaStreamIsCapturing`, :py:obj:`~.cudaStreamUpdateCaptureDependencies` """ cdef cyruntime.cudaStream_t cystream if stream is None: pstream = 0 elif isinstance(stream, (cudaStream_t,driver.CUstream)): pstream = int(stream) else: pstream = int(cudaStream_t(stream)) cystream = pstream cdef cyruntime.cudaStreamCaptureStatus captureStatus_out cdef unsigned long long id_out = 0 cdef cudaGraph_t graph_out = cudaGraph_t() cdef const cyruntime.cudaGraphNode_t* cydependencies_out = NULL pydependencies_out = [] cdef const cyruntime.cudaGraphEdgeData* cyedgeData_out = NULL pyedgeData_out = [] cdef size_t numDependencies_out = 0 with nogil: err = cyruntime.cudaStreamGetCaptureInfo_v3(cystream, &captureStatus_out, &id_out, graph_out._pvt_ptr, &cydependencies_out, &cyedgeData_out, &numDependencies_out) if cudaError_t(err) == cudaError_t(0): pydependencies_out = [cudaGraphNode_t(init_value=cydependencies_out[idx]) for idx in range(numDependencies_out)] if cudaError_t(err) == cudaError_t(0): pyedgeData_out = [cudaGraphEdgeData(_ptr=&cyedgeData_out[idx]) for idx in range(numDependencies_out)] if err != cyruntime.cudaSuccess: return (_dict_cudaError_t[err], None, None, None, None, None, None) return (_dict_cudaError_t[err], cudaStreamCaptureStatus(captureStatus_out), id_out, graph_out, pydependencies_out, pyedgeData_out, numDependencies_out) @cython.embedsignature(True) def cudaStreamUpdateCaptureDependencies(stream, dependencies : Optional[tuple[cudaGraphNode_t] | list[cudaGraphNode_t]], size_t numDependencies, unsigned int flags): """ Update the set of dependencies in a capturing stream (11.3+) Modifies the dependency set of a capturing stream. The dependency set is the set of nodes that the next captured node in the stream will depend on. Valid flags are :py:obj:`~.cudaStreamAddCaptureDependencies` and :py:obj:`~.cudaStreamSetCaptureDependencies`. These control whether the set passed to the API is added to the existing set or replaces it. A flags value of 0 defaults to :py:obj:`~.cudaStreamAddCaptureDependencies`. Nodes that are removed from the dependency set via this API do not result in :py:obj:`~.cudaErrorStreamCaptureUnjoined` if they are unreachable from the stream at :py:obj:`~.cudaStreamEndCapture`. Returns :py:obj:`~.cudaErrorIllegalState` if the stream is not capturing. This API is new in CUDA 11.3. Developers requiring compatibility across minor versions of the CUDA driver to 11.0 should not use this API or provide a fallback. Parameters ---------- stream : :py:obj:`~.CUstream` or :py:obj:`~.cudaStream_t` The stream to update dependencies : list[:py:obj:`~.cudaGraphNode_t`] The set of dependencies to add numDependencies : size_t The size of the dependencies array flags : unsigned int See above Returns ------- cudaError_t :py:obj:`~.cudaSuccess`, :py:obj:`~.cudaErrorInvalidValue`, :py:obj:`~.cudaErrorIllegalState` See Also -------- :py:obj:`~.cudaStreamBeginCapture`, :py:obj:`~.cudaStreamGetCaptureInfo`, """ dependencies = [] if dependencies is None else dependencies if not all(isinstance(_x, (cudaGraphNode_t,driver.CUgraphNode)) for _x in dependencies): raise TypeError("Argument 'dependencies' is not instance of type (expected tuple[cyruntime.cudaGraphNode_t,driver.CUgraphNode] or list[cyruntime.cudaGraphNode_t,driver.CUgraphNode]") cdef cyruntime.cudaStream_t cystream if stream is None: pstream = 0 elif isinstance(stream, (cudaStream_t,driver.CUstream)): pstream = int(stream) else: pstream = int(cudaStream_t(stream)) cystream = pstream cdef cyruntime.cudaGraphNode_t* cydependencies = NULL if len(dependencies) > 1: cydependencies = calloc(len(dependencies), sizeof(cyruntime.cudaGraphNode_t)) if cydependencies is NULL: raise MemoryError('Failed to allocate length x size memory: ' + str(len(dependencies)) + 'x' + str(sizeof(cyruntime.cudaGraphNode_t))) else: for idx in range(len(dependencies)): cydependencies[idx] = (dependencies[idx])._pvt_ptr[0] elif len(dependencies) == 1: cydependencies = (dependencies[0])._pvt_ptr if numDependencies > len(dependencies): raise RuntimeError("List is too small: " + str(len(dependencies)) + " < " + str(numDependencies)) with nogil: err = cyruntime.cudaStreamUpdateCaptureDependencies(cystream, cydependencies, numDependencies, flags) if len(dependencies) > 1 and cydependencies is not NULL: free(cydependencies) return (_dict_cudaError_t[err],) @cython.embedsignature(True) def cudaStreamUpdateCaptureDependencies_v2(stream, dependencies : Optional[tuple[cudaGraphNode_t] | list[cudaGraphNode_t]], dependencyData : Optional[tuple[cudaGraphEdgeData] | list[cudaGraphEdgeData]], size_t numDependencies, unsigned int flags): """ Update the set of dependencies in a capturing stream (12.3+) Modifies the dependency set of a capturing stream. The dependency set is the set of nodes that the next captured node in the stream will depend on. Valid flags are :py:obj:`~.cudaStreamAddCaptureDependencies` and :py:obj:`~.cudaStreamSetCaptureDependencies`. These control whether the set passed to the API is added to the existing set or replaces it. A flags value of 0 defaults to :py:obj:`~.cudaStreamAddCaptureDependencies`. Nodes that are removed from the dependency set via this API do not result in :py:obj:`~.cudaErrorStreamCaptureUnjoined` if they are unreachable from the stream at :py:obj:`~.cudaStreamEndCapture`. Returns :py:obj:`~.cudaErrorIllegalState` if the stream is not capturing. Parameters ---------- stream : :py:obj:`~.CUstream` or :py:obj:`~.cudaStream_t` The stream to update dependencies : list[:py:obj:`~.cudaGraphNode_t`] The set of dependencies to add dependencyData : list[:py:obj:`~.cudaGraphEdgeData`] Optional array of data associated with each dependency. numDependencies : size_t The size of the dependencies array flags : unsigned int See above Returns ------- cudaError_t :py:obj:`~.cudaSuccess`, :py:obj:`~.cudaErrorInvalidValue`, :py:obj:`~.cudaErrorIllegalState` See Also -------- :py:obj:`~.cudaStreamBeginCapture`, :py:obj:`~.cudaStreamGetCaptureInfo`, """ dependencyData = [] if dependencyData is None else dependencyData if not all(isinstance(_x, (cudaGraphEdgeData,)) for _x in dependencyData): raise TypeError("Argument 'dependencyData' is not instance of type (expected tuple[cyruntime.cudaGraphEdgeData,] or list[cyruntime.cudaGraphEdgeData,]") dependencies = [] if dependencies is None else dependencies if not all(isinstance(_x, (cudaGraphNode_t,driver.CUgraphNode)) for _x in dependencies): raise TypeError("Argument 'dependencies' is not instance of type (expected tuple[cyruntime.cudaGraphNode_t,driver.CUgraphNode] or list[cyruntime.cudaGraphNode_t,driver.CUgraphNode]") cdef cyruntime.cudaStream_t cystream if stream is None: pstream = 0 elif isinstance(stream, (cudaStream_t,driver.CUstream)): pstream = int(stream) else: pstream = int(cudaStream_t(stream)) cystream = pstream cdef cyruntime.cudaGraphNode_t* cydependencies = NULL if len(dependencies) > 1: cydependencies = calloc(len(dependencies), sizeof(cyruntime.cudaGraphNode_t)) if cydependencies is NULL: raise MemoryError('Failed to allocate length x size memory: ' + str(len(dependencies)) + 'x' + str(sizeof(cyruntime.cudaGraphNode_t))) else: for idx in range(len(dependencies)): cydependencies[idx] = (dependencies[idx])._pvt_ptr[0] elif len(dependencies) == 1: cydependencies = (dependencies[0])._pvt_ptr cdef cyruntime.cudaGraphEdgeData* cydependencyData = NULL if len(dependencyData) > 1: cydependencyData = calloc(len(dependencyData), sizeof(cyruntime.cudaGraphEdgeData)) if cydependencyData is NULL: raise MemoryError('Failed to allocate length x size memory: ' + str(len(dependencyData)) + 'x' + str(sizeof(cyruntime.cudaGraphEdgeData))) for idx in range(len(dependencyData)): string.memcpy(&cydependencyData[idx], (dependencyData[idx])._pvt_ptr, sizeof(cyruntime.cudaGraphEdgeData)) elif len(dependencyData) == 1: cydependencyData = (dependencyData[0])._pvt_ptr with nogil: err = cyruntime.cudaStreamUpdateCaptureDependencies_v2(cystream, cydependencies, cydependencyData, numDependencies, flags) if len(dependencies) > 1 and cydependencies is not NULL: free(cydependencies) if len(dependencyData) > 1 and cydependencyData is not NULL: free(cydependencyData) return (_dict_cudaError_t[err],) @cython.embedsignature(True) def cudaEventCreate(): """ Creates an event object. Creates an event object for the current device using :py:obj:`~.cudaEventDefault`. Returns ------- cudaError_t :py:obj:`~.cudaSuccess`, :py:obj:`~.cudaErrorInvalidValue`, :py:obj:`~.cudaErrorLaunchFailure`, :py:obj:`~.cudaErrorMemoryAllocation` event : :py:obj:`~.cudaEvent_t` Newly created event See Also -------- cudaEventCreate (C++ API), :py:obj:`~.cudaEventCreateWithFlags`, :py:obj:`~.cudaEventRecord`, :py:obj:`~.cudaEventQuery`, :py:obj:`~.cudaEventSynchronize`, :py:obj:`~.cudaEventDestroy`, :py:obj:`~.cudaEventElapsedTime`, :py:obj:`~.cudaStreamWaitEvent`, :py:obj:`~.cuEventCreate` """ cdef cudaEvent_t event = cudaEvent_t() with nogil: err = cyruntime.cudaEventCreate(event._pvt_ptr) if err != cyruntime.cudaSuccess: return (_dict_cudaError_t[err], None) return (_dict_cudaError_t[err], event) @cython.embedsignature(True) def cudaEventCreateWithFlags(unsigned int flags): """ Creates an event object with the specified flags. Creates an event object for the current device with the specified flags. Valid flags include: - :py:obj:`~.cudaEventDefault`: Default event creation flag. - :py:obj:`~.cudaEventBlockingSync`: Specifies that event should use blocking synchronization. A host thread that uses :py:obj:`~.cudaEventSynchronize()` to wait on an event created with this flag will block until the event actually completes. - :py:obj:`~.cudaEventDisableTiming`: Specifies that the created event does not need to record timing data. Events created with this flag specified and the :py:obj:`~.cudaEventBlockingSync` flag not specified will provide the best performance when used with :py:obj:`~.cudaStreamWaitEvent()` and :py:obj:`~.cudaEventQuery()`. - :py:obj:`~.cudaEventInterprocess`: Specifies that the created event may be used as an interprocess event by :py:obj:`~.cudaIpcGetEventHandle()`. :py:obj:`~.cudaEventInterprocess` must be specified along with :py:obj:`~.cudaEventDisableTiming`. Parameters ---------- flags : unsigned int Flags for new event Returns ------- cudaError_t :py:obj:`~.cudaSuccess`, :py:obj:`~.cudaErrorInvalidValue`, :py:obj:`~.cudaErrorLaunchFailure`, :py:obj:`~.cudaErrorMemoryAllocation` event : :py:obj:`~.cudaEvent_t` Newly created event See Also -------- :py:obj:`~.cudaEventCreate (C API)`, :py:obj:`~.cudaEventSynchronize`, :py:obj:`~.cudaEventDestroy`, :py:obj:`~.cudaEventElapsedTime`, :py:obj:`~.cudaStreamWaitEvent`, :py:obj:`~.cuEventCreate` """ cdef cudaEvent_t event = cudaEvent_t() with nogil: err = cyruntime.cudaEventCreateWithFlags(event._pvt_ptr, flags) if err != cyruntime.cudaSuccess: return (_dict_cudaError_t[err], None) return (_dict_cudaError_t[err], event) @cython.embedsignature(True) def cudaEventRecord(event, stream): """ Records an event. Captures in `event` the contents of `stream` at the time of this call. `event` and `stream` must be on the same CUDA context. Calls such as :py:obj:`~.cudaEventQuery()` or :py:obj:`~.cudaStreamWaitEvent()` will then examine or wait for completion of the work that was captured. Uses of `stream` after this call do not modify `event`. See note on default stream behavior for what is captured in the default case. :py:obj:`~.cudaEventRecord()` can be called multiple times on the same event and will overwrite the previously captured state. Other APIs such as :py:obj:`~.cudaStreamWaitEvent()` use the most recently captured state at the time of the API call, and are not affected by later calls to :py:obj:`~.cudaEventRecord()`. Before the first call to :py:obj:`~.cudaEventRecord()`, an event represents an empty set of work, so for example :py:obj:`~.cudaEventQuery()` would return :py:obj:`~.cudaSuccess`. Parameters ---------- event : :py:obj:`~.CUevent` or :py:obj:`~.cudaEvent_t` Event to record stream : :py:obj:`~.CUstream` or :py:obj:`~.cudaStream_t` Stream in which to record event Returns ------- cudaError_t :py:obj:`~.cudaSuccess`, :py:obj:`~.cudaErrorInvalidValue`, :py:obj:`~.cudaErrorInvalidResourceHandle`, :py:obj:`~.cudaErrorLaunchFailure` See Also -------- :py:obj:`~.cudaEventCreate (C API)`, :py:obj:`~.cudaEventCreateWithFlags`, :py:obj:`~.cudaEventQuery`, :py:obj:`~.cudaEventSynchronize`, :py:obj:`~.cudaEventDestroy`, :py:obj:`~.cudaEventElapsedTime`, :py:obj:`~.cudaStreamWaitEvent`, :py:obj:`~.cudaEventRecordWithFlags`, :py:obj:`~.cuEventRecord` """ cdef cyruntime.cudaStream_t cystream if stream is None: pstream = 0 elif isinstance(stream, (cudaStream_t,driver.CUstream)): pstream = int(stream) else: pstream = int(cudaStream_t(stream)) cystream = pstream cdef cyruntime.cudaEvent_t cyevent if event is None: pevent = 0 elif isinstance(event, (cudaEvent_t,driver.CUevent)): pevent = int(event) else: pevent = int(cudaEvent_t(event)) cyevent = pevent with nogil: err = cyruntime.cudaEventRecord(cyevent, cystream) return (_dict_cudaError_t[err],) @cython.embedsignature(True) def cudaEventRecordWithFlags(event, stream, unsigned int flags): """ Records an event. Captures in `event` the contents of `stream` at the time of this call. `event` and `stream` must be on the same CUDA context. Calls such as :py:obj:`~.cudaEventQuery()` or :py:obj:`~.cudaStreamWaitEvent()` will then examine or wait for completion of the work that was captured. Uses of `stream` after this call do not modify `event`. See note on default stream behavior for what is captured in the default case. :py:obj:`~.cudaEventRecordWithFlags()` can be called multiple times on the same event and will overwrite the previously captured state. Other APIs such as :py:obj:`~.cudaStreamWaitEvent()` use the most recently captured state at the time of the API call, and are not affected by later calls to :py:obj:`~.cudaEventRecordWithFlags()`. Before the first call to :py:obj:`~.cudaEventRecordWithFlags()`, an event represents an empty set of work, so for example :py:obj:`~.cudaEventQuery()` would return :py:obj:`~.cudaSuccess`. flags include: - :py:obj:`~.cudaEventRecordDefault`: Default event creation flag. - :py:obj:`~.cudaEventRecordExternal`: Event is captured in the graph as an external event node when performing stream capture. Parameters ---------- event : :py:obj:`~.CUevent` or :py:obj:`~.cudaEvent_t` Event to record stream : :py:obj:`~.CUstream` or :py:obj:`~.cudaStream_t` Stream in which to record event flags : unsigned int Parameters for the operation(See above) Returns ------- cudaError_t :py:obj:`~.cudaSuccess`, :py:obj:`~.cudaErrorInvalidValue`, :py:obj:`~.cudaErrorInvalidResourceHandle`, :py:obj:`~.cudaErrorLaunchFailure` See Also -------- :py:obj:`~.cudaEventCreate (C API)`, :py:obj:`~.cudaEventCreateWithFlags`, :py:obj:`~.cudaEventQuery`, :py:obj:`~.cudaEventSynchronize`, :py:obj:`~.cudaEventDestroy`, :py:obj:`~.cudaEventElapsedTime`, :py:obj:`~.cudaStreamWaitEvent`, :py:obj:`~.cudaEventRecord`, :py:obj:`~.cuEventRecord`, """ cdef cyruntime.cudaStream_t cystream if stream is None: pstream = 0 elif isinstance(stream, (cudaStream_t,driver.CUstream)): pstream = int(stream) else: pstream = int(cudaStream_t(stream)) cystream = pstream cdef cyruntime.cudaEvent_t cyevent if event is None: pevent = 0 elif isinstance(event, (cudaEvent_t,driver.CUevent)): pevent = int(event) else: pevent = int(cudaEvent_t(event)) cyevent = pevent with nogil: err = cyruntime.cudaEventRecordWithFlags(cyevent, cystream, flags) return (_dict_cudaError_t[err],) @cython.embedsignature(True) def cudaEventQuery(event): """ Queries an event's status. Queries the status of all work currently captured by `event`. See :py:obj:`~.cudaEventRecord()` for details on what is captured by an event. Returns :py:obj:`~.cudaSuccess` if all captured work has been completed, or :py:obj:`~.cudaErrorNotReady` if any captured work is incomplete. For the purposes of Unified Memory, a return value of :py:obj:`~.cudaSuccess` is equivalent to having called :py:obj:`~.cudaEventSynchronize()`. Parameters ---------- event : :py:obj:`~.CUevent` or :py:obj:`~.cudaEvent_t` Event to query Returns ------- cudaError_t :py:obj:`~.cudaSuccess`, :py:obj:`~.cudaErrorNotReady`, :py:obj:`~.cudaErrorInvalidValue`, :py:obj:`~.cudaErrorInvalidResourceHandle`, :py:obj:`~.cudaErrorLaunchFailure` See Also -------- :py:obj:`~.cudaEventCreate (C API)`, :py:obj:`~.cudaEventCreateWithFlags`, :py:obj:`~.cudaEventRecord`, :py:obj:`~.cudaEventSynchronize`, :py:obj:`~.cudaEventDestroy`, :py:obj:`~.cudaEventElapsedTime`, :py:obj:`~.cuEventQuery` """ cdef cyruntime.cudaEvent_t cyevent if event is None: pevent = 0 elif isinstance(event, (cudaEvent_t,driver.CUevent)): pevent = int(event) else: pevent = int(cudaEvent_t(event)) cyevent = pevent with nogil: err = cyruntime.cudaEventQuery(cyevent) return (_dict_cudaError_t[err],) @cython.embedsignature(True) def cudaEventSynchronize(event): """ Waits for an event to complete. Waits until the completion of all work currently captured in `event`. See :py:obj:`~.cudaEventRecord()` for details on what is captured by an event. Waiting for an event that was created with the :py:obj:`~.cudaEventBlockingSync` flag will cause the calling CPU thread to block until the event has been completed by the device. If the :py:obj:`~.cudaEventBlockingSync` flag has not been set, then the CPU thread will busy-wait until the event has been completed by the device. Parameters ---------- event : :py:obj:`~.CUevent` or :py:obj:`~.cudaEvent_t` Event to wait for Returns ------- cudaError_t :py:obj:`~.cudaSuccess`, :py:obj:`~.cudaErrorInvalidValue`, :py:obj:`~.cudaErrorInvalidResourceHandle`, :py:obj:`~.cudaErrorLaunchFailure` See Also -------- :py:obj:`~.cudaEventCreate (C API)`, :py:obj:`~.cudaEventCreateWithFlags`, :py:obj:`~.cudaEventRecord`, :py:obj:`~.cudaEventQuery`, :py:obj:`~.cudaEventDestroy`, :py:obj:`~.cudaEventElapsedTime`, :py:obj:`~.cuEventSynchronize` """ cdef cyruntime.cudaEvent_t cyevent if event is None: pevent = 0 elif isinstance(event, (cudaEvent_t,driver.CUevent)): pevent = int(event) else: pevent = int(cudaEvent_t(event)) cyevent = pevent with nogil: err = cyruntime.cudaEventSynchronize(cyevent) return (_dict_cudaError_t[err],) @cython.embedsignature(True) def cudaEventDestroy(event): """ Destroys an event object. Destroys the event specified by `event`. An event may be destroyed before it is complete (i.e., while :py:obj:`~.cudaEventQuery()` would return :py:obj:`~.cudaErrorNotReady`). In this case, the call does not block on completion of the event, and any associated resources will automatically be released asynchronously at completion. Parameters ---------- event : :py:obj:`~.CUevent` or :py:obj:`~.cudaEvent_t` Event to destroy Returns ------- cudaError_t :py:obj:`~.cudaSuccess`, :py:obj:`~.cudaErrorInvalidValue`, :py:obj:`~.cudaErrorInvalidResourceHandle`, :py:obj:`~.cudaErrorLaunchFailure` See Also -------- :py:obj:`~.cudaEventCreate (C API)`, :py:obj:`~.cudaEventCreateWithFlags`, :py:obj:`~.cudaEventQuery`, :py:obj:`~.cudaEventSynchronize`, :py:obj:`~.cudaEventRecord`, :py:obj:`~.cudaEventElapsedTime`, :py:obj:`~.cuEventDestroy` """ cdef cyruntime.cudaEvent_t cyevent if event is None: pevent = 0 elif isinstance(event, (cudaEvent_t,driver.CUevent)): pevent = int(event) else: pevent = int(cudaEvent_t(event)) cyevent = pevent with nogil: err = cyruntime.cudaEventDestroy(cyevent) return (_dict_cudaError_t[err],) @cython.embedsignature(True) def cudaEventElapsedTime(start, end): """ Computes the elapsed time between events. Computes the elapsed time between two events (in milliseconds with a resolution of around 0.5 microseconds). If either event was last recorded in a non-NULL stream, the resulting time may be greater than expected (even if both used the same stream handle). This happens because the :py:obj:`~.cudaEventRecord()` operation takes place asynchronously and there is no guarantee that the measured latency is actually just between the two events. Any number of other different stream operations could execute in between the two measured events, thus altering the timing in a significant way. If :py:obj:`~.cudaEventRecord()` has not been called on either event, then :py:obj:`~.cudaErrorInvalidResourceHandle` is returned. If :py:obj:`~.cudaEventRecord()` has been called on both events but one or both of them has not yet been completed (that is, :py:obj:`~.cudaEventQuery()` would return :py:obj:`~.cudaErrorNotReady` on at least one of the events), :py:obj:`~.cudaErrorNotReady` is returned. If either event was created with the :py:obj:`~.cudaEventDisableTiming` flag, then this function will return :py:obj:`~.cudaErrorInvalidResourceHandle`. Parameters ---------- start : :py:obj:`~.CUevent` or :py:obj:`~.cudaEvent_t` Starting event end : :py:obj:`~.CUevent` or :py:obj:`~.cudaEvent_t` Ending event Returns ------- cudaError_t :py:obj:`~.cudaSuccess`, :py:obj:`~.cudaErrorNotReady`, :py:obj:`~.cudaErrorInvalidValue`, :py:obj:`~.cudaErrorInvalidResourceHandle`, :py:obj:`~.cudaErrorLaunchFailure`, :py:obj:`~.cudaErrorUnknown` ms : float Time between `start` and `end` in ms See Also -------- :py:obj:`~.cudaEventCreate (C API)`, :py:obj:`~.cudaEventCreateWithFlags`, :py:obj:`~.cudaEventQuery`, :py:obj:`~.cudaEventSynchronize`, :py:obj:`~.cudaEventDestroy`, :py:obj:`~.cudaEventRecord`, :py:obj:`~.cuEventElapsedTime` """ cdef cyruntime.cudaEvent_t cyend if end is None: pend = 0 elif isinstance(end, (cudaEvent_t,driver.CUevent)): pend = int(end) else: pend = int(cudaEvent_t(end)) cyend = pend cdef cyruntime.cudaEvent_t cystart if start is None: pstart = 0 elif isinstance(start, (cudaEvent_t,driver.CUevent)): pstart = int(start) else: pstart = int(cudaEvent_t(start)) cystart = pstart cdef float ms = 0 with nogil: err = cyruntime.cudaEventElapsedTime(&ms, cystart, cyend) if err != cyruntime.cudaSuccess: return (_dict_cudaError_t[err], None) return (_dict_cudaError_t[err], ms) @cython.embedsignature(True) def cudaEventElapsedTime_v2(start, end): """ Computes the elapsed time between events. Computes the elapsed time between two events (in milliseconds with a resolution of around 0.5 microseconds). Note this API is not guaranteed to return the latest errors for pending work. As such this API is intended to serve as a elapsed time calculation only and polling for completion on the events to be compared should be done with :py:obj:`~.cudaEventQuery` instead. If either event was last recorded in a non-NULL stream, the resulting time may be greater than expected (even if both used the same stream handle). This happens because the :py:obj:`~.cudaEventRecord()` operation takes place asynchronously and there is no guarantee that the measured latency is actually just between the two events. Any number of other different stream operations could execute in between the two measured events, thus altering the timing in a significant way. If :py:obj:`~.cudaEventRecord()` has not been called on either event, then :py:obj:`~.cudaErrorInvalidResourceHandle` is returned. If :py:obj:`~.cudaEventRecord()` has been called on both events but one or both of them has not yet been completed (that is, :py:obj:`~.cudaEventQuery()` would return :py:obj:`~.cudaErrorNotReady` on at least one of the events), :py:obj:`~.cudaErrorNotReady` is returned. If either event was created with the :py:obj:`~.cudaEventDisableTiming` flag, then this function will return :py:obj:`~.cudaErrorInvalidResourceHandle`. Parameters ---------- start : :py:obj:`~.CUevent` or :py:obj:`~.cudaEvent_t` Starting event end : :py:obj:`~.CUevent` or :py:obj:`~.cudaEvent_t` Ending event Returns ------- cudaError_t :py:obj:`~.cudaSuccess`, :py:obj:`~.cudaErrorNotReady`, :py:obj:`~.cudaErrorInvalidValue`, :py:obj:`~.cudaErrorInvalidResourceHandle`, :py:obj:`~.cudaErrorLaunchFailure`, :py:obj:`~.cudaErrorUnknown` ms : float Time between `start` and `end` in ms See Also -------- :py:obj:`~.cudaEventCreate (C API)`, :py:obj:`~.cudaEventCreateWithFlags`, :py:obj:`~.cudaEventQuery`, :py:obj:`~.cudaEventSynchronize`, :py:obj:`~.cudaEventDestroy`, :py:obj:`~.cudaEventRecord`, :py:obj:`~.cuEventElapsedTime` """ cdef cyruntime.cudaEvent_t cyend if end is None: pend = 0 elif isinstance(end, (cudaEvent_t,driver.CUevent)): pend = int(end) else: pend = int(cudaEvent_t(end)) cyend = pend cdef cyruntime.cudaEvent_t cystart if start is None: pstart = 0 elif isinstance(start, (cudaEvent_t,driver.CUevent)): pstart = int(start) else: pstart = int(cudaEvent_t(start)) cystart = pstart cdef float ms = 0 with nogil: err = cyruntime.cudaEventElapsedTime_v2(&ms, cystart, cyend) if err != cyruntime.cudaSuccess: return (_dict_cudaError_t[err], None) return (_dict_cudaError_t[err], ms) @cython.embedsignature(True) def cudaImportExternalMemory(memHandleDesc : Optional[cudaExternalMemoryHandleDesc]): """ Imports an external memory object. Imports an externally allocated memory object and returns a handle to that in `extMem_out`. The properties of the handle being imported must be described in `memHandleDesc`. The :py:obj:`~.cudaExternalMemoryHandleDesc` structure is defined as follows: **View CUDA Toolkit Documentation for a C++ code example** where :py:obj:`~.cudaExternalMemoryHandleDesc.type` specifies the type of handle being imported. :py:obj:`~.cudaExternalMemoryHandleType` is defined as: **View CUDA Toolkit Documentation for a C++ code example** If :py:obj:`~.cudaExternalMemoryHandleDesc.type` is :py:obj:`~.cudaExternalMemoryHandleTypeOpaqueFd`, then :py:obj:`~.cudaExternalMemoryHandleDesc`::handle::fd must be a valid file descriptor referencing a memory object. Ownership of the file descriptor is transferred to the CUDA driver when the handle is imported successfully. Performing any operations on the file descriptor after it is imported results in undefined behavior. If :py:obj:`~.cudaExternalMemoryHandleDesc.type` is :py:obj:`~.cudaExternalMemoryHandleTypeOpaqueWin32`, then exactly one of :py:obj:`~.cudaExternalMemoryHandleDesc`::handle::win32::handle and :py:obj:`~.cudaExternalMemoryHandleDesc`::handle::win32::name must not be NULL. If :py:obj:`~.cudaExternalMemoryHandleDesc`::handle::win32::handle is not NULL, then it must represent a valid shared NT handle that references a memory object. Ownership of this handle is not transferred to CUDA after the import operation, so the application must release the handle using the appropriate system call. If :py:obj:`~.cudaExternalMemoryHandleDesc`::handle::win32::name is not NULL, then it must point to a NULL-terminated array of UTF-16 characters that refers to a memory object. If :py:obj:`~.cudaExternalMemoryHandleDesc.type` is :py:obj:`~.cudaExternalMemoryHandleTypeOpaqueWin32Kmt`, then :py:obj:`~.cudaExternalMemoryHandleDesc`::handle::win32::handle must be non-NULL and :py:obj:`~.cudaExternalMemoryHandleDesc`::handle::win32::name must be NULL. The handle specified must be a globally shared KMT handle. This handle does not hold a reference to the underlying object, and thus will be invalid when all references to the memory object are destroyed. If :py:obj:`~.cudaExternalMemoryHandleDesc.type` is :py:obj:`~.cudaExternalMemoryHandleTypeD3D12Heap`, then exactly one of :py:obj:`~.cudaExternalMemoryHandleDesc`::handle::win32::handle and :py:obj:`~.cudaExternalMemoryHandleDesc`::handle::win32::name must not be NULL. If :py:obj:`~.cudaExternalMemoryHandleDesc`::handle::win32::handle is not NULL, then it must represent a valid shared NT handle that is returned by ID3D12Device::CreateSharedHandle when referring to a ID3D12Heap object. This handle holds a reference to the underlying object. If :py:obj:`~.cudaExternalMemoryHandleDesc`::handle::win32::name is not NULL, then it must point to a NULL-terminated array of UTF-16 characters that refers to a ID3D12Heap object. If :py:obj:`~.cudaExternalMemoryHandleDesc.type` is :py:obj:`~.cudaExternalMemoryHandleTypeD3D12Resource`, then exactly one of :py:obj:`~.cudaExternalMemoryHandleDesc`::handle::win32::handle and :py:obj:`~.cudaExternalMemoryHandleDesc`::handle::win32::name must not be NULL. If :py:obj:`~.cudaExternalMemoryHandleDesc`::handle::win32::handle is not NULL, then it must represent a valid shared NT handle that is returned by ID3D12Device::CreateSharedHandle when referring to a ID3D12Resource object. This handle holds a reference to the underlying object. If :py:obj:`~.cudaExternalMemoryHandleDesc`::handle::win32::name is not NULL, then it must point to a NULL-terminated array of UTF-16 characters that refers to a ID3D12Resource object. If :py:obj:`~.cudaExternalMemoryHandleDesc.type` is :py:obj:`~.cudaExternalMemoryHandleTypeD3D11Resource`,then exactly one of :py:obj:`~.cudaExternalMemoryHandleDesc`::handle::win32::handle and :py:obj:`~.cudaExternalMemoryHandleDesc`::handle::win32::name must not be NULL. If :py:obj:`~.cudaExternalMemoryHandleDesc`::handle::win32::handle is not NULL, then it must represent a valid shared NT handle that is returned by IDXGIResource1::CreateSharedHandle when referring to a ID3D11Resource object. If :py:obj:`~.cudaExternalMemoryHandleDesc`::handle::win32::name is not NULL, then it must point to a NULL-terminated array of UTF-16 characters that refers to a ID3D11Resource object. If :py:obj:`~.cudaExternalMemoryHandleDesc.type` is :py:obj:`~.cudaExternalMemoryHandleTypeD3D11ResourceKmt`, then :py:obj:`~.cudaExternalMemoryHandleDesc`::handle::win32::handle must be non-NULL and :py:obj:`~.cudaExternalMemoryHandleDesc`::handle::win32::name must be NULL. The handle specified must be a valid shared KMT handle that is returned by IDXGIResource::GetSharedHandle when referring to a ID3D11Resource object. If :py:obj:`~.cudaExternalMemoryHandleDesc.type` is :py:obj:`~.cudaExternalMemoryHandleTypeNvSciBuf`, then :py:obj:`~.cudaExternalMemoryHandleDesc`::handle::nvSciBufObject must be NON-NULL and reference a valid NvSciBuf object. If the NvSciBuf object imported into CUDA is also mapped by other drivers, then the application must use :py:obj:`~.cudaWaitExternalSemaphoresAsync` or :py:obj:`~.cudaSignalExternalSemaphoresAsync` as approprriate barriers to maintain coherence between CUDA and the other drivers. See :py:obj:`~.cudaExternalSemaphoreWaitSkipNvSciBufMemSync` and :py:obj:`~.cudaExternalSemaphoreSignalSkipNvSciBufMemSync` for memory synchronization. The size of the memory object must be specified in :py:obj:`~.cudaExternalMemoryHandleDesc.size`. Specifying the flag :py:obj:`~.cudaExternalMemoryDedicated` in :py:obj:`~.cudaExternalMemoryHandleDesc.flags` indicates that the resource is a dedicated resource. The definition of what a dedicated resource is outside the scope of this extension. This flag must be set if :py:obj:`~.cudaExternalMemoryHandleDesc.type` is one of the following: :py:obj:`~.cudaExternalMemoryHandleTypeD3D12Resource` :py:obj:`~.cudaExternalMemoryHandleTypeD3D11Resource` :py:obj:`~.cudaExternalMemoryHandleTypeD3D11ResourceKmt` Parameters ---------- memHandleDesc : :py:obj:`~.cudaExternalMemoryHandleDesc` Memory import handle descriptor Returns ------- cudaError_t :py:obj:`~.cudaSuccess`, :py:obj:`~.cudaErrorInvalidValue`, :py:obj:`~.cudaErrorInvalidResourceHandle`, :py:obj:`~.cudaErrorOperatingSystem` extMem_out : :py:obj:`~.cudaExternalMemory_t` Returned handle to an external memory object See Also -------- :py:obj:`~.cudaDestroyExternalMemory`, :py:obj:`~.cudaExternalMemoryGetMappedBuffer`, :py:obj:`~.cudaExternalMemoryGetMappedMipmappedArray` Notes ----- If the Vulkan memory imported into CUDA is mapped on the CPU then the application must use vkInvalidateMappedMemoryRanges/vkFlushMappedMemoryRanges as well as appropriate Vulkan pipeline barriers to maintain coherence between CPU and GPU. For more information on these APIs, please refer to "Synchronization and Cache Control" chapter from Vulkan specification. """ cdef cudaExternalMemory_t extMem_out = cudaExternalMemory_t() cdef cyruntime.cudaExternalMemoryHandleDesc* cymemHandleDesc_ptr = memHandleDesc._pvt_ptr if memHandleDesc is not None else NULL with nogil: err = cyruntime.cudaImportExternalMemory(extMem_out._pvt_ptr, cymemHandleDesc_ptr) if err != cyruntime.cudaSuccess: return (_dict_cudaError_t[err], None) return (_dict_cudaError_t[err], extMem_out) @cython.embedsignature(True) def cudaExternalMemoryGetMappedBuffer(extMem, bufferDesc : Optional[cudaExternalMemoryBufferDesc]): """ Maps a buffer onto an imported memory object. Maps a buffer onto an imported memory object and returns a device pointer in `devPtr`. The properties of the buffer being mapped must be described in `bufferDesc`. The :py:obj:`~.cudaExternalMemoryBufferDesc` structure is defined as follows: **View CUDA Toolkit Documentation for a C++ code example** where :py:obj:`~.cudaExternalMemoryBufferDesc.offset` is the offset in the memory object where the buffer's base address is. :py:obj:`~.cudaExternalMemoryBufferDesc.size` is the size of the buffer. :py:obj:`~.cudaExternalMemoryBufferDesc.flags` must be zero. The offset and size have to be suitably aligned to match the requirements of the external API. Mapping two buffers whose ranges overlap may or may not result in the same virtual address being returned for the overlapped portion. In such cases, the application must ensure that all accesses to that region from the GPU are volatile. Otherwise writes made via one address are not guaranteed to be visible via the other address, even if they're issued by the same thread. It is recommended that applications map the combined range instead of mapping separate buffers and then apply the appropriate offsets to the returned pointer to derive the individual buffers. The returned pointer `devPtr` must be freed using :py:obj:`~.cudaFree`. Parameters ---------- extMem : :py:obj:`~.cudaExternalMemory_t` Handle to external memory object bufferDesc : :py:obj:`~.cudaExternalMemoryBufferDesc` Buffer descriptor Returns ------- cudaError_t :py:obj:`~.cudaSuccess`, :py:obj:`~.cudaErrorInvalidValue`, :py:obj:`~.cudaErrorInvalidResourceHandle` devPtr : Any Returned device pointer to buffer See Also -------- :py:obj:`~.cudaImportExternalMemory`, :py:obj:`~.cudaDestroyExternalMemory`, :py:obj:`~.cudaExternalMemoryGetMappedMipmappedArray` """ cdef cyruntime.cudaExternalMemory_t cyextMem if extMem is None: pextMem = 0 elif isinstance(extMem, (cudaExternalMemory_t,)): pextMem = int(extMem) else: pextMem = int(cudaExternalMemory_t(extMem)) cyextMem = pextMem cdef void_ptr devPtr = 0 cdef cyruntime.cudaExternalMemoryBufferDesc* cybufferDesc_ptr = bufferDesc._pvt_ptr if bufferDesc is not None else NULL with nogil: err = cyruntime.cudaExternalMemoryGetMappedBuffer(&devPtr, cyextMem, cybufferDesc_ptr) if err != cyruntime.cudaSuccess: return (_dict_cudaError_t[err], None) return (_dict_cudaError_t[err], devPtr) @cython.embedsignature(True) def cudaExternalMemoryGetMappedMipmappedArray(extMem, mipmapDesc : Optional[cudaExternalMemoryMipmappedArrayDesc]): """ Maps a CUDA mipmapped array onto an external memory object. Maps a CUDA mipmapped array onto an external object and returns a handle to it in `mipmap`. The properties of the CUDA mipmapped array being mapped must be described in `mipmapDesc`. The structure :py:obj:`~.cudaExternalMemoryMipmappedArrayDesc` is defined as follows: **View CUDA Toolkit Documentation for a C++ code example** where :py:obj:`~.cudaExternalMemoryMipmappedArrayDesc.offset` is the offset in the memory object where the base level of the mipmap chain is. :py:obj:`~.cudaExternalMemoryMipmappedArrayDesc.formatDesc` describes the format of the data. :py:obj:`~.cudaExternalMemoryMipmappedArrayDesc.extent` specifies the dimensions of the base level of the mipmap chain. :py:obj:`~.cudaExternalMemoryMipmappedArrayDesc.flags` are flags associated with CUDA mipmapped arrays. For further details, please refer to the documentation for :py:obj:`~.cudaMalloc3DArray`. Note that if the mipmapped array is bound as a color target in the graphics API, then the flag :py:obj:`~.cudaArrayColorAttachment` must be specified in :py:obj:`~.cudaExternalMemoryMipmappedArrayDesc.flags`. :py:obj:`~.cudaExternalMemoryMipmappedArrayDesc.numLevels` specifies the total number of levels in the mipmap chain. The returned CUDA mipmapped array must be freed using :py:obj:`~.cudaFreeMipmappedArray`. Parameters ---------- extMem : :py:obj:`~.cudaExternalMemory_t` Handle to external memory object mipmapDesc : :py:obj:`~.cudaExternalMemoryMipmappedArrayDesc` CUDA array descriptor Returns ------- cudaError_t :py:obj:`~.cudaSuccess`, :py:obj:`~.cudaErrorInvalidValue`, :py:obj:`~.cudaErrorInvalidResourceHandle` mipmap : :py:obj:`~.cudaMipmappedArray_t` Returned CUDA mipmapped array See Also -------- :py:obj:`~.cudaImportExternalMemory`, :py:obj:`~.cudaDestroyExternalMemory`, :py:obj:`~.cudaExternalMemoryGetMappedBuffer` Notes ----- If :py:obj:`~.cudaExternalMemoryHandleDesc.type` is :py:obj:`~.cudaExternalMemoryHandleTypeNvSciBuf`, then :py:obj:`~.cudaExternalMemoryMipmappedArrayDesc.numLevels` must not be greater than 1. """ cdef cyruntime.cudaExternalMemory_t cyextMem if extMem is None: pextMem = 0 elif isinstance(extMem, (cudaExternalMemory_t,)): pextMem = int(extMem) else: pextMem = int(cudaExternalMemory_t(extMem)) cyextMem = pextMem cdef cudaMipmappedArray_t mipmap = cudaMipmappedArray_t() cdef cyruntime.cudaExternalMemoryMipmappedArrayDesc* cymipmapDesc_ptr = mipmapDesc._pvt_ptr if mipmapDesc is not None else NULL with nogil: err = cyruntime.cudaExternalMemoryGetMappedMipmappedArray(mipmap._pvt_ptr, cyextMem, cymipmapDesc_ptr) if err != cyruntime.cudaSuccess: return (_dict_cudaError_t[err], None) return (_dict_cudaError_t[err], mipmap) @cython.embedsignature(True) def cudaDestroyExternalMemory(extMem): """ Destroys an external memory object. Destroys the specified external memory object. Any existing buffers and CUDA mipmapped arrays mapped onto this object must no longer be used and must be explicitly freed using :py:obj:`~.cudaFree` and :py:obj:`~.cudaFreeMipmappedArray` respectively. Parameters ---------- extMem : :py:obj:`~.cudaExternalMemory_t` External memory object to be destroyed Returns ------- cudaError_t :py:obj:`~.cudaSuccess`, :py:obj:`~.cudaErrorInvalidResourceHandle` See Also -------- :py:obj:`~.cudaImportExternalMemory`, :py:obj:`~.cudaExternalMemoryGetMappedBuffer`, :py:obj:`~.cudaExternalMemoryGetMappedMipmappedArray` """ cdef cyruntime.cudaExternalMemory_t cyextMem if extMem is None: pextMem = 0 elif isinstance(extMem, (cudaExternalMemory_t,)): pextMem = int(extMem) else: pextMem = int(cudaExternalMemory_t(extMem)) cyextMem = pextMem with nogil: err = cyruntime.cudaDestroyExternalMemory(cyextMem) return (_dict_cudaError_t[err],) @cython.embedsignature(True) def cudaImportExternalSemaphore(semHandleDesc : Optional[cudaExternalSemaphoreHandleDesc]): """ Imports an external semaphore. Imports an externally allocated synchronization object and returns a handle to that in `extSem_out`. The properties of the handle being imported must be described in `semHandleDesc`. The :py:obj:`~.cudaExternalSemaphoreHandleDesc` is defined as follows: **View CUDA Toolkit Documentation for a C++ code example** where :py:obj:`~.cudaExternalSemaphoreHandleDesc.type` specifies the type of handle being imported. :py:obj:`~.cudaExternalSemaphoreHandleType` is defined as: **View CUDA Toolkit Documentation for a C++ code example** If :py:obj:`~.cudaExternalSemaphoreHandleDesc.type` is :py:obj:`~.cudaExternalSemaphoreHandleTypeOpaqueFd`, then :py:obj:`~.cudaExternalSemaphoreHandleDesc`::handle::fd must be a valid file descriptor referencing a synchronization object. Ownership of the file descriptor is transferred to the CUDA driver when the handle is imported successfully. Performing any operations on the file descriptor after it is imported results in undefined behavior. If :py:obj:`~.cudaExternalSemaphoreHandleDesc.type` is :py:obj:`~.cudaExternalSemaphoreHandleTypeOpaqueWin32`, then exactly one of :py:obj:`~.cudaExternalSemaphoreHandleDesc`::handle::win32::handle and :py:obj:`~.cudaExternalSemaphoreHandleDesc`::handle::win32::name must not be NULL. If :py:obj:`~.cudaExternalSemaphoreHandleDesc`::handle::win32::handle is not NULL, then it must represent a valid shared NT handle that references a synchronization object. Ownership of this handle is not transferred to CUDA after the import operation, so the application must release the handle using the appropriate system call. If :py:obj:`~.cudaExternalSemaphoreHandleDesc`::handle::win32::name is not NULL, then it must name a valid synchronization object. If :py:obj:`~.cudaExternalSemaphoreHandleDesc.type` is :py:obj:`~.cudaExternalSemaphoreHandleTypeOpaqueWin32Kmt`, then :py:obj:`~.cudaExternalSemaphoreHandleDesc`::handle::win32::handle must be non-NULL and :py:obj:`~.cudaExternalSemaphoreHandleDesc`::handle::win32::name must be NULL. The handle specified must be a globally shared KMT handle. This handle does not hold a reference to the underlying object, and thus will be invalid when all references to the synchronization object are destroyed. If :py:obj:`~.cudaExternalSemaphoreHandleDesc.type` is :py:obj:`~.cudaExternalSemaphoreHandleTypeD3D12Fence`, then exactly one of :py:obj:`~.cudaExternalSemaphoreHandleDesc`::handle::win32::handle and :py:obj:`~.cudaExternalSemaphoreHandleDesc`::handle::win32::name must not be NULL. If :py:obj:`~.cudaExternalSemaphoreHandleDesc`::handle::win32::handle is not NULL, then it must represent a valid shared NT handle that is returned by ID3D12Device::CreateSharedHandle when referring to a ID3D12Fence object. This handle holds a reference to the underlying object. If :py:obj:`~.cudaExternalSemaphoreHandleDesc`::handle::win32::name is not NULL, then it must name a valid synchronization object that refers to a valid ID3D12Fence object. If :py:obj:`~.cudaExternalSemaphoreHandleDesc.type` is :py:obj:`~.cudaExternalSemaphoreHandleTypeD3D11Fence`, then exactly one of :py:obj:`~.cudaExternalSemaphoreHandleDesc`::handle::win32::handle and :py:obj:`~.cudaExternalSemaphoreHandleDesc`::handle::win32::name must not be NULL. If :py:obj:`~.cudaExternalSemaphoreHandleDesc`::handle::win32::handle is not NULL, then it must represent a valid shared NT handle that is returned by ID3D11Fence::CreateSharedHandle. If :py:obj:`~.cudaExternalSemaphoreHandleDesc`::handle::win32::name is not NULL, then it must name a valid synchronization object that refers to a valid ID3D11Fence object. If :py:obj:`~.cudaExternalSemaphoreHandleDesc.type` is :py:obj:`~.cudaExternalSemaphoreHandleTypeNvSciSync`, then :py:obj:`~.cudaExternalSemaphoreHandleDesc`::handle::nvSciSyncObj represents a valid NvSciSyncObj. :py:obj:`~.cudaExternalSemaphoreHandleTypeKeyedMutex`, then exactly one of :py:obj:`~.cudaExternalSemaphoreHandleDesc`::handle::win32::handle and :py:obj:`~.cudaExternalSemaphoreHandleDesc`::handle::win32::name must not be NULL. If :py:obj:`~.cudaExternalSemaphoreHandleDesc`::handle::win32::handle is not NULL, then it represent a valid shared NT handle that is returned by IDXGIResource1::CreateSharedHandle when referring to a IDXGIKeyedMutex object. If :py:obj:`~.cudaExternalSemaphoreHandleDesc.type` is :py:obj:`~.cudaExternalSemaphoreHandleTypeKeyedMutexKmt`, then :py:obj:`~.cudaExternalSemaphoreHandleDesc`::handle::win32::handle must be non-NULL and :py:obj:`~.cudaExternalSemaphoreHandleDesc`::handle::win32::name must be NULL. The handle specified must represent a valid KMT handle that is returned by IDXGIResource::GetSharedHandle when referring to a IDXGIKeyedMutex object. If :py:obj:`~.cudaExternalSemaphoreHandleDesc.type` is :py:obj:`~.cudaExternalSemaphoreHandleTypeTimelineSemaphoreFd`, then :py:obj:`~.cudaExternalSemaphoreHandleDesc`::handle::fd must be a valid file descriptor referencing a synchronization object. Ownership of the file descriptor is transferred to the CUDA driver when the handle is imported successfully. Performing any operations on the file descriptor after it is imported results in undefined behavior. If :py:obj:`~.cudaExternalSemaphoreHandleDesc.type` is :py:obj:`~.cudaExternalSemaphoreHandleTypeTimelineSemaphoreWin32`, then exactly one of :py:obj:`~.cudaExternalSemaphoreHandleDesc`::handle::win32::handle and :py:obj:`~.cudaExternalSemaphoreHandleDesc`::handle::win32::name must not be NULL. If :py:obj:`~.cudaExternalSemaphoreHandleDesc`::handle::win32::handle is not NULL, then it must represent a valid shared NT handle that references a synchronization object. Ownership of this handle is not transferred to CUDA after the import operation, so the application must release the handle using the appropriate system call. If :py:obj:`~.cudaExternalSemaphoreHandleDesc`::handle::win32::name is not NULL, then it must name a valid synchronization object. Parameters ---------- semHandleDesc : :py:obj:`~.cudaExternalSemaphoreHandleDesc` Semaphore import handle descriptor Returns ------- cudaError_t :py:obj:`~.cudaSuccess`, :py:obj:`~.cudaErrorInvalidResourceHandle`, :py:obj:`~.cudaErrorOperatingSystem` extSem_out : :py:obj:`~.cudaExternalSemaphore_t` Returned handle to an external semaphore See Also -------- :py:obj:`~.cudaDestroyExternalSemaphore`, :py:obj:`~.cudaSignalExternalSemaphoresAsync`, :py:obj:`~.cudaWaitExternalSemaphoresAsync` """ cdef cudaExternalSemaphore_t extSem_out = cudaExternalSemaphore_t() cdef cyruntime.cudaExternalSemaphoreHandleDesc* cysemHandleDesc_ptr = semHandleDesc._pvt_ptr if semHandleDesc is not None else NULL with nogil: err = cyruntime.cudaImportExternalSemaphore(extSem_out._pvt_ptr, cysemHandleDesc_ptr) if err != cyruntime.cudaSuccess: return (_dict_cudaError_t[err], None) return (_dict_cudaError_t[err], extSem_out) @cython.embedsignature(True) def cudaSignalExternalSemaphoresAsync(extSemArray : Optional[tuple[cudaExternalSemaphore_t] | list[cudaExternalSemaphore_t]], paramsArray : Optional[tuple[cudaExternalSemaphoreSignalParams] | list[cudaExternalSemaphoreSignalParams]], unsigned int numExtSems, stream): """ Signals a set of external semaphore objects. Enqueues a signal operation on a set of externally allocated semaphore object in the specified stream. The operations will be executed when all prior operations in the stream complete. The exact semantics of signaling a semaphore depends on the type of the object. If the semaphore object is any one of the following types: :py:obj:`~.cudaExternalSemaphoreHandleTypeOpaqueFd`, :py:obj:`~.cudaExternalSemaphoreHandleTypeOpaqueWin32`, :py:obj:`~.cudaExternalSemaphoreHandleTypeOpaqueWin32Kmt` then signaling the semaphore will set it to the signaled state. If the semaphore object is any one of the following types: :py:obj:`~.cudaExternalSemaphoreHandleTypeD3D12Fence`, :py:obj:`~.cudaExternalSemaphoreHandleTypeD3D11Fence`, :py:obj:`~.cudaExternalSemaphoreHandleTypeTimelineSemaphoreFd`, :py:obj:`~.cudaExternalSemaphoreHandleTypeTimelineSemaphoreWin32` then the semaphore will be set to the value specified in :py:obj:`~.cudaExternalSemaphoreSignalParams`::params::fence::value. If the semaphore object is of the type :py:obj:`~.cudaExternalSemaphoreHandleTypeNvSciSync` this API sets :py:obj:`~.cudaExternalSemaphoreSignalParams`::params::nvSciSync::fence to a value that can be used by subsequent waiters of the same NvSciSync object to order operations with those currently submitted in `stream`. Such an update will overwrite previous contents of :py:obj:`~.cudaExternalSemaphoreSignalParams`::params::nvSciSync::fence. By default, signaling such an external semaphore object causes appropriate memory synchronization operations to be performed over all the external memory objects that are imported as :py:obj:`~.cudaExternalMemoryHandleTypeNvSciBuf`. This ensures that any subsequent accesses made by other importers of the same set of NvSciBuf memory object(s) are coherent. These operations can be skipped by specifying the flag :py:obj:`~.cudaExternalSemaphoreSignalSkipNvSciBufMemSync`, which can be used as a performance optimization when data coherency is not required. But specifying this flag in scenarios where data coherency is required results in undefined behavior. Also, for semaphore object of the type :py:obj:`~.cudaExternalSemaphoreHandleTypeNvSciSync`, if the NvSciSyncAttrList used to create the NvSciSyncObj had not set the flags in :py:obj:`~.cudaDeviceGetNvSciSyncAttributes` to cudaNvSciSyncAttrSignal, this API will return cudaErrorNotSupported. :py:obj:`~.cudaExternalSemaphoreSignalParams`::params::nvSciSync::fence associated with semaphore object of the type :py:obj:`~.cudaExternalSemaphoreHandleTypeNvSciSync` can be deterministic. For this the NvSciSyncAttrList used to create the semaphore object must have value of NvSciSyncAttrKey_RequireDeterministicFences key set to true. Deterministic fences allow users to enqueue a wait over the semaphore object even before corresponding signal is enqueued. For such a semaphore object, CUDA guarantees that each signal operation will increment the fence value by '1'. Users are expected to track count of signals enqueued on the semaphore object and insert waits accordingly. When such a semaphore object is signaled from multiple streams, due to concurrent stream execution, it is possible that the order in which the semaphore gets signaled is indeterministic. This could lead to waiters of the semaphore getting unblocked incorrectly. Users are expected to handle such situations, either by not using the same semaphore object with deterministic fence support enabled in different streams or by adding explicit dependency amongst such streams so that the semaphore is signaled in order. If the semaphore object is any one of the following types: :py:obj:`~.cudaExternalSemaphoreHandleTypeKeyedMutex`, :py:obj:`~.cudaExternalSemaphoreHandleTypeKeyedMutexKmt`, then the keyed mutex will be released with the key specified in :py:obj:`~.cudaExternalSemaphoreSignalParams`::params::keyedmutex::key. Parameters ---------- extSemArray : list[:py:obj:`~.cudaExternalSemaphore_t`] Set of external semaphores to be signaled paramsArray : list[:py:obj:`~.cudaExternalSemaphoreSignalParams`] Array of semaphore parameters numExtSems : unsigned int Number of semaphores to signal stream : :py:obj:`~.CUstream` or :py:obj:`~.cudaStream_t` Stream to enqueue the signal operations in Returns ------- cudaError_t :py:obj:`~.cudaSuccess`, :py:obj:`~.cudaErrorInvalidResourceHandle` See Also -------- :py:obj:`~.cudaImportExternalSemaphore`, :py:obj:`~.cudaDestroyExternalSemaphore`, :py:obj:`~.cudaWaitExternalSemaphoresAsync` """ cdef cyruntime.cudaStream_t cystream if stream is None: pstream = 0 elif isinstance(stream, (cudaStream_t,driver.CUstream)): pstream = int(stream) else: pstream = int(cudaStream_t(stream)) cystream = pstream paramsArray = [] if paramsArray is None else paramsArray if not all(isinstance(_x, (cudaExternalSemaphoreSignalParams,)) for _x in paramsArray): raise TypeError("Argument 'paramsArray' is not instance of type (expected tuple[cyruntime.cudaExternalSemaphoreSignalParams,] or list[cyruntime.cudaExternalSemaphoreSignalParams,]") extSemArray = [] if extSemArray is None else extSemArray if not all(isinstance(_x, (cudaExternalSemaphore_t,)) for _x in extSemArray): raise TypeError("Argument 'extSemArray' is not instance of type (expected tuple[cyruntime.cudaExternalSemaphore_t,] or list[cyruntime.cudaExternalSemaphore_t,]") cdef cyruntime.cudaExternalSemaphore_t* cyextSemArray = NULL if len(extSemArray) > 1: cyextSemArray = calloc(len(extSemArray), sizeof(cyruntime.cudaExternalSemaphore_t)) if cyextSemArray is NULL: raise MemoryError('Failed to allocate length x size memory: ' + str(len(extSemArray)) + 'x' + str(sizeof(cyruntime.cudaExternalSemaphore_t))) else: for idx in range(len(extSemArray)): cyextSemArray[idx] = (extSemArray[idx])._pvt_ptr[0] elif len(extSemArray) == 1: cyextSemArray = (extSemArray[0])._pvt_ptr cdef cyruntime.cudaExternalSemaphoreSignalParams* cyparamsArray = NULL if len(paramsArray) > 1: cyparamsArray = calloc(len(paramsArray), sizeof(cyruntime.cudaExternalSemaphoreSignalParams)) if cyparamsArray is NULL: raise MemoryError('Failed to allocate length x size memory: ' + str(len(paramsArray)) + 'x' + str(sizeof(cyruntime.cudaExternalSemaphoreSignalParams))) for idx in range(len(paramsArray)): string.memcpy(&cyparamsArray[idx], (paramsArray[idx])._pvt_ptr, sizeof(cyruntime.cudaExternalSemaphoreSignalParams)) elif len(paramsArray) == 1: cyparamsArray = (paramsArray[0])._pvt_ptr if numExtSems > len(extSemArray): raise RuntimeError("List is too small: " + str(len(extSemArray)) + " < " + str(numExtSems)) if numExtSems > len(paramsArray): raise RuntimeError("List is too small: " + str(len(paramsArray)) + " < " + str(numExtSems)) with nogil: err = cyruntime.cudaSignalExternalSemaphoresAsync(cyextSemArray, cyparamsArray, numExtSems, cystream) if len(extSemArray) > 1 and cyextSemArray is not NULL: free(cyextSemArray) if len(paramsArray) > 1 and cyparamsArray is not NULL: free(cyparamsArray) return (_dict_cudaError_t[err],) @cython.embedsignature(True) def cudaWaitExternalSemaphoresAsync(extSemArray : Optional[tuple[cudaExternalSemaphore_t] | list[cudaExternalSemaphore_t]], paramsArray : Optional[tuple[cudaExternalSemaphoreWaitParams] | list[cudaExternalSemaphoreWaitParams]], unsigned int numExtSems, stream): """ Waits on a set of external semaphore objects. Enqueues a wait operation on a set of externally allocated semaphore object in the specified stream. The operations will be executed when all prior operations in the stream complete. The exact semantics of waiting on a semaphore depends on the type of the object. If the semaphore object is any one of the following types: :py:obj:`~.cudaExternalSemaphoreHandleTypeOpaqueFd`, :py:obj:`~.cudaExternalSemaphoreHandleTypeOpaqueWin32`, :py:obj:`~.cudaExternalSemaphoreHandleTypeOpaqueWin32Kmt` then waiting on the semaphore will wait until the semaphore reaches the signaled state. The semaphore will then be reset to the unsignaled state. Therefore for every signal operation, there can only be one wait operation. If the semaphore object is any one of the following types: :py:obj:`~.cudaExternalSemaphoreHandleTypeD3D12Fence`, :py:obj:`~.cudaExternalSemaphoreHandleTypeD3D11Fence`, :py:obj:`~.cudaExternalSemaphoreHandleTypeTimelineSemaphoreFd`, :py:obj:`~.cudaExternalSemaphoreHandleTypeTimelineSemaphoreWin32` then waiting on the semaphore will wait until the value of the semaphore is greater than or equal to :py:obj:`~.cudaExternalSemaphoreWaitParams`::params::fence::value. If the semaphore object is of the type :py:obj:`~.cudaExternalSemaphoreHandleTypeNvSciSync` then, waiting on the semaphore will wait until the :py:obj:`~.cudaExternalSemaphoreSignalParams`::params::nvSciSync::fence is signaled by the signaler of the NvSciSyncObj that was associated with this semaphore object. By default, waiting on such an external semaphore object causes appropriate memory synchronization operations to be performed over all external memory objects that are imported as :py:obj:`~.cudaExternalMemoryHandleTypeNvSciBuf`. This ensures that any subsequent accesses made by other importers of the same set of NvSciBuf memory object(s) are coherent. These operations can be skipped by specifying the flag :py:obj:`~.cudaExternalSemaphoreWaitSkipNvSciBufMemSync`, which can be used as a performance optimization when data coherency is not required. But specifying this flag in scenarios where data coherency is required results in undefined behavior. Also, for semaphore object of the type :py:obj:`~.cudaExternalSemaphoreHandleTypeNvSciSync`, if the NvSciSyncAttrList used to create the NvSciSyncObj had not set the flags in :py:obj:`~.cudaDeviceGetNvSciSyncAttributes` to cudaNvSciSyncAttrWait, this API will return cudaErrorNotSupported. If the semaphore object is any one of the following types: :py:obj:`~.cudaExternalSemaphoreHandleTypeKeyedMutex`, :py:obj:`~.cudaExternalSemaphoreHandleTypeKeyedMutexKmt`, then the keyed mutex will be acquired when it is released with the key specified in :py:obj:`~.cudaExternalSemaphoreSignalParams`::params::keyedmutex::key or until the timeout specified by :py:obj:`~.cudaExternalSemaphoreSignalParams`::params::keyedmutex::timeoutMs has lapsed. The timeout interval can either be a finite value specified in milliseconds or an infinite value. In case an infinite value is specified the timeout never elapses. The windows INFINITE macro must be used to specify infinite timeout Parameters ---------- extSemArray : list[:py:obj:`~.cudaExternalSemaphore_t`] External semaphores to be waited on paramsArray : list[:py:obj:`~.cudaExternalSemaphoreWaitParams`] Array of semaphore parameters numExtSems : unsigned int Number of semaphores to wait on stream : :py:obj:`~.CUstream` or :py:obj:`~.cudaStream_t` Stream to enqueue the wait operations in Returns ------- cudaError_t :py:obj:`~.cudaSuccess`, :py:obj:`~.cudaErrorInvalidResourceHandle` :py:obj:`~.cudaErrorTimeout` See Also -------- :py:obj:`~.cudaImportExternalSemaphore`, :py:obj:`~.cudaDestroyExternalSemaphore`, :py:obj:`~.cudaSignalExternalSemaphoresAsync` """ cdef cyruntime.cudaStream_t cystream if stream is None: pstream = 0 elif isinstance(stream, (cudaStream_t,driver.CUstream)): pstream = int(stream) else: pstream = int(cudaStream_t(stream)) cystream = pstream paramsArray = [] if paramsArray is None else paramsArray if not all(isinstance(_x, (cudaExternalSemaphoreWaitParams,)) for _x in paramsArray): raise TypeError("Argument 'paramsArray' is not instance of type (expected tuple[cyruntime.cudaExternalSemaphoreWaitParams,] or list[cyruntime.cudaExternalSemaphoreWaitParams,]") extSemArray = [] if extSemArray is None else extSemArray if not all(isinstance(_x, (cudaExternalSemaphore_t,)) for _x in extSemArray): raise TypeError("Argument 'extSemArray' is not instance of type (expected tuple[cyruntime.cudaExternalSemaphore_t,] or list[cyruntime.cudaExternalSemaphore_t,]") cdef cyruntime.cudaExternalSemaphore_t* cyextSemArray = NULL if len(extSemArray) > 1: cyextSemArray = calloc(len(extSemArray), sizeof(cyruntime.cudaExternalSemaphore_t)) if cyextSemArray is NULL: raise MemoryError('Failed to allocate length x size memory: ' + str(len(extSemArray)) + 'x' + str(sizeof(cyruntime.cudaExternalSemaphore_t))) else: for idx in range(len(extSemArray)): cyextSemArray[idx] = (extSemArray[idx])._pvt_ptr[0] elif len(extSemArray) == 1: cyextSemArray = (extSemArray[0])._pvt_ptr cdef cyruntime.cudaExternalSemaphoreWaitParams* cyparamsArray = NULL if len(paramsArray) > 1: cyparamsArray = calloc(len(paramsArray), sizeof(cyruntime.cudaExternalSemaphoreWaitParams)) if cyparamsArray is NULL: raise MemoryError('Failed to allocate length x size memory: ' + str(len(paramsArray)) + 'x' + str(sizeof(cyruntime.cudaExternalSemaphoreWaitParams))) for idx in range(len(paramsArray)): string.memcpy(&cyparamsArray[idx], (paramsArray[idx])._pvt_ptr, sizeof(cyruntime.cudaExternalSemaphoreWaitParams)) elif len(paramsArray) == 1: cyparamsArray = (paramsArray[0])._pvt_ptr if numExtSems > len(extSemArray): raise RuntimeError("List is too small: " + str(len(extSemArray)) + " < " + str(numExtSems)) if numExtSems > len(paramsArray): raise RuntimeError("List is too small: " + str(len(paramsArray)) + " < " + str(numExtSems)) with nogil: err = cyruntime.cudaWaitExternalSemaphoresAsync(cyextSemArray, cyparamsArray, numExtSems, cystream) if len(extSemArray) > 1 and cyextSemArray is not NULL: free(cyextSemArray) if len(paramsArray) > 1 and cyparamsArray is not NULL: free(cyparamsArray) return (_dict_cudaError_t[err],) @cython.embedsignature(True) def cudaDestroyExternalSemaphore(extSem): """ Destroys an external semaphore. Destroys an external semaphore object and releases any references to the underlying resource. Any outstanding signals or waits must have completed before the semaphore is destroyed. Parameters ---------- extSem : :py:obj:`~.cudaExternalSemaphore_t` External semaphore to be destroyed Returns ------- cudaError_t :py:obj:`~.cudaSuccess`, :py:obj:`~.cudaErrorInvalidResourceHandle` See Also -------- :py:obj:`~.cudaImportExternalSemaphore`, :py:obj:`~.cudaSignalExternalSemaphoresAsync`, :py:obj:`~.cudaWaitExternalSemaphoresAsync` """ cdef cyruntime.cudaExternalSemaphore_t cyextSem if extSem is None: pextSem = 0 elif isinstance(extSem, (cudaExternalSemaphore_t,)): pextSem = int(extSem) else: pextSem = int(cudaExternalSemaphore_t(extSem)) cyextSem = pextSem with nogil: err = cyruntime.cudaDestroyExternalSemaphore(cyextSem) return (_dict_cudaError_t[err],) @cython.embedsignature(True) def cudaFuncSetCacheConfig(func, cacheConfig not None : cudaFuncCache): """ Sets the preferred cache configuration for a device function. On devices where the L1 cache and shared memory use the same hardware resources, this sets through `cacheConfig` the preferred cache configuration for the function specified via `func`. This is only a preference. The runtime will use the requested configuration if possible, but it is free to choose a different configuration if required to execute `func`. `func` is a device function symbol and must be declared as a `None` function. If the specified function does not exist, then :py:obj:`~.cudaErrorInvalidDeviceFunction` is returned. For templated functions, pass the function symbol as follows: func_name This setting does nothing on devices where the size of the L1 cache and shared memory are fixed. Launching a kernel with a different preference than the most recent preference setting may insert a device-side synchronization point. The supported cache configurations are: - :py:obj:`~.cudaFuncCachePreferNone`: no preference for shared memory or L1 (default) - :py:obj:`~.cudaFuncCachePreferShared`: prefer larger shared memory and smaller L1 cache - :py:obj:`~.cudaFuncCachePreferL1`: prefer larger L1 cache and smaller shared memory - :py:obj:`~.cudaFuncCachePreferEqual`: prefer equal size L1 cache and shared memory Parameters ---------- func : Any Device function symbol cacheConfig : :py:obj:`~.cudaFuncCache` Requested cache configuration Returns ------- cudaError_t :py:obj:`~.cudaSuccess`, :py:obj:`~.cudaErrorInvalidDeviceFunction`2 See Also -------- cudaFuncSetCacheConfig (C++ API), :py:obj:`~.cudaFuncGetAttributes (C API)`, :py:obj:`~.cudaLaunchKernel (C API)`, :py:obj:`~.cuFuncSetCacheConfig` Notes ----- This API does not accept a :py:obj:`~.cudaKernel_t` casted as void*. If cache config modification is required for a :py:obj:`~.cudaKernel_t` (or a global function), it can be replaced with a call to :py:obj:`~.cudaFuncSetAttributes` with the attribute :py:obj:`~.cudaFuncAttributePreferredSharedMemoryCarveout` to specify a more granular L1 cache and shared memory split configuration. """ cyfunc = _HelperInputVoidPtr(func) cdef void* cyfunc_ptr = cyfunc.cptr cdef cyruntime.cudaFuncCache cycacheConfig = cacheConfig.value with nogil: err = cyruntime.cudaFuncSetCacheConfig(cyfunc_ptr, cycacheConfig) return (_dict_cudaError_t[err],) @cython.embedsignature(True) def cudaFuncGetAttributes(func): """ Find out attributes for a given function. This function obtains the attributes of a function specified via `func`. `func` is a device function symbol and must be declared as a `None` function. The fetched attributes are placed in `attr`. If the specified function does not exist, then it is assumed to be a :py:obj:`~.cudaKernel_t` and used as is. For templated functions, pass the function symbol as follows: func_name Note that some function attributes such as :py:obj:`~.maxThreadsPerBlock` may vary based on the device that is currently being used. Parameters ---------- func : Any Device function symbol Returns ------- cudaError_t :py:obj:`~.cudaSuccess`, :py:obj:`~.cudaErrorInvalidDeviceFunction`2 attr : :py:obj:`~.cudaFuncAttributes` Return pointer to function's attributes See Also -------- :py:obj:`~.cudaFuncSetCacheConfig (C API)`, cudaFuncGetAttributes (C++ API), :py:obj:`~.cudaLaunchKernel (C API)`, :py:obj:`~.cuFuncGetAttribute` """ cdef cudaFuncAttributes attr = cudaFuncAttributes() cyfunc = _HelperInputVoidPtr(func) cdef void* cyfunc_ptr = cyfunc.cptr with nogil: err = cyruntime.cudaFuncGetAttributes(attr._pvt_ptr, cyfunc_ptr) if err != cyruntime.cudaSuccess: return (_dict_cudaError_t[err], None) return (_dict_cudaError_t[err], attr) @cython.embedsignature(True) def cudaFuncSetAttribute(func, attr not None : cudaFuncAttribute, int value): """ Set attributes for a given function. This function sets the attributes of a function specified via `func`. The parameter `func` must be a pointer to a function that executes on the device. The parameter specified by `func` must be declared as a `None` function. The enumeration defined by `attr` is set to the value defined by `value`. If the specified function does not exist, then it is assumed to be a :py:obj:`~.cudaKernel_t` and used as is. If the specified attribute cannot be written, or if the value is incorrect, then :py:obj:`~.cudaErrorInvalidValue` is returned. Valid values for `attr` are: - :py:obj:`~.cudaFuncAttributeMaxDynamicSharedMemorySize` - The requested maximum size in bytes of dynamically-allocated shared memory. The sum of this value and the function attribute :py:obj:`~.sharedSizeBytes` cannot exceed the device attribute :py:obj:`~.cudaDevAttrMaxSharedMemoryPerBlockOptin`. The maximal size of requestable dynamic shared memory may differ by GPU architecture. - :py:obj:`~.cudaFuncAttributePreferredSharedMemoryCarveout` - On devices where the L1 cache and shared memory use the same hardware resources, this sets the shared memory carveout preference, in percent of the total shared memory. See :py:obj:`~.cudaDevAttrMaxSharedMemoryPerMultiprocessor`. This is only a hint, and the driver can choose a different ratio if required to execute the function. - :py:obj:`~.cudaFuncAttributeRequiredClusterWidth`: The required cluster width in blocks. The width, height, and depth values must either all be 0 or all be positive. The validity of the cluster dimensions is checked at launch time. If the value is set during compile time, it cannot be set at runtime. Setting it at runtime will return cudaErrorNotPermitted. - :py:obj:`~.cudaFuncAttributeRequiredClusterHeight`: The required cluster height in blocks. The width, height, and depth values must either all be 0 or all be positive. The validity of the cluster dimensions is checked at launch time. If the value is set during compile time, it cannot be set at runtime. Setting it at runtime will return cudaErrorNotPermitted. - :py:obj:`~.cudaFuncAttributeRequiredClusterDepth`: The required cluster depth in blocks. The width, height, and depth values must either all be 0 or all be positive. The validity of the cluster dimensions is checked at launch time. If the value is set during compile time, it cannot be set at runtime. Setting it at runtime will return cudaErrorNotPermitted. - :py:obj:`~.cudaFuncAttributeNonPortableClusterSizeAllowed`: Indicates whether the function can be launched with non-portable cluster size. 1 is allowed, 0 is disallowed. - :py:obj:`~.cudaFuncAttributeClusterSchedulingPolicyPreference`: The block scheduling policy of a function. The value type is cudaClusterSchedulingPolicy. cudaLaunchKernel (C++ API), cudaFuncSetCacheConfig (C++ API), :py:obj:`~.cudaFuncGetAttributes (C API)`, Parameters ---------- func : Any Function to get attributes of attr : :py:obj:`~.cudaFuncAttribute` Attribute to set value : int Value to set Returns ------- cudaError_t :py:obj:`~.cudaSuccess`, :py:obj:`~.cudaErrorInvalidDeviceFunction`, :py:obj:`~.cudaErrorInvalidValue` """ cyfunc = _HelperInputVoidPtr(func) cdef void* cyfunc_ptr = cyfunc.cptr cdef cyruntime.cudaFuncAttribute cyattr = attr.value with nogil: err = cyruntime.cudaFuncSetAttribute(cyfunc_ptr, cyattr, value) return (_dict_cudaError_t[err],) ctypedef struct cudaStreamHostCallbackData_st: cyruntime.cudaHostFn_t callback void *userData ctypedef cudaStreamHostCallbackData_st cudaStreamHostCallbackData @cython.show_performance_hints(False) cdef void cudaStreamRtHostCallbackWrapper(void *data) nogil: cdef cudaStreamHostCallbackData *cbData = data with gil: cbData.callback(cbData.userData) free(cbData) @cython.embedsignature(True) def cudaLaunchHostFunc(stream, fn, userData): """ Enqueues a host function call in a stream. Enqueues a host function to run in a stream. The function will be called after currently enqueued work and will block work added after it. The host function must not make any CUDA API calls. Attempting to use a CUDA API may result in :py:obj:`~.cudaErrorNotPermitted`, but this is not required. The host function must not perform any synchronization that may depend on outstanding CUDA work not mandated to run earlier. Host functions without a mandated order (such as in independent streams) execute in undefined order and may be serialized. For the purposes of Unified Memory, execution makes a number of guarantees: - The stream is considered idle for the duration of the function's execution. Thus, for example, the function may always use memory attached to the stream it was enqueued in. - The start of execution of the function has the same effect as synchronizing an event recorded in the same stream immediately prior to the function. It thus synchronizes streams which have been "joined" prior to the function. - Adding device work to any stream does not have the effect of making the stream active until all preceding host functions and stream callbacks have executed. Thus, for example, a function might use global attached memory even if work has been added to another stream, if the work has been ordered behind the function call with an event. - Completion of the function does not cause a stream to become active except as described above. The stream will remain idle if no device work follows the function, and will remain idle across consecutive host functions or stream callbacks without device work in between. Thus, for example, stream synchronization can be done by signaling from a host function at the end of the stream. Note that, in constrast to :py:obj:`~.cuStreamAddCallback`, the function will not be called in the event of an error in the CUDA context. Parameters ---------- hStream : :py:obj:`~.CUstream` or :py:obj:`~.cudaStream_t` Stream to enqueue function call in fn : :py:obj:`~.cudaHostFn_t` The function to call once preceding stream operations are complete userData : Any User-specified data to be passed to the function Returns ------- cudaError_t :py:obj:`~.cudaSuccess`, :py:obj:`~.cudaErrorInvalidResourceHandle`, :py:obj:`~.cudaErrorInvalidValue`, :py:obj:`~.cudaErrorNotSupported` See Also -------- :py:obj:`~.cudaStreamCreate`, :py:obj:`~.cudaStreamQuery`, :py:obj:`~.cudaStreamSynchronize`, :py:obj:`~.cudaStreamWaitEvent`, :py:obj:`~.cudaStreamDestroy`, :py:obj:`~.cudaMallocManaged`, :py:obj:`~.cudaStreamAttachMemAsync`, :py:obj:`~.cudaStreamAddCallback`, :py:obj:`~.cuLaunchHostFunc` """ cdef cyruntime.cudaHostFn_t cyfn if fn is None: pfn = 0 elif isinstance(fn, (cudaHostFn_t,)): pfn = int(fn) else: pfn = int(cudaHostFn_t(fn)) cyfn = pfn cdef cyruntime.cudaStream_t cystream if stream is None: pstream = 0 elif isinstance(stream, (cudaStream_t,driver.CUstream)): pstream = int(stream) else: pstream = int(cudaStream_t(stream)) cystream = pstream cyuserData = _HelperInputVoidPtr(userData) cdef void* cyuserData_ptr = cyuserData.cptr cdef cudaStreamHostCallbackData *cbData = NULL cbData = malloc(sizeof(cbData[0])) if cbData == NULL: return (cudaError_t.cudaErrorMemoryAllocation,) cbData.callback = cyfn cbData.userData = cyuserData_ptr with nogil: err = cyruntime.cudaLaunchHostFunc(cystream, cudaStreamRtHostCallbackWrapper, cbData) if err != cyruntime.cudaSuccess: free(cbData) return (_dict_cudaError_t[err],) @cython.embedsignature(True) def cudaFuncSetSharedMemConfig(func, config not None : cudaSharedMemConfig): """ Sets the shared memory configuration for a device function. [Deprecated] On devices with configurable shared memory banks, this function will force all subsequent launches of the specified device function to have the given shared memory bank size configuration. On any given launch of the function, the shared memory configuration of the device will be temporarily changed if needed to suit the function's preferred configuration. Changes in shared memory configuration between subsequent launches of functions, may introduce a device side synchronization point. Any per-function setting of shared memory bank size set via :py:obj:`~.cudaFuncSetSharedMemConfig` will override the device wide setting set by :py:obj:`~.cudaDeviceSetSharedMemConfig`. Changing the shared memory bank size will not increase shared memory usage or affect occupancy of kernels, but may have major effects on performance. Larger bank sizes will allow for greater potential bandwidth to shared memory, but will change what kinds of accesses to shared memory will result in bank conflicts. This function will do nothing on devices with fixed shared memory bank size. For templated functions, pass the function symbol as follows: func_name The supported bank configurations are: - :py:obj:`~.cudaSharedMemBankSizeDefault`: use the device's shared memory configuration when launching this function. - :py:obj:`~.cudaSharedMemBankSizeFourByte`: set shared memory bank width to be four bytes natively when launching this function. - :py:obj:`~.cudaSharedMemBankSizeEightByte`: set shared memory bank width to be eight bytes natively when launching this function. Parameters ---------- func : Any Device function symbol config : :py:obj:`~.cudaSharedMemConfig` Requested shared memory configuration Returns ------- cudaError_t :py:obj:`~.cudaSuccess`, :py:obj:`~.cudaErrorInvalidDeviceFunction`, :py:obj:`~.cudaErrorInvalidValue`,2 See Also -------- :py:obj:`~.cudaDeviceSetSharedMemConfig`, :py:obj:`~.cudaDeviceGetSharedMemConfig`, :py:obj:`~.cudaDeviceSetCacheConfig`, :py:obj:`~.cudaDeviceGetCacheConfig`, :py:obj:`~.cudaFuncSetCacheConfig`, :py:obj:`~.cuFuncSetSharedMemConfig` """ cyfunc = _HelperInputVoidPtr(func) cdef void* cyfunc_ptr = cyfunc.cptr cdef cyruntime.cudaSharedMemConfig cyconfig = config.value with nogil: err = cyruntime.cudaFuncSetSharedMemConfig(cyfunc_ptr, cyconfig) return (_dict_cudaError_t[err],) @cython.embedsignature(True) def cudaOccupancyMaxActiveBlocksPerMultiprocessor(func, int blockSize, size_t dynamicSMemSize): """ Returns occupancy for a device function. Returns in `*numBlocks` the maximum number of active blocks per streaming multiprocessor for the device function. Parameters ---------- func : Any Kernel function for which occupancy is calculated blockSize : int Block size the kernel is intended to be launched with dynamicSMemSize : size_t Per-block dynamic shared memory usage intended, in bytes Returns ------- cudaError_t :py:obj:`~.cudaSuccess`, :py:obj:`~.cudaErrorInvalidDevice`, :py:obj:`~.cudaErrorInvalidDeviceFunction`, :py:obj:`~.cudaErrorInvalidValue`, :py:obj:`~.cudaErrorUnknown`, numBlocks : int Returned occupancy See Also -------- :py:obj:`~.cudaOccupancyMaxActiveBlocksPerMultiprocessorWithFlags`, cudaOccupancyMaxPotentialBlockSize (C++ API), cudaOccupancyMaxPotentialBlockSizeWithFlags (C++ API), cudaOccupancyMaxPotentialBlockSizeVariableSMem (C++ API), cudaOccupancyMaxPotentialBlockSizeVariableSMemWithFlags (C++ API), cudaOccupancyAvailableDynamicSMemPerBlock (C++ API), :py:obj:`~.cuOccupancyMaxActiveBlocksPerMultiprocessor` """ cdef int numBlocks = 0 cyfunc = _HelperInputVoidPtr(func) cdef void* cyfunc_ptr = cyfunc.cptr with nogil: err = cyruntime.cudaOccupancyMaxActiveBlocksPerMultiprocessor(&numBlocks, cyfunc_ptr, blockSize, dynamicSMemSize) if err != cyruntime.cudaSuccess: return (_dict_cudaError_t[err], None) return (_dict_cudaError_t[err], numBlocks) @cython.embedsignature(True) def cudaOccupancyAvailableDynamicSMemPerBlock(func, int numBlocks, int blockSize): """ Returns dynamic shared memory available per block when launching `numBlocks` blocks on SM. Returns in `*dynamicSmemSize` the maximum size of dynamic shared memory to allow `numBlocks` blocks per SM. Parameters ---------- func : Any Kernel function for which occupancy is calculated numBlocks : int Number of blocks to fit on SM blockSize : int Size of the block Returns ------- cudaError_t :py:obj:`~.cudaSuccess`, :py:obj:`~.cudaErrorInvalidDevice`, :py:obj:`~.cudaErrorInvalidDeviceFunction`, :py:obj:`~.cudaErrorInvalidValue`, :py:obj:`~.cudaErrorUnknown`, dynamicSmemSize : int Returned maximum dynamic shared memory See Also -------- :py:obj:`~.cudaOccupancyMaxActiveBlocksPerMultiprocessorWithFlags`, cudaOccupancyMaxPotentialBlockSize (C++ API), cudaOccupancyMaxPotentialBlockSizeWithFlags (C++ API), cudaOccupancyMaxPotentialBlockSizeVariableSMem (C++ API), cudaOccupancyMaxPotentialBlockSizeVariableSMemWithFlags (C++ API), :py:obj:`~.cudaOccupancyAvailableDynamicSMemPerBlock` """ cdef size_t dynamicSmemSize = 0 cyfunc = _HelperInputVoidPtr(func) cdef void* cyfunc_ptr = cyfunc.cptr with nogil: err = cyruntime.cudaOccupancyAvailableDynamicSMemPerBlock(&dynamicSmemSize, cyfunc_ptr, numBlocks, blockSize) if err != cyruntime.cudaSuccess: return (_dict_cudaError_t[err], None) return (_dict_cudaError_t[err], dynamicSmemSize) @cython.embedsignature(True) def cudaOccupancyMaxActiveBlocksPerMultiprocessorWithFlags(func, int blockSize, size_t dynamicSMemSize, unsigned int flags): """ Returns occupancy for a device function with the specified flags. Returns in `*numBlocks` the maximum number of active blocks per streaming multiprocessor for the device function. The `flags` parameter controls how special cases are handled. Valid flags include: - :py:obj:`~.cudaOccupancyDefault`: keeps the default behavior as :py:obj:`~.cudaOccupancyMaxActiveBlocksPerMultiprocessor` - :py:obj:`~.cudaOccupancyDisableCachingOverride`: This flag suppresses the default behavior on platform where global caching affects occupancy. On such platforms, if caching is enabled, but per-block SM resource usage would result in zero occupancy, the occupancy calculator will calculate the occupancy as if caching is disabled. Setting this flag makes the occupancy calculator to return 0 in such cases. More information can be found about this feature in the "Unified L1/Texture Cache" section of the Maxwell tuning guide. Parameters ---------- func : Any Kernel function for which occupancy is calculated blockSize : int Block size the kernel is intended to be launched with dynamicSMemSize : size_t Per-block dynamic shared memory usage intended, in bytes flags : unsigned int Requested behavior for the occupancy calculator Returns ------- cudaError_t :py:obj:`~.cudaSuccess`, :py:obj:`~.cudaErrorInvalidDevice`, :py:obj:`~.cudaErrorInvalidDeviceFunction`, :py:obj:`~.cudaErrorInvalidValue`, :py:obj:`~.cudaErrorUnknown`, numBlocks : int Returned occupancy See Also -------- :py:obj:`~.cudaOccupancyMaxActiveBlocksPerMultiprocessor`, cudaOccupancyMaxPotentialBlockSize (C++ API), cudaOccupancyMaxPotentialBlockSizeWithFlags (C++ API), cudaOccupancyMaxPotentialBlockSizeVariableSMem (C++ API), cudaOccupancyMaxPotentialBlockSizeVariableSMemWithFlags (C++ API), cudaOccupancyAvailableDynamicSMemPerBlock (C++ API), :py:obj:`~.cuOccupancyMaxActiveBlocksPerMultiprocessorWithFlags` """ cdef int numBlocks = 0 cyfunc = _HelperInputVoidPtr(func) cdef void* cyfunc_ptr = cyfunc.cptr with nogil: err = cyruntime.cudaOccupancyMaxActiveBlocksPerMultiprocessorWithFlags(&numBlocks, cyfunc_ptr, blockSize, dynamicSMemSize, flags) if err != cyruntime.cudaSuccess: return (_dict_cudaError_t[err], None) return (_dict_cudaError_t[err], numBlocks) @cython.embedsignature(True) def cudaMallocManaged(size_t size, unsigned int flags): """ Allocates memory that will be automatically managed by the Unified Memory system. Allocates `size` bytes of managed memory on the device and returns in `*devPtr` a pointer to the allocated memory. If the device doesn't support allocating managed memory, :py:obj:`~.cudaErrorNotSupported` is returned. Support for managed memory can be queried using the device attribute :py:obj:`~.cudaDevAttrManagedMemory`. The allocated memory is suitably aligned for any kind of variable. The memory is not cleared. If `size` is 0, :py:obj:`~.cudaMallocManaged` returns :py:obj:`~.cudaErrorInvalidValue`. The pointer is valid on the CPU and on all GPUs in the system that support managed memory. All accesses to this pointer must obey the Unified Memory programming model. `flags` specifies the default stream association for this allocation. `flags` must be one of :py:obj:`~.cudaMemAttachGlobal` or :py:obj:`~.cudaMemAttachHost`. The default value for `flags` is :py:obj:`~.cudaMemAttachGlobal`. If :py:obj:`~.cudaMemAttachGlobal` is specified, then this memory is accessible from any stream on any device. If :py:obj:`~.cudaMemAttachHost` is specified, then the allocation should not be accessed from devices that have a zero value for the device attribute :py:obj:`~.cudaDevAttrConcurrentManagedAccess`; an explicit call to :py:obj:`~.cudaStreamAttachMemAsync` will be required to enable access on such devices. If the association is later changed via :py:obj:`~.cudaStreamAttachMemAsync` to a single stream, the default association, as specifed during :py:obj:`~.cudaMallocManaged`, is restored when that stream is destroyed. For managed variables, the default association is always :py:obj:`~.cudaMemAttachGlobal`. Note that destroying a stream is an asynchronous operation, and as a result, the change to default association won't happen until all work in the stream has completed. Memory allocated with :py:obj:`~.cudaMallocManaged` should be released with :py:obj:`~.cudaFree`. Device memory oversubscription is possible for GPUs that have a non- zero value for the device attribute :py:obj:`~.cudaDevAttrConcurrentManagedAccess`. Managed memory on such GPUs may be evicted from device memory to host memory at any time by the Unified Memory driver in order to make room for other allocations. In a system where all GPUs have a non-zero value for the device attribute :py:obj:`~.cudaDevAttrConcurrentManagedAccess`, managed memory may not be populated when this API returns and instead may be populated on access. In such systems, managed memory can migrate to any processor's memory at any time. The Unified Memory driver will employ heuristics to maintain data locality and prevent excessive page faults to the extent possible. The application can also guide the driver about memory usage patterns via :py:obj:`~.cudaMemAdvise`. The application can also explicitly migrate memory to a desired processor's memory via :py:obj:`~.cudaMemPrefetchAsync`. In a multi-GPU system where all of the GPUs have a zero value for the device attribute :py:obj:`~.cudaDevAttrConcurrentManagedAccess` and all the GPUs have peer-to-peer support with each other, the physical storage for managed memory is created on the GPU which is active at the time :py:obj:`~.cudaMallocManaged` is called. All other GPUs will reference the data at reduced bandwidth via peer mappings over the PCIe bus. The Unified Memory driver does not migrate memory among such GPUs. In a multi-GPU system where not all GPUs have peer-to-peer support with each other and where the value of the device attribute :py:obj:`~.cudaDevAttrConcurrentManagedAccess` is zero for at least one of those GPUs, the location chosen for physical storage of managed memory is system-dependent. - On Linux, the location chosen will be device memory as long as the current set of active contexts are on devices that either have peer- to-peer support with each other or have a non-zero value for the device attribute :py:obj:`~.cudaDevAttrConcurrentManagedAccess`. If there is an active context on a GPU that does not have a non-zero value for that device attribute and it does not have peer-to-peer support with the other devices that have active contexts on them, then the location for physical storage will be 'zero-copy' or host memory. Note that this means that managed memory that is located in device memory is migrated to host memory if a new context is created on a GPU that doesn't have a non-zero value for the device attribute and does not support peer-to-peer with at least one of the other devices that has an active context. This in turn implies that context creation may fail if there is insufficient host memory to migrate all managed allocations. - On Windows, the physical storage is always created in 'zero-copy' or host memory. All GPUs will reference the data at reduced bandwidth over the PCIe bus. In these circumstances, use of the environment variable CUDA_VISIBLE_DEVICES is recommended to restrict CUDA to only use those GPUs that have peer-to-peer support. Alternatively, users can also set CUDA_MANAGED_FORCE_DEVICE_ALLOC to a non-zero value to force the driver to always use device memory for physical storage. When this environment variable is set to a non-zero value, all devices used in that process that support managed memory have to be peer-to-peer compatible with each other. The error :py:obj:`~.cudaErrorInvalidDevice` will be returned if a device that supports managed memory is used and it is not peer-to-peer compatible with any of the other managed memory supporting devices that were previously used in that process, even if :py:obj:`~.cudaDeviceReset` has been called on those devices. These environment variables are described in the CUDA programming guide under the "CUDA environment variables" section. Parameters ---------- size : size_t Requested allocation size in bytes flags : unsigned int Must be either :py:obj:`~.cudaMemAttachGlobal` or :py:obj:`~.cudaMemAttachHost` (defaults to :py:obj:`~.cudaMemAttachGlobal`) Returns ------- cudaError_t :py:obj:`~.cudaSuccess`, :py:obj:`~.cudaErrorMemoryAllocation`, :py:obj:`~.cudaErrorNotSupported`, :py:obj:`~.cudaErrorInvalidValue` devPtr : Any Pointer to allocated device memory See Also -------- :py:obj:`~.cudaMallocPitch`, :py:obj:`~.cudaFree`, :py:obj:`~.cudaMallocArray`, :py:obj:`~.cudaFreeArray`, :py:obj:`~.cudaMalloc3D`, :py:obj:`~.cudaMalloc3DArray`, :py:obj:`~.cudaMallocHost (C API)`, :py:obj:`~.cudaFreeHost`, :py:obj:`~.cudaHostAlloc`, :py:obj:`~.cudaDeviceGetAttribute`, :py:obj:`~.cudaStreamAttachMemAsync`, :py:obj:`~.cuMemAllocManaged` """ cdef void_ptr devPtr = 0 with nogil: err = cyruntime.cudaMallocManaged(&devPtr, size, flags) if err != cyruntime.cudaSuccess: return (_dict_cudaError_t[err], None) return (_dict_cudaError_t[err], devPtr) @cython.embedsignature(True) def cudaMalloc(size_t size): """ Allocate memory on the device. Allocates `size` bytes of linear memory on the device and returns in `*devPtr` a pointer to the allocated memory. The allocated memory is suitably aligned for any kind of variable. The memory is not cleared. :py:obj:`~.cudaMalloc()` returns :py:obj:`~.cudaErrorMemoryAllocation` in case of failure. The device version of :py:obj:`~.cudaFree` cannot be used with a `*devPtr` allocated using the host API, and vice versa. Parameters ---------- size : size_t Requested allocation size in bytes Returns ------- cudaError_t :py:obj:`~.cudaSuccess`, :py:obj:`~.cudaErrorInvalidValue`, :py:obj:`~.cudaErrorMemoryAllocation` devPtr : Any Pointer to allocated device memory See Also -------- :py:obj:`~.cudaMallocPitch`, :py:obj:`~.cudaFree`, :py:obj:`~.cudaMallocArray`, :py:obj:`~.cudaFreeArray`, :py:obj:`~.cudaMalloc3D`, :py:obj:`~.cudaMalloc3DArray`, :py:obj:`~.cudaMallocHost (C API)`, :py:obj:`~.cudaFreeHost`, :py:obj:`~.cudaHostAlloc`, :py:obj:`~.cuMemAlloc` """ cdef void_ptr devPtr = 0 with nogil: err = cyruntime.cudaMalloc(&devPtr, size) if err != cyruntime.cudaSuccess: return (_dict_cudaError_t[err], None) return (_dict_cudaError_t[err], devPtr) @cython.embedsignature(True) def cudaMallocHost(size_t size): """ Allocates page-locked memory on the host. Allocates `size` bytes of host memory that is page-locked and accessible to the device. The driver tracks the virtual memory ranges allocated with this function and automatically accelerates calls to functions such as :py:obj:`~.cudaMemcpy`*(). Since the memory can be accessed directly by the device, it can be read or written with much higher bandwidth than pageable memory obtained with functions such as :py:obj:`~.malloc()`. On systems where :py:obj:`~.pageableMemoryAccessUsesHostPageTables` is true, :py:obj:`~.cudaMallocHost` may not page-lock the allocated memory. Page-locking excessive amounts of memory with :py:obj:`~.cudaMallocHost()` may degrade system performance, since it reduces the amount of memory available to the system for paging. As a result, this function is best used sparingly to allocate staging areas for data exchange between host and device. Parameters ---------- size : size_t Requested allocation size in bytes Returns ------- cudaError_t :py:obj:`~.cudaSuccess`, :py:obj:`~.cudaErrorInvalidValue`, :py:obj:`~.cudaErrorMemoryAllocation` ptr : Any Pointer to allocated host memory See Also -------- :py:obj:`~.cudaMalloc`, :py:obj:`~.cudaMallocPitch`, :py:obj:`~.cudaMallocArray`, :py:obj:`~.cudaMalloc3D`, :py:obj:`~.cudaMalloc3DArray`, :py:obj:`~.cudaHostAlloc`, :py:obj:`~.cudaFree`, :py:obj:`~.cudaFreeArray`, cudaMallocHost (C++ API), :py:obj:`~.cudaFreeHost`, :py:obj:`~.cudaHostAlloc`, :py:obj:`~.cuMemAllocHost` """ cdef void_ptr ptr = 0 with nogil: err = cyruntime.cudaMallocHost(&ptr, size) if err != cyruntime.cudaSuccess: return (_dict_cudaError_t[err], None) return (_dict_cudaError_t[err], ptr) @cython.embedsignature(True) def cudaMallocPitch(size_t width, size_t height): """ Allocates pitched memory on the device. Allocates at least `width` (in bytes) * `height` bytes of linear memory on the device and returns in `*devPtr` a pointer to the allocated memory. The function may pad the allocation to ensure that corresponding pointers in any given row will continue to meet the alignment requirements for coalescing as the address is updated from row to row. The pitch returned in `*pitch` by :py:obj:`~.cudaMallocPitch()` is the width in bytes of the allocation. The intended usage of `pitch` is as a separate parameter of the allocation, used to compute addresses within the 2D array. Given the row and column of an array element of type `T`, the address is computed as: **View CUDA Toolkit Documentation for a C++ code example** For allocations of 2D arrays, it is recommended that programmers consider performing pitch allocations using :py:obj:`~.cudaMallocPitch()`. Due to pitch alignment restrictions in the hardware, this is especially true if the application will be performing 2D memory copies between different regions of device memory (whether linear memory or CUDA arrays). Parameters ---------- width : size_t Requested pitched allocation width (in bytes) height : size_t Requested pitched allocation height Returns ------- cudaError_t :py:obj:`~.cudaSuccess`, :py:obj:`~.cudaErrorInvalidValue`, :py:obj:`~.cudaErrorMemoryAllocation` devPtr : Any Pointer to allocated pitched device memory pitch : int Pitch for allocation See Also -------- :py:obj:`~.cudaMalloc`, :py:obj:`~.cudaFree`, :py:obj:`~.cudaMallocArray`, :py:obj:`~.cudaFreeArray`, :py:obj:`~.cudaMallocHost (C API)`, :py:obj:`~.cudaFreeHost`, :py:obj:`~.cudaMalloc3D`, :py:obj:`~.cudaMalloc3DArray`, :py:obj:`~.cudaHostAlloc`, :py:obj:`~.cuMemAllocPitch` """ cdef void_ptr devPtr = 0 cdef size_t pitch = 0 with nogil: err = cyruntime.cudaMallocPitch(&devPtr, &pitch, width, height) if err != cyruntime.cudaSuccess: return (_dict_cudaError_t[err], None, None) return (_dict_cudaError_t[err], devPtr, pitch) @cython.embedsignature(True) def cudaMallocArray(desc : Optional[cudaChannelFormatDesc], size_t width, size_t height, unsigned int flags): """ Allocate an array on the device. Allocates a CUDA array according to the :py:obj:`~.cudaChannelFormatDesc` structure `desc` and returns a handle to the new CUDA array in `*array`. The :py:obj:`~.cudaChannelFormatDesc` is defined as: **View CUDA Toolkit Documentation for a C++ code example** where :py:obj:`~.cudaChannelFormatKind` is one of :py:obj:`~.cudaChannelFormatKindSigned`, :py:obj:`~.cudaChannelFormatKindUnsigned`, or :py:obj:`~.cudaChannelFormatKindFloat`. The `flags` parameter enables different options to be specified that affect the allocation, as follows. - :py:obj:`~.cudaArrayDefault`: This flag's value is defined to be 0 and provides default array allocation - :py:obj:`~.cudaArraySurfaceLoadStore`: Allocates an array that can be read from or written to using a surface reference - :py:obj:`~.cudaArrayTextureGather`: This flag indicates that texture gather operations will be performed on the array. - :py:obj:`~.cudaArraySparse`: Allocates a CUDA array without physical backing memory. The subregions within this sparse array can later be mapped onto a physical memory allocation by calling :py:obj:`~.cuMemMapArrayAsync`. The physical backing memory must be allocated via :py:obj:`~.cuMemCreate`. - :py:obj:`~.cudaArrayDeferredMapping`: Allocates a CUDA array without physical backing memory. The entire array can later be mapped onto a physical memory allocation by calling :py:obj:`~.cuMemMapArrayAsync`. The physical backing memory must be allocated via :py:obj:`~.cuMemCreate`. `width` and `height` must meet certain size requirements. See :py:obj:`~.cudaMalloc3DArray()` for more details. Parameters ---------- desc : :py:obj:`~.cudaChannelFormatDesc` Requested channel format width : size_t Requested array allocation width height : size_t Requested array allocation height flags : unsigned int Requested properties of allocated array Returns ------- cudaError_t :py:obj:`~.cudaSuccess`, :py:obj:`~.cudaErrorInvalidValue`, :py:obj:`~.cudaErrorMemoryAllocation` array : :py:obj:`~.cudaArray_t` Pointer to allocated array in device memory See Also -------- :py:obj:`~.cudaMalloc`, :py:obj:`~.cudaMallocPitch`, :py:obj:`~.cudaFree`, :py:obj:`~.cudaFreeArray`, :py:obj:`~.cudaMallocHost (C API)`, :py:obj:`~.cudaFreeHost`, :py:obj:`~.cudaMalloc3D`, :py:obj:`~.cudaMalloc3DArray`, :py:obj:`~.cudaHostAlloc`, :py:obj:`~.cuArrayCreate` """ cdef cudaArray_t array = cudaArray_t() cdef cyruntime.cudaChannelFormatDesc* cydesc_ptr = desc._pvt_ptr if desc is not None else NULL with nogil: err = cyruntime.cudaMallocArray(array._pvt_ptr, cydesc_ptr, width, height, flags) if err != cyruntime.cudaSuccess: return (_dict_cudaError_t[err], None) return (_dict_cudaError_t[err], array) @cython.embedsignature(True) def cudaFree(devPtr): """ Frees memory on the device. Frees the memory space pointed to by `devPtr`, which must have been returned by a previous call to one of the following memory allocation APIs - :py:obj:`~.cudaMalloc()`, :py:obj:`~.cudaMallocPitch()`, :py:obj:`~.cudaMallocManaged()`, :py:obj:`~.cudaMallocAsync()`, :py:obj:`~.cudaMallocFromPoolAsync()`. Note - This API will not perform any implicit synchronization when the pointer was allocated with :py:obj:`~.cudaMallocAsync` or :py:obj:`~.cudaMallocFromPoolAsync`. Callers must ensure that all accesses to these pointer have completed before invoking :py:obj:`~.cudaFree`. For best performance and memory reuse, users should use :py:obj:`~.cudaFreeAsync` to free memory allocated via the stream ordered memory allocator. For all other pointers, this API may perform implicit synchronization. If :py:obj:`~.cudaFree`(`devPtr`) has already been called before, an error is returned. If `devPtr` is 0, no operation is performed. :py:obj:`~.cudaFree()` returns :py:obj:`~.cudaErrorValue` in case of failure. The device version of :py:obj:`~.cudaFree` cannot be used with a `*devPtr` allocated using the host API, and vice versa. Parameters ---------- devPtr : Any Device pointer to memory to free Returns ------- cudaError_t :py:obj:`~.cudaSuccess`, :py:obj:`~.cudaErrorInvalidValue` See Also -------- :py:obj:`~.cudaMalloc`, :py:obj:`~.cudaMallocPitch`, :py:obj:`~.cudaMallocManaged`, :py:obj:`~.cudaMallocArray`, :py:obj:`~.cudaFreeArray`, :py:obj:`~.cudaMallocAsync`, :py:obj:`~.cudaMallocFromPoolAsync` :py:obj:`~.cudaMallocHost (C API)`, :py:obj:`~.cudaFreeHost`, :py:obj:`~.cudaMalloc3D`, :py:obj:`~.cudaMalloc3DArray`, :py:obj:`~.cudaFreeAsync` :py:obj:`~.cudaHostAlloc`, :py:obj:`~.cuMemFree` """ cydevPtr = _HelperInputVoidPtr(devPtr) cdef void* cydevPtr_ptr = cydevPtr.cptr with nogil: err = cyruntime.cudaFree(cydevPtr_ptr) return (_dict_cudaError_t[err],) @cython.embedsignature(True) def cudaFreeHost(ptr): """ Frees page-locked memory. Frees the memory space pointed to by `hostPtr`, which must have been returned by a previous call to :py:obj:`~.cudaMallocHost()` or :py:obj:`~.cudaHostAlloc()`. Parameters ---------- ptr : Any Pointer to memory to free Returns ------- cudaError_t :py:obj:`~.cudaSuccess`, :py:obj:`~.cudaErrorInvalidValue` See Also -------- :py:obj:`~.cudaMalloc`, :py:obj:`~.cudaMallocPitch`, :py:obj:`~.cudaFree`, :py:obj:`~.cudaMallocArray`, :py:obj:`~.cudaFreeArray`, :py:obj:`~.cudaMallocHost (C API)`, :py:obj:`~.cudaMalloc3D`, :py:obj:`~.cudaMalloc3DArray`, :py:obj:`~.cudaHostAlloc`, :py:obj:`~.cuMemFreeHost` """ cyptr = _HelperInputVoidPtr(ptr) cdef void* cyptr_ptr = cyptr.cptr with nogil: err = cyruntime.cudaFreeHost(cyptr_ptr) return (_dict_cudaError_t[err],) @cython.embedsignature(True) def cudaFreeArray(array): """ Frees an array on the device. Frees the CUDA array `array`, which must have been returned by a previous call to :py:obj:`~.cudaMallocArray()`. If `devPtr` is 0, no operation is performed. Parameters ---------- array : :py:obj:`~.cudaArray_t` Pointer to array to free Returns ------- cudaError_t :py:obj:`~.cudaSuccess`, :py:obj:`~.cudaErrorInvalidValue` See Also -------- :py:obj:`~.cudaMalloc`, :py:obj:`~.cudaMallocPitch`, :py:obj:`~.cudaFree`, :py:obj:`~.cudaMallocArray`, :py:obj:`~.cudaMallocHost (C API)`, :py:obj:`~.cudaFreeHost`, :py:obj:`~.cudaHostAlloc`, :py:obj:`~.cuArrayDestroy` """ cdef cyruntime.cudaArray_t cyarray if array is None: parray = 0 elif isinstance(array, (cudaArray_t,)): parray = int(array) else: parray = int(cudaArray_t(array)) cyarray = parray with nogil: err = cyruntime.cudaFreeArray(cyarray) return (_dict_cudaError_t[err],) @cython.embedsignature(True) def cudaFreeMipmappedArray(mipmappedArray): """ Frees a mipmapped array on the device. Frees the CUDA mipmapped array `mipmappedArray`, which must have been returned by a previous call to :py:obj:`~.cudaMallocMipmappedArray()`. If `devPtr` is 0, no operation is performed. Parameters ---------- mipmappedArray : :py:obj:`~.cudaMipmappedArray_t` Pointer to mipmapped array to free Returns ------- cudaError_t :py:obj:`~.cudaSuccess`, :py:obj:`~.cudaErrorInvalidValue` See Also -------- :py:obj:`~.cudaMalloc`, :py:obj:`~.cudaMallocPitch`, :py:obj:`~.cudaFree`, :py:obj:`~.cudaMallocArray`, :py:obj:`~.cudaMallocHost (C API)`, :py:obj:`~.cudaFreeHost`, :py:obj:`~.cudaHostAlloc`, :py:obj:`~.cuMipmappedArrayDestroy` """ cdef cyruntime.cudaMipmappedArray_t cymipmappedArray if mipmappedArray is None: pmipmappedArray = 0 elif isinstance(mipmappedArray, (cudaMipmappedArray_t,)): pmipmappedArray = int(mipmappedArray) else: pmipmappedArray = int(cudaMipmappedArray_t(mipmappedArray)) cymipmappedArray = pmipmappedArray with nogil: err = cyruntime.cudaFreeMipmappedArray(cymipmappedArray) return (_dict_cudaError_t[err],) @cython.embedsignature(True) def cudaHostAlloc(size_t size, unsigned int flags): """ Allocates page-locked memory on the host. Allocates `size` bytes of host memory that is page-locked and accessible to the device. The driver tracks the virtual memory ranges allocated with this function and automatically accelerates calls to functions such as :py:obj:`~.cudaMemcpy()`. Since the memory can be accessed directly by the device, it can be read or written with much higher bandwidth than pageable memory obtained with functions such as :py:obj:`~.malloc()`. Allocating excessive amounts of pinned memory may degrade system performance, since it reduces the amount of memory available to the system for paging. As a result, this function is best used sparingly to allocate staging areas for data exchange between host and device. The `flags` parameter enables different options to be specified that affect the allocation, as follows. - :py:obj:`~.cudaHostAllocDefault`: This flag's value is defined to be 0 and causes :py:obj:`~.cudaHostAlloc()` to emulate :py:obj:`~.cudaMallocHost()`. - :py:obj:`~.cudaHostAllocPortable`: The memory returned by this call will be considered as pinned memory by all CUDA contexts, not just the one that performed the allocation. - :py:obj:`~.cudaHostAllocMapped`: Maps the allocation into the CUDA address space. The device pointer to the memory may be obtained by calling :py:obj:`~.cudaHostGetDevicePointer()`. - :py:obj:`~.cudaHostAllocWriteCombined`: Allocates the memory as write-combined (WC). WC memory can be transferred across the PCI Express bus more quickly on some system configurations, but cannot be read efficiently by most CPUs. WC memory is a good option for buffers that will be written by the CPU and read by the device via mapped pinned memory or host->device transfers. All of these flags are orthogonal to one another: a developer may allocate memory that is portable, mapped and/or write-combined with no restrictions. In order for the :py:obj:`~.cudaHostAllocMapped` flag to have any effect, the CUDA context must support the :py:obj:`~.cudaDeviceMapHost` flag, which can be checked via :py:obj:`~.cudaGetDeviceFlags()`. The :py:obj:`~.cudaDeviceMapHost` flag is implicitly set for contexts created via the runtime API. The :py:obj:`~.cudaHostAllocMapped` flag may be specified on CUDA contexts for devices that do not support mapped pinned memory. The failure is deferred to :py:obj:`~.cudaHostGetDevicePointer()` because the memory may be mapped into other CUDA contexts via the :py:obj:`~.cudaHostAllocPortable` flag. Memory allocated by this function must be freed with :py:obj:`~.cudaFreeHost()`. Parameters ---------- size : size_t Requested allocation size in bytes flags : unsigned int Requested properties of allocated memory Returns ------- cudaError_t :py:obj:`~.cudaSuccess`, :py:obj:`~.cudaErrorInvalidValue`, :py:obj:`~.cudaErrorMemoryAllocation` pHost : Any Device pointer to allocated memory See Also -------- :py:obj:`~.cudaSetDeviceFlags`, :py:obj:`~.cudaMallocHost (C API)`, :py:obj:`~.cudaFreeHost`, :py:obj:`~.cudaGetDeviceFlags`, :py:obj:`~.cuMemHostAlloc` """ cdef void_ptr pHost = 0 with nogil: err = cyruntime.cudaHostAlloc(&pHost, size, flags) if err != cyruntime.cudaSuccess: return (_dict_cudaError_t[err], None) return (_dict_cudaError_t[err], pHost) @cython.embedsignature(True) def cudaHostRegister(ptr, size_t size, unsigned int flags): """ Registers an existing host memory range for use by CUDA. Page-locks the memory range specified by `ptr` and `size` and maps it for the device(s) as specified by `flags`. This memory range also is added to the same tracking mechanism as :py:obj:`~.cudaHostAlloc()` to automatically accelerate calls to functions such as :py:obj:`~.cudaMemcpy()`. Since the memory can be accessed directly by the device, it can be read or written with much higher bandwidth than pageable memory that has not been registered. Page-locking excessive amounts of memory may degrade system performance, since it reduces the amount of memory available to the system for paging. As a result, this function is best used sparingly to register staging areas for data exchange between host and device. On systems where :py:obj:`~.pageableMemoryAccessUsesHostPageTables` is true, :py:obj:`~.cudaHostRegister` will not page-lock the memory range specified by `ptr` but only populate unpopulated pages. :py:obj:`~.cudaHostRegister` is supported only on I/O coherent devices that have a non-zero value for the device attribute :py:obj:`~.cudaDevAttrHostRegisterSupported`. The `flags` parameter enables different options to be specified that affect the allocation, as follows. - :py:obj:`~.cudaHostRegisterDefault`: On a system with unified virtual addressing, the memory will be both mapped and portable. On a system with no unified virtual addressing, the memory will be neither mapped nor portable. - :py:obj:`~.cudaHostRegisterPortable`: The memory returned by this call will be considered as pinned memory by all CUDA contexts, not just the one that performed the allocation. - :py:obj:`~.cudaHostRegisterMapped`: Maps the allocation into the CUDA address space. The device pointer to the memory may be obtained by calling :py:obj:`~.cudaHostGetDevicePointer()`. - :py:obj:`~.cudaHostRegisterIoMemory`: The passed memory pointer is treated as pointing to some memory-mapped I/O space, e.g. belonging to a third-party PCIe device, and it will marked as non cache- coherent and contiguous. - :py:obj:`~.cudaHostRegisterReadOnly`: The passed memory pointer is treated as pointing to memory that is considered read-only by the device. On platforms without :py:obj:`~.cudaDevAttrPageableMemoryAccessUsesHostPageTables`, this flag is required in order to register memory mapped to the CPU as read-only. Support for the use of this flag can be queried from the device attribute :py:obj:`~.cudaDevAttrHostRegisterReadOnlySupported`. Using this flag with a current context associated with a device that does not have this attribute set will cause :py:obj:`~.cudaHostRegister` to error with cudaErrorNotSupported. All of these flags are orthogonal to one another: a developer may page- lock memory that is portable or mapped with no restrictions. The CUDA context must have been created with the :py:obj:`~.cudaMapHost` flag in order for the :py:obj:`~.cudaHostRegisterMapped` flag to have any effect. The :py:obj:`~.cudaHostRegisterMapped` flag may be specified on CUDA contexts for devices that do not support mapped pinned memory. The failure is deferred to :py:obj:`~.cudaHostGetDevicePointer()` because the memory may be mapped into other CUDA contexts via the :py:obj:`~.cudaHostRegisterPortable` flag. For devices that have a non-zero value for the device attribute :py:obj:`~.cudaDevAttrCanUseHostPointerForRegisteredMem`, the memory can also be accessed from the device using the host pointer `ptr`. The device pointer returned by :py:obj:`~.cudaHostGetDevicePointer()` may or may not match the original host pointer `ptr` and depends on the devices visible to the application. If all devices visible to the application have a non-zero value for the device attribute, the device pointer returned by :py:obj:`~.cudaHostGetDevicePointer()` will match the original pointer `ptr`. If any device visible to the application has a zero value for the device attribute, the device pointer returned by :py:obj:`~.cudaHostGetDevicePointer()` will not match the original host pointer `ptr`, but it will be suitable for use on all devices provided Unified Virtual Addressing is enabled. In such systems, it is valid to access the memory using either pointer on devices that have a non-zero value for the device attribute. Note however that such devices should access the memory using only of the two pointers and not both. The memory page-locked by this function must be unregistered with :py:obj:`~.cudaHostUnregister()`. Parameters ---------- ptr : Any Host pointer to memory to page-lock size : size_t Size in bytes of the address range to page-lock in bytes flags : unsigned int Flags for allocation request Returns ------- cudaError_t :py:obj:`~.cudaSuccess`, :py:obj:`~.cudaErrorInvalidValue`, :py:obj:`~.cudaErrorMemoryAllocation`, :py:obj:`~.cudaErrorHostMemoryAlreadyRegistered`, :py:obj:`~.cudaErrorNotSupported` See Also -------- :py:obj:`~.cudaHostUnregister`, :py:obj:`~.cudaHostGetFlags`, :py:obj:`~.cudaHostGetDevicePointer`, :py:obj:`~.cuMemHostRegister` """ cyptr = _HelperInputVoidPtr(ptr) cdef void* cyptr_ptr = cyptr.cptr with nogil: err = cyruntime.cudaHostRegister(cyptr_ptr, size, flags) return (_dict_cudaError_t[err],) @cython.embedsignature(True) def cudaHostUnregister(ptr): """ Unregisters a memory range that was registered with cudaHostRegister. Unmaps the memory range whose base address is specified by `ptr`, and makes it pageable again. The base address must be the same one specified to :py:obj:`~.cudaHostRegister()`. Parameters ---------- ptr : Any Host pointer to memory to unregister Returns ------- cudaError_t :py:obj:`~.cudaSuccess`, :py:obj:`~.cudaErrorInvalidValue`, :py:obj:`~.cudaErrorHostMemoryNotRegistered` See Also -------- :py:obj:`~.cudaHostUnregister`, :py:obj:`~.cuMemHostUnregister` """ cyptr = _HelperInputVoidPtr(ptr) cdef void* cyptr_ptr = cyptr.cptr with nogil: err = cyruntime.cudaHostUnregister(cyptr_ptr) return (_dict_cudaError_t[err],) @cython.embedsignature(True) def cudaHostGetDevicePointer(pHost, unsigned int flags): """ Passes back device pointer of mapped host memory allocated by cudaHostAlloc or registered by cudaHostRegister. Passes back the device pointer corresponding to the mapped, pinned host buffer allocated by :py:obj:`~.cudaHostAlloc()` or registered by :py:obj:`~.cudaHostRegister()`. :py:obj:`~.cudaHostGetDevicePointer()` will fail if the :py:obj:`~.cudaDeviceMapHost` flag was not specified before deferred context creation occurred, or if called on a device that does not support mapped, pinned memory. For devices that have a non-zero value for the device attribute :py:obj:`~.cudaDevAttrCanUseHostPointerForRegisteredMem`, the memory can also be accessed from the device using the host pointer `pHost`. The device pointer returned by :py:obj:`~.cudaHostGetDevicePointer()` may or may not match the original host pointer `pHost` and depends on the devices visible to the application. If all devices visible to the application have a non-zero value for the device attribute, the device pointer returned by :py:obj:`~.cudaHostGetDevicePointer()` will match the original pointer `pHost`. If any device visible to the application has a zero value for the device attribute, the device pointer returned by :py:obj:`~.cudaHostGetDevicePointer()` will not match the original host pointer `pHost`, but it will be suitable for use on all devices provided Unified Virtual Addressing is enabled. In such systems, it is valid to access the memory using either pointer on devices that have a non-zero value for the device attribute. Note however that such devices should access the memory using only of the two pointers and not both. `flags` provides for future releases. For now, it must be set to 0. Parameters ---------- pHost : Any Requested host pointer mapping flags : unsigned int Flags for extensions (must be 0 for now) Returns ------- cudaError_t :py:obj:`~.cudaSuccess`, :py:obj:`~.cudaErrorInvalidValue`, :py:obj:`~.cudaErrorMemoryAllocation` pDevice : Any Returned device pointer for mapped memory See Also -------- :py:obj:`~.cudaSetDeviceFlags`, :py:obj:`~.cudaHostAlloc`, :py:obj:`~.cuMemHostGetDevicePointer` """ cdef void_ptr pDevice = 0 cypHost = _HelperInputVoidPtr(pHost) cdef void* cypHost_ptr = cypHost.cptr with nogil: err = cyruntime.cudaHostGetDevicePointer(&pDevice, cypHost_ptr, flags) if err != cyruntime.cudaSuccess: return (_dict_cudaError_t[err], None) return (_dict_cudaError_t[err], pDevice) @cython.embedsignature(True) def cudaHostGetFlags(pHost): """ Passes back flags used to allocate pinned host memory allocated by cudaHostAlloc. :py:obj:`~.cudaHostGetFlags()` will fail if the input pointer does not reside in an address range allocated by :py:obj:`~.cudaHostAlloc()`. Parameters ---------- pHost : Any Host pointer Returns ------- cudaError_t :py:obj:`~.cudaSuccess`, :py:obj:`~.cudaErrorInvalidValue` pFlags : unsigned int Returned flags word See Also -------- :py:obj:`~.cudaHostAlloc`, :py:obj:`~.cuMemHostGetFlags` """ cdef unsigned int pFlags = 0 cypHost = _HelperInputVoidPtr(pHost) cdef void* cypHost_ptr = cypHost.cptr with nogil: err = cyruntime.cudaHostGetFlags(&pFlags, cypHost_ptr) if err != cyruntime.cudaSuccess: return (_dict_cudaError_t[err], None) return (_dict_cudaError_t[err], pFlags) @cython.embedsignature(True) def cudaMalloc3D(extent not None : cudaExtent): """ Allocates logical 1D, 2D, or 3D memory objects on the device. Allocates at least `width` * `height` * `depth` bytes of linear memory on the device and returns a :py:obj:`~.cudaPitchedPtr` in which `ptr` is a pointer to the allocated memory. The function may pad the allocation to ensure hardware alignment requirements are met. The pitch returned in the `pitch` field of `pitchedDevPtr` is the width in bytes of the allocation. The returned :py:obj:`~.cudaPitchedPtr` contains additional fields `xsize` and `ysize`, the logical width and height of the allocation, which are equivalent to the `width` and `height` `extent` parameters provided by the programmer during allocation. For allocations of 2D and 3D objects, it is highly recommended that programmers perform allocations using :py:obj:`~.cudaMalloc3D()` or :py:obj:`~.cudaMallocPitch()`. Due to alignment restrictions in the hardware, this is especially true if the application will be performing memory copies involving 2D or 3D objects (whether linear memory or CUDA arrays). Parameters ---------- extent : :py:obj:`~.cudaExtent` Requested allocation size (`width` field in bytes) Returns ------- cudaError_t :py:obj:`~.cudaSuccess`, :py:obj:`~.cudaErrorInvalidValue`, :py:obj:`~.cudaErrorMemoryAllocation` pitchedDevPtr : :py:obj:`~.cudaPitchedPtr` Pointer to allocated pitched device memory See Also -------- :py:obj:`~.cudaMallocPitch`, :py:obj:`~.cudaFree`, :py:obj:`~.cudaMemcpy3D`, :py:obj:`~.cudaMemset3D`, :py:obj:`~.cudaMalloc3DArray`, :py:obj:`~.cudaMallocArray`, :py:obj:`~.cudaFreeArray`, :py:obj:`~.cudaMallocHost (C API)`, :py:obj:`~.cudaFreeHost`, :py:obj:`~.cudaHostAlloc`, :py:obj:`~.make_cudaPitchedPtr`, :py:obj:`~.make_cudaExtent`, :py:obj:`~.cuMemAllocPitch` """ cdef cudaPitchedPtr pitchedDevPtr = cudaPitchedPtr() with nogil: err = cyruntime.cudaMalloc3D(pitchedDevPtr._pvt_ptr, extent._pvt_ptr[0]) if err != cyruntime.cudaSuccess: return (_dict_cudaError_t[err], None) return (_dict_cudaError_t[err], pitchedDevPtr) @cython.embedsignature(True) def cudaMalloc3DArray(desc : Optional[cudaChannelFormatDesc], extent not None : cudaExtent, unsigned int flags): """ Allocate an array on the device. Allocates a CUDA array according to the :py:obj:`~.cudaChannelFormatDesc` structure `desc` and returns a handle to the new CUDA array in `*array`. The :py:obj:`~.cudaChannelFormatDesc` is defined as: **View CUDA Toolkit Documentation for a C++ code example** where :py:obj:`~.cudaChannelFormatKind` is one of :py:obj:`~.cudaChannelFormatKindSigned`, :py:obj:`~.cudaChannelFormatKindUnsigned`, or :py:obj:`~.cudaChannelFormatKindFloat`. :py:obj:`~.cudaMalloc3DArray()` can allocate the following: - A 1D array is allocated if the height and depth extents are both zero. - A 2D array is allocated if only the depth extent is zero. - A 3D array is allocated if all three extents are non-zero. - A 1D layered CUDA array is allocated if only the height extent is zero and the cudaArrayLayered flag is set. Each layer is a 1D array. The number of layers is determined by the depth extent. - A 2D layered CUDA array is allocated if all three extents are non- zero and the cudaArrayLayered flag is set. Each layer is a 2D array. The number of layers is determined by the depth extent. - A cubemap CUDA array is allocated if all three extents are non-zero and the cudaArrayCubemap flag is set. Width must be equal to height, and depth must be six. A cubemap is a special type of 2D layered CUDA array, where the six layers represent the six faces of a cube. The order of the six layers in memory is the same as that listed in :py:obj:`~.cudaGraphicsCubeFace`. - A cubemap layered CUDA array is allocated if all three extents are non-zero, and both, cudaArrayCubemap and cudaArrayLayered flags are set. Width must be equal to height, and depth must be a multiple of six. A cubemap layered CUDA array is a special type of 2D layered CUDA array that consists of a collection of cubemaps. The first six layers represent the first cubemap, the next six layers form the second cubemap, and so on. The `flags` parameter enables different options to be specified that affect the allocation, as follows. - :py:obj:`~.cudaArrayDefault`: This flag's value is defined to be 0 and provides default array allocation - :py:obj:`~.cudaArrayLayered`: Allocates a layered CUDA array, with the depth extent indicating the number of layers - :py:obj:`~.cudaArrayCubemap`: Allocates a cubemap CUDA array. Width must be equal to height, and depth must be six. If the cudaArrayLayered flag is also set, depth must be a multiple of six. - :py:obj:`~.cudaArraySurfaceLoadStore`: Allocates a CUDA array that could be read from or written to using a surface reference. - :py:obj:`~.cudaArrayTextureGather`: This flag indicates that texture gather operations will be performed on the CUDA array. Texture gather can only be performed on 2D CUDA arrays. - :py:obj:`~.cudaArraySparse`: Allocates a CUDA array without physical backing memory. The subregions within this sparse array can later be mapped onto a physical memory allocation by calling :py:obj:`~.cuMemMapArrayAsync`. This flag can only be used for creating 2D, 3D or 2D layered sparse CUDA arrays. The physical backing memory must be allocated via :py:obj:`~.cuMemCreate`. - :py:obj:`~.cudaArrayDeferredMapping`: Allocates a CUDA array without physical backing memory. The entire array can later be mapped onto a physical memory allocation by calling :py:obj:`~.cuMemMapArrayAsync`. The physical backing memory must be allocated via :py:obj:`~.cuMemCreate`. The width, height and depth extents must meet certain size requirements as listed in the following table. All values are specified in elements. Note that 2D CUDA arrays have different size requirements if the :py:obj:`~.cudaArrayTextureGather` flag is set. In that case, the valid range for (width, height, depth) is ((1,maxTexture2DGather[0]), (1,maxTexture2DGather[1]), 0). **View CUDA Toolkit Documentation for a table example** Parameters ---------- desc : :py:obj:`~.cudaChannelFormatDesc` Requested channel format extent : :py:obj:`~.cudaExtent` Requested allocation size (`width` field in elements) flags : unsigned int Flags for extensions Returns ------- cudaError_t :py:obj:`~.cudaSuccess`, :py:obj:`~.cudaErrorInvalidValue`, :py:obj:`~.cudaErrorMemoryAllocation` array : :py:obj:`~.cudaArray_t` Pointer to allocated array in device memory See Also -------- :py:obj:`~.cudaMalloc3D`, :py:obj:`~.cudaMalloc`, :py:obj:`~.cudaMallocPitch`, :py:obj:`~.cudaFree`, :py:obj:`~.cudaFreeArray`, :py:obj:`~.cudaMallocHost (C API)`, :py:obj:`~.cudaFreeHost`, :py:obj:`~.cudaHostAlloc`, :py:obj:`~.make_cudaExtent`, :py:obj:`~.cuArray3DCreate` """ cdef cudaArray_t array = cudaArray_t() cdef cyruntime.cudaChannelFormatDesc* cydesc_ptr = desc._pvt_ptr if desc is not None else NULL with nogil: err = cyruntime.cudaMalloc3DArray(array._pvt_ptr, cydesc_ptr, extent._pvt_ptr[0], flags) if err != cyruntime.cudaSuccess: return (_dict_cudaError_t[err], None) return (_dict_cudaError_t[err], array) @cython.embedsignature(True) def cudaMallocMipmappedArray(desc : Optional[cudaChannelFormatDesc], extent not None : cudaExtent, unsigned int numLevels, unsigned int flags): """ Allocate a mipmapped array on the device. Allocates a CUDA mipmapped array according to the :py:obj:`~.cudaChannelFormatDesc` structure `desc` and returns a handle to the new CUDA mipmapped array in `*mipmappedArray`. `numLevels` specifies the number of mipmap levels to be allocated. This value is clamped to the range [1, 1 + floor(log2(max(width, height, depth)))]. The :py:obj:`~.cudaChannelFormatDesc` is defined as: **View CUDA Toolkit Documentation for a C++ code example** where :py:obj:`~.cudaChannelFormatKind` is one of :py:obj:`~.cudaChannelFormatKindSigned`, :py:obj:`~.cudaChannelFormatKindUnsigned`, or :py:obj:`~.cudaChannelFormatKindFloat`. :py:obj:`~.cudaMallocMipmappedArray()` can allocate the following: - A 1D mipmapped array is allocated if the height and depth extents are both zero. - A 2D mipmapped array is allocated if only the depth extent is zero. - A 3D mipmapped array is allocated if all three extents are non-zero. - A 1D layered CUDA mipmapped array is allocated if only the height extent is zero and the cudaArrayLayered flag is set. Each layer is a 1D mipmapped array. The number of layers is determined by the depth extent. - A 2D layered CUDA mipmapped array is allocated if all three extents are non-zero and the cudaArrayLayered flag is set. Each layer is a 2D mipmapped array. The number of layers is determined by the depth extent. - A cubemap CUDA mipmapped array is allocated if all three extents are non-zero and the cudaArrayCubemap flag is set. Width must be equal to height, and depth must be six. The order of the six layers in memory is the same as that listed in :py:obj:`~.cudaGraphicsCubeFace`. - A cubemap layered CUDA mipmapped array is allocated if all three extents are non-zero, and both, cudaArrayCubemap and cudaArrayLayered flags are set. Width must be equal to height, and depth must be a multiple of six. A cubemap layered CUDA mipmapped array is a special type of 2D layered CUDA mipmapped array that consists of a collection of cubemap mipmapped arrays. The first six layers represent the first cubemap mipmapped array, the next six layers form the second cubemap mipmapped array, and so on. The `flags` parameter enables different options to be specified that affect the allocation, as follows. - :py:obj:`~.cudaArrayDefault`: This flag's value is defined to be 0 and provides default mipmapped array allocation - :py:obj:`~.cudaArrayLayered`: Allocates a layered CUDA mipmapped array, with the depth extent indicating the number of layers - :py:obj:`~.cudaArrayCubemap`: Allocates a cubemap CUDA mipmapped array. Width must be equal to height, and depth must be six. If the cudaArrayLayered flag is also set, depth must be a multiple of six. - :py:obj:`~.cudaArraySurfaceLoadStore`: This flag indicates that individual mipmap levels of the CUDA mipmapped array will be read from or written to using a surface reference. - :py:obj:`~.cudaArrayTextureGather`: This flag indicates that texture gather operations will be performed on the CUDA array. Texture gather can only be performed on 2D CUDA mipmapped arrays, and the gather operations are performed only on the most detailed mipmap level. - :py:obj:`~.cudaArraySparse`: Allocates a CUDA mipmapped array without physical backing memory. The subregions within this sparse array can later be mapped onto a physical memory allocation by calling :py:obj:`~.cuMemMapArrayAsync`. This flag can only be used for creating 2D, 3D or 2D layered sparse CUDA mipmapped arrays. The physical backing memory must be allocated via :py:obj:`~.cuMemCreate`. - :py:obj:`~.cudaArrayDeferredMapping`: Allocates a CUDA mipmapped array without physical backing memory. The entire array can later be mapped onto a physical memory allocation by calling :py:obj:`~.cuMemMapArrayAsync`. The physical backing memory must be allocated via :py:obj:`~.cuMemCreate`. The width, height and depth extents must meet certain size requirements as listed in the following table. All values are specified in elements. **View CUDA Toolkit Documentation for a table example** Parameters ---------- desc : :py:obj:`~.cudaChannelFormatDesc` Requested channel format extent : :py:obj:`~.cudaExtent` Requested allocation size (`width` field in elements) numLevels : unsigned int Number of mipmap levels to allocate flags : unsigned int Flags for extensions Returns ------- cudaError_t :py:obj:`~.cudaSuccess`, :py:obj:`~.cudaErrorInvalidValue`, :py:obj:`~.cudaErrorMemoryAllocation` mipmappedArray : :py:obj:`~.cudaMipmappedArray_t` Pointer to allocated mipmapped array in device memory See Also -------- :py:obj:`~.cudaMalloc3D`, :py:obj:`~.cudaMalloc`, :py:obj:`~.cudaMallocPitch`, :py:obj:`~.cudaFree`, :py:obj:`~.cudaFreeArray`, :py:obj:`~.cudaMallocHost (C API)`, :py:obj:`~.cudaFreeHost`, :py:obj:`~.cudaHostAlloc`, :py:obj:`~.make_cudaExtent`, :py:obj:`~.cuMipmappedArrayCreate` """ cdef cudaMipmappedArray_t mipmappedArray = cudaMipmappedArray_t() cdef cyruntime.cudaChannelFormatDesc* cydesc_ptr = desc._pvt_ptr if desc is not None else NULL with nogil: err = cyruntime.cudaMallocMipmappedArray(mipmappedArray._pvt_ptr, cydesc_ptr, extent._pvt_ptr[0], numLevels, flags) if err != cyruntime.cudaSuccess: return (_dict_cudaError_t[err], None) return (_dict_cudaError_t[err], mipmappedArray) @cython.embedsignature(True) def cudaGetMipmappedArrayLevel(mipmappedArray, unsigned int level): """ Gets a mipmap level of a CUDA mipmapped array. Returns in `*levelArray` a CUDA array that represents a single mipmap level of the CUDA mipmapped array `mipmappedArray`. If `level` is greater than the maximum number of levels in this mipmapped array, :py:obj:`~.cudaErrorInvalidValue` is returned. If `mipmappedArray` is NULL, :py:obj:`~.cudaErrorInvalidResourceHandle` is returned. Parameters ---------- mipmappedArray : :py:obj:`~.cudaMipmappedArray_const_t` CUDA mipmapped array level : unsigned int Mipmap level Returns ------- cudaError_t :py:obj:`~.cudaSuccess`, :py:obj:`~.cudaErrorInvalidValue` :py:obj:`~.cudaErrorInvalidResourceHandle` levelArray : :py:obj:`~.cudaArray_t` Returned mipmap level CUDA array See Also -------- :py:obj:`~.cudaMalloc3D`, :py:obj:`~.cudaMalloc`, :py:obj:`~.cudaMallocPitch`, :py:obj:`~.cudaFree`, :py:obj:`~.cudaFreeArray`, :py:obj:`~.cudaMallocHost (C API)`, :py:obj:`~.cudaFreeHost`, :py:obj:`~.cudaHostAlloc`, :py:obj:`~.make_cudaExtent`, :py:obj:`~.cuMipmappedArrayGetLevel` """ cdef cyruntime.cudaMipmappedArray_const_t cymipmappedArray if mipmappedArray is None: pmipmappedArray = 0 elif isinstance(mipmappedArray, (cudaMipmappedArray_const_t,)): pmipmappedArray = int(mipmappedArray) else: pmipmappedArray = int(cudaMipmappedArray_const_t(mipmappedArray)) cymipmappedArray = pmipmappedArray cdef cudaArray_t levelArray = cudaArray_t() with nogil: err = cyruntime.cudaGetMipmappedArrayLevel(levelArray._pvt_ptr, cymipmappedArray, level) if err != cyruntime.cudaSuccess: return (_dict_cudaError_t[err], None) return (_dict_cudaError_t[err], levelArray) @cython.embedsignature(True) def cudaMemcpy3D(p : Optional[cudaMemcpy3DParms]): """ Copies data between 3D objects. **View CUDA Toolkit Documentation for a C++ code example** :py:obj:`~.cudaMemcpy3D()` copies data betwen two 3D objects. The source and destination objects may be in either host memory, device memory, or a CUDA array. The source, destination, extent, and kind of copy performed is specified by the :py:obj:`~.cudaMemcpy3DParms` struct which should be initialized to zero before use: **View CUDA Toolkit Documentation for a C++ code example** The struct passed to :py:obj:`~.cudaMemcpy3D()` must specify one of `srcArray` or `srcPtr` and one of `dstArray` or `dstPtr`. Passing more than one non-zero source or destination will cause :py:obj:`~.cudaMemcpy3D()` to return an error. The `srcPos` and `dstPos` fields are optional offsets into the source and destination objects and are defined in units of each object's elements. The element for a host or device pointer is assumed to be unsigned char. The `extent` field defines the dimensions of the transferred area in elements. If a CUDA array is participating in the copy, the extent is defined in terms of that array's elements. If no CUDA array is participating in the copy then the extents are defined in elements of unsigned char. The `kind` field defines the direction of the copy. It must be one of :py:obj:`~.cudaMemcpyHostToHost`, :py:obj:`~.cudaMemcpyHostToDevice`, :py:obj:`~.cudaMemcpyDeviceToHost`, :py:obj:`~.cudaMemcpyDeviceToDevice`, or :py:obj:`~.cudaMemcpyDefault`. Passing :py:obj:`~.cudaMemcpyDefault` is recommended, in which case the type of transfer is inferred from the pointer values. However, :py:obj:`~.cudaMemcpyDefault` is only allowed on systems that support unified virtual addressing. For :py:obj:`~.cudaMemcpyHostToHost` or :py:obj:`~.cudaMemcpyHostToDevice` or :py:obj:`~.cudaMemcpyDeviceToHost` passed as kind and cudaArray type passed as source or destination, if the kind implies cudaArray type to be present on the host, :py:obj:`~.cudaMemcpy3D()` will disregard that implication and silently correct the kind based on the fact that cudaArray type can only be present on the device. If the source and destination are both arrays, :py:obj:`~.cudaMemcpy3D()` will return an error if they do not have the same element size. The source and destination object may not overlap. If overlapping source and destination objects are specified, undefined behavior will result. The source object must entirely contain the region defined by `srcPos` and `extent`. The destination object must entirely contain the region defined by `dstPos` and `extent`. :py:obj:`~.cudaMemcpy3D()` returns an error if the pitch of `srcPtr` or `dstPtr` exceeds the maximum allowed. The pitch of a :py:obj:`~.cudaPitchedPtr` allocated with :py:obj:`~.cudaMalloc3D()` will always be valid. Parameters ---------- p : :py:obj:`~.cudaMemcpy3DParms` 3D memory copy parameters Returns ------- cudaError_t :py:obj:`~.cudaSuccess`, :py:obj:`~.cudaErrorInvalidValue`, :py:obj:`~.cudaErrorInvalidPitchValue`, :py:obj:`~.cudaErrorInvalidMemcpyDirection` See Also -------- :py:obj:`~.cudaMalloc3D`, :py:obj:`~.cudaMalloc3DArray`, :py:obj:`~.cudaMemset3D`, :py:obj:`~.cudaMemcpy3DAsync`, :py:obj:`~.cudaMemcpy`, :py:obj:`~.cudaMemcpy2D`, :py:obj:`~.cudaMemcpy2DToArray`, :py:obj:`~.cudaMemcpy2DFromArray`, :py:obj:`~.cudaMemcpy2DArrayToArray`, :py:obj:`~.cudaMemcpyToSymbol`, :py:obj:`~.cudaMemcpyFromSymbol`, :py:obj:`~.cudaMemcpyAsync`, :py:obj:`~.cudaMemcpy2DAsync`, :py:obj:`~.cudaMemcpy2DToArrayAsync`, :py:obj:`~.cudaMemcpy2DFromArrayAsync`, :py:obj:`~.cudaMemcpyToSymbolAsync`, :py:obj:`~.cudaMemcpyFromSymbolAsync`, :py:obj:`~.make_cudaExtent`, :py:obj:`~.make_cudaPos`, :py:obj:`~.cuMemcpy3D` """ cdef cyruntime.cudaMemcpy3DParms* cyp_ptr = p._pvt_ptr if p is not None else NULL with nogil: err = cyruntime.cudaMemcpy3D(cyp_ptr) return (_dict_cudaError_t[err],) @cython.embedsignature(True) def cudaMemcpy3DPeer(p : Optional[cudaMemcpy3DPeerParms]): """ Copies memory between devices. Perform a 3D memory copy according to the parameters specified in `p`. See the definition of the :py:obj:`~.cudaMemcpy3DPeerParms` structure for documentation of its parameters. Note that this function is synchronous with respect to the host only if the source or destination of the transfer is host memory. Note also that this copy is serialized with respect to all pending and future asynchronous work in to the current device, the copy's source device, and the copy's destination device (use :py:obj:`~.cudaMemcpy3DPeerAsync` to avoid this synchronization). Parameters ---------- p : :py:obj:`~.cudaMemcpy3DPeerParms` Parameters for the memory copy Returns ------- cudaError_t :py:obj:`~.cudaSuccess`, :py:obj:`~.cudaErrorInvalidValue`, :py:obj:`~.cudaErrorInvalidDevice`, :py:obj:`~.cudaErrorInvalidPitchValue` See Also -------- :py:obj:`~.cudaMemcpy`, :py:obj:`~.cudaMemcpyPeer`, :py:obj:`~.cudaMemcpyAsync`, :py:obj:`~.cudaMemcpyPeerAsync`, :py:obj:`~.cudaMemcpy3DPeerAsync`, :py:obj:`~.cuMemcpy3DPeer` """ cdef cyruntime.cudaMemcpy3DPeerParms* cyp_ptr = p._pvt_ptr if p is not None else NULL with nogil: err = cyruntime.cudaMemcpy3DPeer(cyp_ptr) return (_dict_cudaError_t[err],) @cython.embedsignature(True) def cudaMemcpy3DAsync(p : Optional[cudaMemcpy3DParms], stream): """ Copies data between 3D objects. **View CUDA Toolkit Documentation for a C++ code example** :py:obj:`~.cudaMemcpy3DAsync()` copies data betwen two 3D objects. The source and destination objects may be in either host memory, device memory, or a CUDA array. The source, destination, extent, and kind of copy performed is specified by the :py:obj:`~.cudaMemcpy3DParms` struct which should be initialized to zero before use: **View CUDA Toolkit Documentation for a C++ code example** The struct passed to :py:obj:`~.cudaMemcpy3DAsync()` must specify one of `srcArray` or `srcPtr` and one of `dstArray` or `dstPtr`. Passing more than one non-zero source or destination will cause :py:obj:`~.cudaMemcpy3DAsync()` to return an error. The `srcPos` and `dstPos` fields are optional offsets into the source and destination objects and are defined in units of each object's elements. The element for a host or device pointer is assumed to be unsigned char. For CUDA arrays, positions must be in the range [0, 2048) for any dimension. The `extent` field defines the dimensions of the transferred area in elements. If a CUDA array is participating in the copy, the extent is defined in terms of that array's elements. If no CUDA array is participating in the copy then the extents are defined in elements of unsigned char. The `kind` field defines the direction of the copy. It must be one of :py:obj:`~.cudaMemcpyHostToHost`, :py:obj:`~.cudaMemcpyHostToDevice`, :py:obj:`~.cudaMemcpyDeviceToHost`, :py:obj:`~.cudaMemcpyDeviceToDevice`, or :py:obj:`~.cudaMemcpyDefault`. Passing :py:obj:`~.cudaMemcpyDefault` is recommended, in which case the type of transfer is inferred from the pointer values. However, :py:obj:`~.cudaMemcpyDefault` is only allowed on systems that support unified virtual addressing. For :py:obj:`~.cudaMemcpyHostToHost` or :py:obj:`~.cudaMemcpyHostToDevice` or :py:obj:`~.cudaMemcpyDeviceToHost` passed as kind and cudaArray type passed as source or destination, if the kind implies cudaArray type to be present on the host, :py:obj:`~.cudaMemcpy3DAsync()` will disregard that implication and silently correct the kind based on the fact that cudaArray type can only be present on the device. If the source and destination are both arrays, :py:obj:`~.cudaMemcpy3DAsync()` will return an error if they do not have the same element size. The source and destination object may not overlap. If overlapping source and destination objects are specified, undefined behavior will result. The source object must lie entirely within the region defined by `srcPos` and `extent`. The destination object must lie entirely within the region defined by `dstPos` and `extent`. :py:obj:`~.cudaMemcpy3DAsync()` returns an error if the pitch of `srcPtr` or `dstPtr` exceeds the maximum allowed. The pitch of a :py:obj:`~.cudaPitchedPtr` allocated with :py:obj:`~.cudaMalloc3D()` will always be valid. :py:obj:`~.cudaMemcpy3DAsync()` is asynchronous with respect to the host, so the call may return before the copy is complete. The copy can optionally be associated to a stream by passing a non-zero `stream` argument. If `kind` is :py:obj:`~.cudaMemcpyHostToDevice` or :py:obj:`~.cudaMemcpyDeviceToHost` and `stream` is non-zero, the copy may overlap with operations in other streams. The device version of this function only handles device to device copies and cannot be given local or shared pointers. Parameters ---------- p : :py:obj:`~.cudaMemcpy3DParms` 3D memory copy parameters stream : :py:obj:`~.CUstream` or :py:obj:`~.cudaStream_t` Stream identifier Returns ------- cudaError_t :py:obj:`~.cudaSuccess`, :py:obj:`~.cudaErrorInvalidValue`, :py:obj:`~.cudaErrorInvalidPitchValue`, :py:obj:`~.cudaErrorInvalidMemcpyDirection` See Also -------- :py:obj:`~.cudaMalloc3D`, :py:obj:`~.cudaMalloc3DArray`, :py:obj:`~.cudaMemset3D`, :py:obj:`~.cudaMemcpy3D`, :py:obj:`~.cudaMemcpy`, :py:obj:`~.cudaMemcpy2D`, :py:obj:`~.cudaMemcpy2DToArray`, ::::py:obj:`~.cudaMemcpy2DFromArray`, :py:obj:`~.cudaMemcpy2DArrayToArray`, :py:obj:`~.cudaMemcpyToSymbol`, :py:obj:`~.cudaMemcpyFromSymbol`, :py:obj:`~.cudaMemcpyAsync`, :py:obj:`~.cudaMemcpy2DAsync`, :py:obj:`~.cudaMemcpy2DToArrayAsync`, :py:obj:`~.cudaMemcpy2DFromArrayAsync`, :py:obj:`~.cudaMemcpyToSymbolAsync`, :py:obj:`~.cudaMemcpyFromSymbolAsync`, :py:obj:`~.make_cudaExtent`, :py:obj:`~.make_cudaPos`, :py:obj:`~.cuMemcpy3DAsync` """ cdef cyruntime.cudaStream_t cystream if stream is None: pstream = 0 elif isinstance(stream, (cudaStream_t,driver.CUstream)): pstream = int(stream) else: pstream = int(cudaStream_t(stream)) cystream = pstream cdef cyruntime.cudaMemcpy3DParms* cyp_ptr = p._pvt_ptr if p is not None else NULL with nogil: err = cyruntime.cudaMemcpy3DAsync(cyp_ptr, cystream) return (_dict_cudaError_t[err],) @cython.embedsignature(True) def cudaMemcpy3DPeerAsync(p : Optional[cudaMemcpy3DPeerParms], stream): """ Copies memory between devices asynchronously. Perform a 3D memory copy according to the parameters specified in `p`. See the definition of the :py:obj:`~.cudaMemcpy3DPeerParms` structure for documentation of its parameters. Parameters ---------- p : :py:obj:`~.cudaMemcpy3DPeerParms` Parameters for the memory copy stream : :py:obj:`~.CUstream` or :py:obj:`~.cudaStream_t` Stream identifier Returns ------- cudaError_t :py:obj:`~.cudaSuccess`, :py:obj:`~.cudaErrorInvalidValue`, :py:obj:`~.cudaErrorInvalidDevice`, :py:obj:`~.cudaErrorInvalidPitchValue` See Also -------- :py:obj:`~.cudaMemcpy`, :py:obj:`~.cudaMemcpyPeer`, :py:obj:`~.cudaMemcpyAsync`, :py:obj:`~.cudaMemcpyPeerAsync`, :py:obj:`~.cudaMemcpy3DPeerAsync`, :py:obj:`~.cuMemcpy3DPeerAsync` """ cdef cyruntime.cudaStream_t cystream if stream is None: pstream = 0 elif isinstance(stream, (cudaStream_t,driver.CUstream)): pstream = int(stream) else: pstream = int(cudaStream_t(stream)) cystream = pstream cdef cyruntime.cudaMemcpy3DPeerParms* cyp_ptr = p._pvt_ptr if p is not None else NULL with nogil: err = cyruntime.cudaMemcpy3DPeerAsync(cyp_ptr, cystream) return (_dict_cudaError_t[err],) @cython.embedsignature(True) def cudaMemGetInfo(): """ Gets free and total device memory. Returns in `*total` the total amount of memory available to the the current context. Returns in `*free` the amount of memory on the device that is free according to the OS. CUDA is not guaranteed to be able to allocate all of the memory that the OS reports as free. In a multi- tenet situation, free estimate returned is prone to race condition where a new allocation/free done by a different process or a different thread in the same process between the time when free memory was estimated and reported, will result in deviation in free value reported and actual free memory. The integrated GPU on Tegra shares memory with CPU and other component of the SoC. The free and total values returned by the API excludes the SWAP memory space maintained by the OS on some platforms. The OS may move some of the memory pages into swap area as the GPU or CPU allocate or access memory. See Tegra app note on how to calculate total and free memory on Tegra. Returns ------- cudaError_t :py:obj:`~.cudaSuccess`, :py:obj:`~.cudaErrorInvalidValue`, :py:obj:`~.cudaErrorLaunchFailure` free : int Returned free memory in bytes total : int Returned total memory in bytes See Also -------- :py:obj:`~.cuMemGetInfo` """ cdef size_t free = 0 cdef size_t total = 0 with nogil: err = cyruntime.cudaMemGetInfo(&free, &total) if err != cyruntime.cudaSuccess: return (_dict_cudaError_t[err], None, None) return (_dict_cudaError_t[err], free, total) @cython.embedsignature(True) def cudaArrayGetInfo(array): """ Gets info about the specified cudaArray. Returns in `*desc`, `*extent` and `*flags` respectively, the type, shape and flags of `array`. Any of `*desc`, `*extent` and `*flags` may be specified as NULL. Parameters ---------- array : :py:obj:`~.cudaArray_t` The :py:obj:`~.cudaArray` to get info for Returns ------- cudaError_t :py:obj:`~.cudaSuccess`, :py:obj:`~.cudaErrorInvalidValue` desc : :py:obj:`~.cudaChannelFormatDesc` Returned array type extent : :py:obj:`~.cudaExtent` Returned array shape. 2D arrays will have depth of zero flags : unsigned int Returned array flags See Also -------- :py:obj:`~.cuArrayGetDescriptor`, :py:obj:`~.cuArray3DGetDescriptor` """ cdef cyruntime.cudaArray_t cyarray if array is None: parray = 0 elif isinstance(array, (cudaArray_t,)): parray = int(array) else: parray = int(cudaArray_t(array)) cyarray = parray cdef cudaChannelFormatDesc desc = cudaChannelFormatDesc() cdef cudaExtent extent = cudaExtent() cdef unsigned int flags = 0 with nogil: err = cyruntime.cudaArrayGetInfo(desc._pvt_ptr, extent._pvt_ptr, &flags, cyarray) if err != cyruntime.cudaSuccess: return (_dict_cudaError_t[err], None, None, None) return (_dict_cudaError_t[err], desc, extent, flags) @cython.embedsignature(True) def cudaArrayGetPlane(hArray, unsigned int planeIdx): """ Gets a CUDA array plane from a CUDA array. Returns in `pPlaneArray` a CUDA array that represents a single format plane of the CUDA array `hArray`. If `planeIdx` is greater than the maximum number of planes in this array or if the array does not have a multi-planar format e.g: :py:obj:`~.cudaChannelFormatKindNV12`, then :py:obj:`~.cudaErrorInvalidValue` is returned. Note that if the `hArray` has format :py:obj:`~.cudaChannelFormatKindNV12`, then passing in 0 for `planeIdx` returns a CUDA array of the same size as `hArray` but with one 8-bit channel and :py:obj:`~.cudaChannelFormatKindUnsigned` as its format kind. If 1 is passed for `planeIdx`, then the returned CUDA array has half the height and width of `hArray` with two 8-bit channels and :py:obj:`~.cudaChannelFormatKindUnsigned` as its format kind. Parameters ---------- hArray : :py:obj:`~.cudaArray_t` CUDA array planeIdx : unsigned int Plane index Returns ------- cudaError_t :py:obj:`~.cudaSuccess`, :py:obj:`~.cudaErrorInvalidValue` :py:obj:`~.cudaErrorInvalidResourceHandle` pPlaneArray : :py:obj:`~.cudaArray_t` Returned CUDA array referenced by the `planeIdx` See Also -------- :py:obj:`~.cuArrayGetPlane` """ cdef cyruntime.cudaArray_t cyhArray if hArray is None: phArray = 0 elif isinstance(hArray, (cudaArray_t,)): phArray = int(hArray) else: phArray = int(cudaArray_t(hArray)) cyhArray = phArray cdef cudaArray_t pPlaneArray = cudaArray_t() with nogil: err = cyruntime.cudaArrayGetPlane(pPlaneArray._pvt_ptr, cyhArray, planeIdx) if err != cyruntime.cudaSuccess: return (_dict_cudaError_t[err], None) return (_dict_cudaError_t[err], pPlaneArray) @cython.embedsignature(True) def cudaArrayGetMemoryRequirements(array, int device): """ Returns the memory requirements of a CUDA array. Returns the memory requirements of a CUDA array in `memoryRequirements` If the CUDA array is not allocated with flag :py:obj:`~.cudaArrayDeferredMapping` :py:obj:`~.cudaErrorInvalidValue` will be returned. The returned value in :py:obj:`~.cudaArrayMemoryRequirements.size` represents the total size of the CUDA array. The returned value in :py:obj:`~.cudaArrayMemoryRequirements.alignment` represents the alignment necessary for mapping the CUDA array. Parameters ---------- array : :py:obj:`~.cudaArray_t` CUDA array to get the memory requirements of device : int Device to get the memory requirements for Returns ------- cudaError_t :py:obj:`~.cudaSuccess` :py:obj:`~.cudaErrorInvalidValue` memoryRequirements : :py:obj:`~.cudaArrayMemoryRequirements` Pointer to :py:obj:`~.cudaArrayMemoryRequirements` See Also -------- :py:obj:`~.cudaMipmappedArrayGetMemoryRequirements` """ cdef cyruntime.cudaArray_t cyarray if array is None: parray = 0 elif isinstance(array, (cudaArray_t,)): parray = int(array) else: parray = int(cudaArray_t(array)) cyarray = parray cdef cudaArrayMemoryRequirements memoryRequirements = cudaArrayMemoryRequirements() with nogil: err = cyruntime.cudaArrayGetMemoryRequirements(memoryRequirements._pvt_ptr, cyarray, device) if err != cyruntime.cudaSuccess: return (_dict_cudaError_t[err], None) return (_dict_cudaError_t[err], memoryRequirements) @cython.embedsignature(True) def cudaMipmappedArrayGetMemoryRequirements(mipmap, int device): """ Returns the memory requirements of a CUDA mipmapped array. Returns the memory requirements of a CUDA mipmapped array in `memoryRequirements` If the CUDA mipmapped array is not allocated with flag :py:obj:`~.cudaArrayDeferredMapping` :py:obj:`~.cudaErrorInvalidValue` will be returned. The returned value in :py:obj:`~.cudaArrayMemoryRequirements.size` represents the total size of the CUDA mipmapped array. The returned value in :py:obj:`~.cudaArrayMemoryRequirements.alignment` represents the alignment necessary for mapping the CUDA mipmapped array. Parameters ---------- mipmap : :py:obj:`~.cudaMipmappedArray_t` CUDA mipmapped array to get the memory requirements of device : int Device to get the memory requirements for Returns ------- cudaError_t :py:obj:`~.cudaSuccess` :py:obj:`~.cudaErrorInvalidValue` memoryRequirements : :py:obj:`~.cudaArrayMemoryRequirements` Pointer to :py:obj:`~.cudaArrayMemoryRequirements` See Also -------- :py:obj:`~.cudaArrayGetMemoryRequirements` """ cdef cyruntime.cudaMipmappedArray_t cymipmap if mipmap is None: pmipmap = 0 elif isinstance(mipmap, (cudaMipmappedArray_t,)): pmipmap = int(mipmap) else: pmipmap = int(cudaMipmappedArray_t(mipmap)) cymipmap = pmipmap cdef cudaArrayMemoryRequirements memoryRequirements = cudaArrayMemoryRequirements() with nogil: err = cyruntime.cudaMipmappedArrayGetMemoryRequirements(memoryRequirements._pvt_ptr, cymipmap, device) if err != cyruntime.cudaSuccess: return (_dict_cudaError_t[err], None) return (_dict_cudaError_t[err], memoryRequirements) @cython.embedsignature(True) def cudaArrayGetSparseProperties(array): """ Returns the layout properties of a sparse CUDA array. Returns the layout properties of a sparse CUDA array in `sparseProperties`. If the CUDA array is not allocated with flag :py:obj:`~.cudaArraySparse` :py:obj:`~.cudaErrorInvalidValue` will be returned. If the returned value in :py:obj:`~.cudaArraySparseProperties.flags` contains :py:obj:`~.cudaArraySparsePropertiesSingleMipTail`, then :py:obj:`~.cudaArraySparseProperties.miptailSize` represents the total size of the array. Otherwise, it will be zero. Also, the returned value in :py:obj:`~.cudaArraySparseProperties.miptailFirstLevel` is always zero. Note that the `array` must have been allocated using :py:obj:`~.cudaMallocArray` or :py:obj:`~.cudaMalloc3DArray`. For CUDA arrays obtained using :py:obj:`~.cudaMipmappedArrayGetLevel`, :py:obj:`~.cudaErrorInvalidValue` will be returned. Instead, :py:obj:`~.cudaMipmappedArrayGetSparseProperties` must be used to obtain the sparse properties of the entire CUDA mipmapped array to which `array` belongs to. Parameters ---------- array : :py:obj:`~.cudaArray_t` The CUDA array to get the sparse properties of Returns ------- cudaError_t :py:obj:`~.cudaSuccess` :py:obj:`~.cudaErrorInvalidValue` sparseProperties : :py:obj:`~.cudaArraySparseProperties` Pointer to return the :py:obj:`~.cudaArraySparseProperties` See Also -------- :py:obj:`~.cudaMipmappedArrayGetSparseProperties`, :py:obj:`~.cuMemMapArrayAsync` """ cdef cyruntime.cudaArray_t cyarray if array is None: parray = 0 elif isinstance(array, (cudaArray_t,)): parray = int(array) else: parray = int(cudaArray_t(array)) cyarray = parray cdef cudaArraySparseProperties sparseProperties = cudaArraySparseProperties() with nogil: err = cyruntime.cudaArrayGetSparseProperties(sparseProperties._pvt_ptr, cyarray) if err != cyruntime.cudaSuccess: return (_dict_cudaError_t[err], None) return (_dict_cudaError_t[err], sparseProperties) @cython.embedsignature(True) def cudaMipmappedArrayGetSparseProperties(mipmap): """ Returns the layout properties of a sparse CUDA mipmapped array. Returns the sparse array layout properties in `sparseProperties`. If the CUDA mipmapped array is not allocated with flag :py:obj:`~.cudaArraySparse` :py:obj:`~.cudaErrorInvalidValue` will be returned. For non-layered CUDA mipmapped arrays, :py:obj:`~.cudaArraySparseProperties.miptailSize` returns the size of the mip tail region. The mip tail region includes all mip levels whose width, height or depth is less than that of the tile. For layered CUDA mipmapped arrays, if :py:obj:`~.cudaArraySparseProperties.flags` contains :py:obj:`~.cudaArraySparsePropertiesSingleMipTail`, then :py:obj:`~.cudaArraySparseProperties.miptailSize` specifies the size of the mip tail of all layers combined. Otherwise, :py:obj:`~.cudaArraySparseProperties.miptailSize` specifies mip tail size per layer. The returned value of :py:obj:`~.cudaArraySparseProperties.miptailFirstLevel` is valid only if :py:obj:`~.cudaArraySparseProperties.miptailSize` is non-zero. Parameters ---------- mipmap : :py:obj:`~.cudaMipmappedArray_t` The CUDA mipmapped array to get the sparse properties of Returns ------- cudaError_t :py:obj:`~.cudaSuccess` :py:obj:`~.cudaErrorInvalidValue` sparseProperties : :py:obj:`~.cudaArraySparseProperties` Pointer to return :py:obj:`~.cudaArraySparseProperties` See Also -------- :py:obj:`~.cudaArrayGetSparseProperties`, :py:obj:`~.cuMemMapArrayAsync` """ cdef cyruntime.cudaMipmappedArray_t cymipmap if mipmap is None: pmipmap = 0 elif isinstance(mipmap, (cudaMipmappedArray_t,)): pmipmap = int(mipmap) else: pmipmap = int(cudaMipmappedArray_t(mipmap)) cymipmap = pmipmap cdef cudaArraySparseProperties sparseProperties = cudaArraySparseProperties() with nogil: err = cyruntime.cudaMipmappedArrayGetSparseProperties(sparseProperties._pvt_ptr, cymipmap) if err != cyruntime.cudaSuccess: return (_dict_cudaError_t[err], None) return (_dict_cudaError_t[err], sparseProperties) @cython.embedsignature(True) def cudaMemcpy(dst, src, size_t count, kind not None : cudaMemcpyKind): """ Copies data between host and device. Copies `count` bytes from the memory area pointed to by `src` to the memory area pointed to by `dst`, where `kind` specifies the direction of the copy, and must be one of :py:obj:`~.cudaMemcpyHostToHost`, :py:obj:`~.cudaMemcpyHostToDevice`, :py:obj:`~.cudaMemcpyDeviceToHost`, :py:obj:`~.cudaMemcpyDeviceToDevice`, or :py:obj:`~.cudaMemcpyDefault`. Passing :py:obj:`~.cudaMemcpyDefault` is recommended, in which case the type of transfer is inferred from the pointer values. However, :py:obj:`~.cudaMemcpyDefault` is only allowed on systems that support unified virtual addressing. Calling :py:obj:`~.cudaMemcpy()` with dst and src pointers that do not match the direction of the copy results in an undefined behavior. \note_sync Parameters ---------- dst : Any Destination memory address src : Any Source memory address count : size_t Size in bytes to copy kind : :py:obj:`~.cudaMemcpyKind` Type of transfer Returns ------- cudaError_t :py:obj:`~.cudaSuccess`, :py:obj:`~.cudaErrorInvalidValue`, :py:obj:`~.cudaErrorInvalidMemcpyDirection` See Also -------- :py:obj:`~.cudaMemcpy2D`, :py:obj:`~.cudaMemcpy2DToArray`, :py:obj:`~.cudaMemcpy2DFromArray`, :py:obj:`~.cudaMemcpy2DArrayToArray`, :py:obj:`~.cudaMemcpyToSymbol`, :py:obj:`~.cudaMemcpyFromSymbol`, :py:obj:`~.cudaMemcpyAsync`, :py:obj:`~.cudaMemcpy2DAsync`, :py:obj:`~.cudaMemcpy2DToArrayAsync`, :py:obj:`~.cudaMemcpy2DFromArrayAsync`, :py:obj:`~.cudaMemcpyToSymbolAsync`, :py:obj:`~.cudaMemcpyFromSymbolAsync`, :py:obj:`~.cuMemcpyDtoH`, :py:obj:`~.cuMemcpyHtoD`, :py:obj:`~.cuMemcpyDtoD`, :py:obj:`~.cuMemcpy` """ cydst = _HelperInputVoidPtr(dst) cdef void* cydst_ptr = cydst.cptr cysrc = _HelperInputVoidPtr(src) cdef void* cysrc_ptr = cysrc.cptr cdef cyruntime.cudaMemcpyKind cykind = kind.value with nogil: err = cyruntime.cudaMemcpy(cydst_ptr, cysrc_ptr, count, cykind) return (_dict_cudaError_t[err],) @cython.embedsignature(True) def cudaMemcpyPeer(dst, int dstDevice, src, int srcDevice, size_t count): """ Copies memory between two devices. Copies memory from one device to memory on another device. `dst` is the base device pointer of the destination memory and `dstDevice` is the destination device. `src` is the base device pointer of the source memory and `srcDevice` is the source device. `count` specifies the number of bytes to copy. Note that this function is asynchronous with respect to the host, but serialized with respect all pending and future asynchronous work in to the current device, `srcDevice`, and `dstDevice` (use :py:obj:`~.cudaMemcpyPeerAsync` to avoid this synchronization). Parameters ---------- dst : Any Destination device pointer dstDevice : int Destination device src : Any Source device pointer srcDevice : int Source device count : size_t Size of memory copy in bytes Returns ------- cudaError_t :py:obj:`~.cudaSuccess`, :py:obj:`~.cudaErrorInvalidValue`, :py:obj:`~.cudaErrorInvalidDevice` See Also -------- :py:obj:`~.cudaMemcpy`, :py:obj:`~.cudaMemcpyAsync`, :py:obj:`~.cudaMemcpyPeerAsync`, :py:obj:`~.cudaMemcpy3DPeerAsync`, :py:obj:`~.cuMemcpyPeer` """ cydst = _HelperInputVoidPtr(dst) cdef void* cydst_ptr = cydst.cptr cysrc = _HelperInputVoidPtr(src) cdef void* cysrc_ptr = cysrc.cptr with nogil: err = cyruntime.cudaMemcpyPeer(cydst_ptr, dstDevice, cysrc_ptr, srcDevice, count) return (_dict_cudaError_t[err],) @cython.embedsignature(True) def cudaMemcpy2D(dst, size_t dpitch, src, size_t spitch, size_t width, size_t height, kind not None : cudaMemcpyKind): """ Copies data between host and device. Copies a matrix (`height` rows of `width` bytes each) from the memory area pointed to by `src` to the memory area pointed to by `dst`, where `kind` specifies the direction of the copy, and must be one of :py:obj:`~.cudaMemcpyHostToHost`, :py:obj:`~.cudaMemcpyHostToDevice`, :py:obj:`~.cudaMemcpyDeviceToHost`, :py:obj:`~.cudaMemcpyDeviceToDevice`, or :py:obj:`~.cudaMemcpyDefault`. Passing :py:obj:`~.cudaMemcpyDefault` is recommended, in which case the type of transfer is inferred from the pointer values. However, :py:obj:`~.cudaMemcpyDefault` is only allowed on systems that support unified virtual addressing. `dpitch` and `spitch` are the widths in memory in bytes of the 2D arrays pointed to by `dst` and `src`, including any padding added to the end of each row. The memory areas may not overlap. `width` must not exceed either `dpitch` or `spitch`. Calling :py:obj:`~.cudaMemcpy2D()` with `dst` and `src` pointers that do not match the direction of the copy results in an undefined behavior. :py:obj:`~.cudaMemcpy2D()` returns an error if `dpitch` or `spitch` exceeds the maximum allowed. Parameters ---------- dst : Any Destination memory address dpitch : size_t Pitch of destination memory src : Any Source memory address spitch : size_t Pitch of source memory width : size_t Width of matrix transfer (columns in bytes) height : size_t Height of matrix transfer (rows) kind : :py:obj:`~.cudaMemcpyKind` Type of transfer Returns ------- cudaError_t :py:obj:`~.cudaSuccess`, :py:obj:`~.cudaErrorInvalidValue`, :py:obj:`~.cudaErrorInvalidPitchValue`, :py:obj:`~.cudaErrorInvalidMemcpyDirection` See Also -------- :py:obj:`~.cudaMemcpy`, :py:obj:`~.cudaMemcpy2DToArray`, :py:obj:`~.cudaMemcpy2DFromArray`, :py:obj:`~.cudaMemcpy2DArrayToArray`, :py:obj:`~.cudaMemcpyToSymbol`, :py:obj:`~.cudaMemcpyFromSymbol`, :py:obj:`~.cudaMemcpyAsync`, :py:obj:`~.cudaMemcpy2DAsync`, :py:obj:`~.cudaMemcpy2DToArrayAsync`, :py:obj:`~.cudaMemcpy2DFromArrayAsync`, :py:obj:`~.cudaMemcpyToSymbolAsync`, :py:obj:`~.cudaMemcpyFromSymbolAsync`, :py:obj:`~.cuMemcpy2D`, :py:obj:`~.cuMemcpy2DUnaligned` """ cydst = _HelperInputVoidPtr(dst) cdef void* cydst_ptr = cydst.cptr cysrc = _HelperInputVoidPtr(src) cdef void* cysrc_ptr = cysrc.cptr cdef cyruntime.cudaMemcpyKind cykind = kind.value with nogil: err = cyruntime.cudaMemcpy2D(cydst_ptr, dpitch, cysrc_ptr, spitch, width, height, cykind) return (_dict_cudaError_t[err],) @cython.embedsignature(True) def cudaMemcpy2DToArray(dst, size_t wOffset, size_t hOffset, src, size_t spitch, size_t width, size_t height, kind not None : cudaMemcpyKind): """ Copies data between host and device. Copies a matrix (`height` rows of `width` bytes each) from the memory area pointed to by `src` to the CUDA array `dst` starting at `hOffset` rows and `wOffset` bytes from the upper left corner, where `kind` specifies the direction of the copy, and must be one of :py:obj:`~.cudaMemcpyHostToHost`, :py:obj:`~.cudaMemcpyHostToDevice`, :py:obj:`~.cudaMemcpyDeviceToHost`, :py:obj:`~.cudaMemcpyDeviceToDevice`, or :py:obj:`~.cudaMemcpyDefault`. Passing :py:obj:`~.cudaMemcpyDefault` is recommended, in which case the type of transfer is inferred from the pointer values. However, :py:obj:`~.cudaMemcpyDefault` is only allowed on systems that support unified virtual addressing. `spitch` is the width in memory in bytes of the 2D array pointed to by `src`, including any padding added to the end of each row. `wOffset` + `width` must not exceed the width of the CUDA array `dst`. `width` must not exceed `spitch`. :py:obj:`~.cudaMemcpy2DToArray()` returns an error if `spitch` exceeds the maximum allowed. Parameters ---------- dst : :py:obj:`~.cudaArray_t` Destination memory address wOffset : size_t Destination starting X offset (columns in bytes) hOffset : size_t Destination starting Y offset (rows) src : Any Source memory address spitch : size_t Pitch of source memory width : size_t Width of matrix transfer (columns in bytes) height : size_t Height of matrix transfer (rows) kind : :py:obj:`~.cudaMemcpyKind` Type of transfer Returns ------- cudaError_t :py:obj:`~.cudaSuccess`, :py:obj:`~.cudaErrorInvalidValue`, :py:obj:`~.cudaErrorInvalidPitchValue`, :py:obj:`~.cudaErrorInvalidMemcpyDirection` See Also -------- :py:obj:`~.cudaMemcpy`, :py:obj:`~.cudaMemcpy2D`, :py:obj:`~.cudaMemcpy2DFromArray`, :py:obj:`~.cudaMemcpy2DArrayToArray`, :py:obj:`~.cudaMemcpyToSymbol`, :py:obj:`~.cudaMemcpyFromSymbol`, :py:obj:`~.cudaMemcpyAsync`, :py:obj:`~.cudaMemcpy2DAsync`, :py:obj:`~.cudaMemcpy2DToArrayAsync`, :py:obj:`~.cudaMemcpy2DFromArrayAsync`, :py:obj:`~.cudaMemcpyToSymbolAsync`, :py:obj:`~.cudaMemcpyFromSymbolAsync`, :py:obj:`~.cuMemcpy2D`, :py:obj:`~.cuMemcpy2DUnaligned` """ cdef cyruntime.cudaArray_t cydst if dst is None: pdst = 0 elif isinstance(dst, (cudaArray_t,)): pdst = int(dst) else: pdst = int(cudaArray_t(dst)) cydst = pdst cysrc = _HelperInputVoidPtr(src) cdef void* cysrc_ptr = cysrc.cptr cdef cyruntime.cudaMemcpyKind cykind = kind.value with nogil: err = cyruntime.cudaMemcpy2DToArray(cydst, wOffset, hOffset, cysrc_ptr, spitch, width, height, cykind) return (_dict_cudaError_t[err],) @cython.embedsignature(True) def cudaMemcpy2DFromArray(dst, size_t dpitch, src, size_t wOffset, size_t hOffset, size_t width, size_t height, kind not None : cudaMemcpyKind): """ Copies data between host and device. Copies a matrix (`height` rows of `width` bytes each) from the CUDA array `src` starting at `hOffset` rows and `wOffset` bytes from the upper left corner to the memory area pointed to by `dst`, where `kind` specifies the direction of the copy, and must be one of :py:obj:`~.cudaMemcpyHostToHost`, :py:obj:`~.cudaMemcpyHostToDevice`, :py:obj:`~.cudaMemcpyDeviceToHost`, :py:obj:`~.cudaMemcpyDeviceToDevice`, or :py:obj:`~.cudaMemcpyDefault`. Passing :py:obj:`~.cudaMemcpyDefault` is recommended, in which case the type of transfer is inferred from the pointer values. However, :py:obj:`~.cudaMemcpyDefault` is only allowed on systems that support unified virtual addressing. `dpitch` is the width in memory in bytes of the 2D array pointed to by `dst`, including any padding added to the end of each row. `wOffset` + `width` must not exceed the width of the CUDA array `src`. `width` must not exceed `dpitch`. :py:obj:`~.cudaMemcpy2DFromArray()` returns an error if `dpitch` exceeds the maximum allowed. Parameters ---------- dst : Any Destination memory address dpitch : size_t Pitch of destination memory src : :py:obj:`~.cudaArray_const_t` Source memory address wOffset : size_t Source starting X offset (columns in bytes) hOffset : size_t Source starting Y offset (rows) width : size_t Width of matrix transfer (columns in bytes) height : size_t Height of matrix transfer (rows) kind : :py:obj:`~.cudaMemcpyKind` Type of transfer Returns ------- cudaError_t :py:obj:`~.cudaSuccess`, :py:obj:`~.cudaErrorInvalidValue`, :py:obj:`~.cudaErrorInvalidPitchValue`, :py:obj:`~.cudaErrorInvalidMemcpyDirection` See Also -------- :py:obj:`~.cudaMemcpy`, :py:obj:`~.cudaMemcpy2D`, :py:obj:`~.cudaMemcpy2DToArray`, :py:obj:`~.cudaMemcpy2DArrayToArray`, :py:obj:`~.cudaMemcpyToSymbol`, :py:obj:`~.cudaMemcpyFromSymbol`, :py:obj:`~.cudaMemcpyAsync`, :py:obj:`~.cudaMemcpy2DAsync`, :py:obj:`~.cudaMemcpy2DToArrayAsync`, :py:obj:`~.cudaMemcpy2DFromArrayAsync`, :py:obj:`~.cudaMemcpyToSymbolAsync`, :py:obj:`~.cudaMemcpyFromSymbolAsync`, :py:obj:`~.cuMemcpy2D`, :py:obj:`~.cuMemcpy2DUnaligned` """ cdef cyruntime.cudaArray_const_t cysrc if src is None: psrc = 0 elif isinstance(src, (cudaArray_const_t,)): psrc = int(src) else: psrc = int(cudaArray_const_t(src)) cysrc = psrc cydst = _HelperInputVoidPtr(dst) cdef void* cydst_ptr = cydst.cptr cdef cyruntime.cudaMemcpyKind cykind = kind.value with nogil: err = cyruntime.cudaMemcpy2DFromArray(cydst_ptr, dpitch, cysrc, wOffset, hOffset, width, height, cykind) return (_dict_cudaError_t[err],) @cython.embedsignature(True) def cudaMemcpy2DArrayToArray(dst, size_t wOffsetDst, size_t hOffsetDst, src, size_t wOffsetSrc, size_t hOffsetSrc, size_t width, size_t height, kind not None : cudaMemcpyKind): """ Copies data between host and device. Copies a matrix (`height` rows of `width` bytes each) from the CUDA array `src` starting at `hOffsetSrc` rows and `wOffsetSrc` bytes from the upper left corner to the CUDA array `dst` starting at `hOffsetDst` rows and `wOffsetDst` bytes from the upper left corner, where `kind` specifies the direction of the copy, and must be one of :py:obj:`~.cudaMemcpyHostToHost`, :py:obj:`~.cudaMemcpyHostToDevice`, :py:obj:`~.cudaMemcpyDeviceToHost`, :py:obj:`~.cudaMemcpyDeviceToDevice`, or :py:obj:`~.cudaMemcpyDefault`. Passing :py:obj:`~.cudaMemcpyDefault` is recommended, in which case the type of transfer is inferred from the pointer values. However, :py:obj:`~.cudaMemcpyDefault` is only allowed on systems that support unified virtual addressing. `wOffsetDst` + `width` must not exceed the width of the CUDA array `dst`. `wOffsetSrc` + `width` must not exceed the width of the CUDA array `src`. Parameters ---------- dst : :py:obj:`~.cudaArray_t` Destination memory address wOffsetDst : size_t Destination starting X offset (columns in bytes) hOffsetDst : size_t Destination starting Y offset (rows) src : :py:obj:`~.cudaArray_const_t` Source memory address wOffsetSrc : size_t Source starting X offset (columns in bytes) hOffsetSrc : size_t Source starting Y offset (rows) width : size_t Width of matrix transfer (columns in bytes) height : size_t Height of matrix transfer (rows) kind : :py:obj:`~.cudaMemcpyKind` Type of transfer Returns ------- cudaError_t :py:obj:`~.cudaSuccess`, :py:obj:`~.cudaErrorInvalidValue`, :py:obj:`~.cudaErrorInvalidMemcpyDirection` See Also -------- :py:obj:`~.cudaMemcpy`, :py:obj:`~.cudaMemcpy2D`, :py:obj:`~.cudaMemcpy2DToArray`, :py:obj:`~.cudaMemcpy2DFromArray`, :py:obj:`~.cudaMemcpyToSymbol`, :py:obj:`~.cudaMemcpyFromSymbol`, :py:obj:`~.cudaMemcpyAsync`, :py:obj:`~.cudaMemcpy2DAsync`, :py:obj:`~.cudaMemcpy2DToArrayAsync`, :py:obj:`~.cudaMemcpy2DFromArrayAsync`, :py:obj:`~.cudaMemcpyToSymbolAsync`, :py:obj:`~.cudaMemcpyFromSymbolAsync`, :py:obj:`~.cuMemcpy2D`, :py:obj:`~.cuMemcpy2DUnaligned` """ cdef cyruntime.cudaArray_const_t cysrc if src is None: psrc = 0 elif isinstance(src, (cudaArray_const_t,)): psrc = int(src) else: psrc = int(cudaArray_const_t(src)) cysrc = psrc cdef cyruntime.cudaArray_t cydst if dst is None: pdst = 0 elif isinstance(dst, (cudaArray_t,)): pdst = int(dst) else: pdst = int(cudaArray_t(dst)) cydst = pdst cdef cyruntime.cudaMemcpyKind cykind = kind.value with nogil: err = cyruntime.cudaMemcpy2DArrayToArray(cydst, wOffsetDst, hOffsetDst, cysrc, wOffsetSrc, hOffsetSrc, width, height, cykind) return (_dict_cudaError_t[err],) @cython.embedsignature(True) def cudaMemcpyAsync(dst, src, size_t count, kind not None : cudaMemcpyKind, stream): """ Copies data between host and device. Copies `count` bytes from the memory area pointed to by `src` to the memory area pointed to by `dst`, where `kind` specifies the direction of the copy, and must be one of :py:obj:`~.cudaMemcpyHostToHost`, :py:obj:`~.cudaMemcpyHostToDevice`, :py:obj:`~.cudaMemcpyDeviceToHost`, :py:obj:`~.cudaMemcpyDeviceToDevice`, or :py:obj:`~.cudaMemcpyDefault`. Passing :py:obj:`~.cudaMemcpyDefault` is recommended, in which case the type of transfer is inferred from the pointer values. However, :py:obj:`~.cudaMemcpyDefault` is only allowed on systems that support unified virtual addressing. The memory areas may not overlap. Calling :py:obj:`~.cudaMemcpyAsync()` with `dst` and `src` pointers that do not match the direction of the copy results in an undefined behavior. :py:obj:`~.cudaMemcpyAsync()` is asynchronous with respect to the host, so the call may return before the copy is complete. The copy can optionally be associated to a stream by passing a non-zero `stream` argument. If `kind` is :py:obj:`~.cudaMemcpyHostToDevice` or :py:obj:`~.cudaMemcpyDeviceToHost` and the `stream` is non-zero, the copy may overlap with operations in other streams. The device version of this function only handles device to device copies and cannot be given local or shared pointers. Parameters ---------- dst : Any Destination memory address src : Any Source memory address count : size_t Size in bytes to copy kind : :py:obj:`~.cudaMemcpyKind` Type of transfer stream : :py:obj:`~.CUstream` or :py:obj:`~.cudaStream_t` Stream identifier Returns ------- cudaError_t :py:obj:`~.cudaSuccess`, :py:obj:`~.cudaErrorInvalidValue`, :py:obj:`~.cudaErrorInvalidMemcpyDirection` See Also -------- :py:obj:`~.cudaMemcpy`, :py:obj:`~.cudaMemcpy2D`, :py:obj:`~.cudaMemcpy2DToArray`, :py:obj:`~.cudaMemcpy2DFromArray`, :py:obj:`~.cudaMemcpy2DArrayToArray`, :py:obj:`~.cudaMemcpyToSymbol`, :py:obj:`~.cudaMemcpyFromSymbol`, :py:obj:`~.cudaMemcpy2DAsync`, :py:obj:`~.cudaMemcpy2DToArrayAsync`, :py:obj:`~.cudaMemcpy2DFromArrayAsync`, :py:obj:`~.cudaMemcpyToSymbolAsync`, :py:obj:`~.cudaMemcpyFromSymbolAsync`, :py:obj:`~.cuMemcpyAsync`, :py:obj:`~.cuMemcpyDtoHAsync`, :py:obj:`~.cuMemcpyHtoDAsync`, :py:obj:`~.cuMemcpyDtoDAsync` """ cdef cyruntime.cudaStream_t cystream if stream is None: pstream = 0 elif isinstance(stream, (cudaStream_t,driver.CUstream)): pstream = int(stream) else: pstream = int(cudaStream_t(stream)) cystream = pstream cydst = _HelperInputVoidPtr(dst) cdef void* cydst_ptr = cydst.cptr cysrc = _HelperInputVoidPtr(src) cdef void* cysrc_ptr = cysrc.cptr cdef cyruntime.cudaMemcpyKind cykind = kind.value with nogil: err = cyruntime.cudaMemcpyAsync(cydst_ptr, cysrc_ptr, count, cykind, cystream) return (_dict_cudaError_t[err],) @cython.embedsignature(True) def cudaMemcpyPeerAsync(dst, int dstDevice, src, int srcDevice, size_t count, stream): """ Copies memory between two devices asynchronously. Copies memory from one device to memory on another device. `dst` is the base device pointer of the destination memory and `dstDevice` is the destination device. `src` is the base device pointer of the source memory and `srcDevice` is the source device. `count` specifies the number of bytes to copy. Note that this function is asynchronous with respect to the host and all work on other devices. Parameters ---------- dst : Any Destination device pointer dstDevice : int Destination device src : Any Source device pointer srcDevice : int Source device count : size_t Size of memory copy in bytes stream : :py:obj:`~.CUstream` or :py:obj:`~.cudaStream_t` Stream identifier Returns ------- cudaError_t :py:obj:`~.cudaSuccess`, :py:obj:`~.cudaErrorInvalidValue`, :py:obj:`~.cudaErrorInvalidDevice` See Also -------- :py:obj:`~.cudaMemcpy`, :py:obj:`~.cudaMemcpyPeer`, :py:obj:`~.cudaMemcpyAsync`, :py:obj:`~.cudaMemcpy3DPeerAsync`, :py:obj:`~.cuMemcpyPeerAsync` """ cdef cyruntime.cudaStream_t cystream if stream is None: pstream = 0 elif isinstance(stream, (cudaStream_t,driver.CUstream)): pstream = int(stream) else: pstream = int(cudaStream_t(stream)) cystream = pstream cydst = _HelperInputVoidPtr(dst) cdef void* cydst_ptr = cydst.cptr cysrc = _HelperInputVoidPtr(src) cdef void* cysrc_ptr = cysrc.cptr with nogil: err = cyruntime.cudaMemcpyPeerAsync(cydst_ptr, dstDevice, cysrc_ptr, srcDevice, count, cystream) return (_dict_cudaError_t[err],) @cython.embedsignature(True) def cudaMemcpyBatchAsync(dsts : Optional[tuple[Any] | list[Any]], srcs : Optional[tuple[Any] | list[Any]], sizes : tuple[int] | list[int], size_t count, attrs : Optional[tuple[cudaMemcpyAttributes] | list[cudaMemcpyAttributes]], attrsIdxs : tuple[int] | list[int], size_t numAttrs, stream): """ Performs a batch of memory copies asynchronously. Performs a batch of memory copies. The batch as a whole executes in stream order but copies within a batch are not guaranteed to execute in any specific order. This API only supports pointer-to-pointer copies. For copies involving CUDA arrays, please see :py:obj:`~.cudaMemcpy3DBatchAsync`. Performs memory copies from source buffers specified in `srcs` to destination buffers specified in `dsts`. The size of each copy is specified in `sizes`. All three arrays must be of the same length as specified by `count`. Since there are no ordering guarantees for copies within a batch, specifying any dependent copies within a batch will result in undefined behavior. Every copy in the batch has to be associated with a set of attributes specified in the `attrs` array. Each entry in this array can apply to more than one copy. This can be done by specifying in the `attrsIdxs` array, the index of the first copy that the corresponding entry in the `attrs` array applies to. Both `attrs` and `attrsIdxs` must be of the same length as specified by `numAttrs`. For example, if a batch has 10 copies listed in dst/src/sizes, the first 6 of which have one set of attributes and the remaining 4 another, then `numAttrs` will be 2, `attrsIdxs` will be {0, 6} and `attrs` will contains the two sets of attributes. Note that the first entry in `attrsIdxs` must always be 0. Also, each entry must be greater than the previous entry and the last entry should be less than `count`. Furthermore, `numAttrs` must be lesser than or equal to `count`. The :py:obj:`~.cudaMemcpyAttributes.srcAccessOrder` indicates the source access ordering to be observed for copies associated with the attribute. If the source access order is set to :py:obj:`~.cudaMemcpySrcAccessOrderStream`, then the source will be accessed in stream order. If the source access order is set to :py:obj:`~.cudaMemcpySrcAccessOrderDuringApiCall` then it indicates that access to the source pointer can be out of stream order and all accesses must be complete before the API call returns. This flag is suited for ephemeral sources (ex., stack variables) when it's known that no prior operations in the stream can be accessing the memory and also that the lifetime of the memory is limited to the scope that the source variable was declared in. Specifying this flag allows the driver to optimize the copy and removes the need for the user to synchronize the stream after the API call. If the source access order is set to :py:obj:`~.cudaMemcpySrcAccessOrderAny` then it indicates that access to the source pointer can be out of stream order and the accesses can happen even after the API call returns. This flag is suited for host pointers allocated outside CUDA (ex., via malloc) when it's known that no prior operations in the stream can be accessing the memory. Specifying this flag allows the driver to optimize the copy on certain platforms. Each memcpy operation in the batch must have a valid :py:obj:`~.cudaMemcpyAttributes` corresponding to it including the appropriate srcAccessOrder setting, otherwise the API will return :py:obj:`~.cudaErrorInvalidValue`. The :py:obj:`~.cudaMemcpyAttributes.srcLocHint` and :py:obj:`~.cudaMemcpyAttributes.dstLocHint` allows applications to specify hint locations for operands of a copy when the operand doesn't have a fixed location. That is, these hints are only applicable for managed memory pointers on devices where :py:obj:`~.cudaDevAttrConcurrentManagedAccess` is true or system- allocated pageable memory on devices where :py:obj:`~.cudaDevAttrPageableMemoryAccess` is true. For other cases, these hints are ignored. The :py:obj:`~.cudaMemcpyAttributes.flags` field can be used to specify certain flags for copies. Setting the :py:obj:`~.cudaMemcpyFlagPreferOverlapWithCompute` flag indicates that the associated copies should preferably overlap with any compute work. Note that this flag is a hint and can be ignored depending on the platform and other parameters of the copy. If any error is encountered while parsing the batch, the index within the batch where the error was encountered will be returned in `failIdx`. Parameters ---------- dsts : list[Any] Array of destination pointers. srcs : list[Any] Array of memcpy source pointers. sizes : list[int] Array of sizes for memcpy operations. count : size_t Size of `dsts`, `srcs` and `sizes` arrays attrs : list[:py:obj:`~.cudaMemcpyAttributes`] Array of memcpy attributes. attrsIdxs : list[int] Array of indices to specify which copies each entry in the `attrs` array applies to. The attributes specified in attrs[k] will be applied to copies starting from attrsIdxs[k] through attrsIdxs[k+1] - 1. Also attrs[numAttrs-1] will apply to copies starting from attrsIdxs[numAttrs-1] through count - 1. numAttrs : size_t Size of `attrs` and `attrsIdxs` arrays. hStream : :py:obj:`~.CUstream` or :py:obj:`~.cudaStream_t` The stream to enqueue the operations in. Must not be legacy NULL stream. Returns ------- cudaError_t :py:obj:`~.cudaSuccess` :py:obj:`~.cudaErrorInvalidValue` failIdx : int Pointer to a location to return the index of the copy where a failure was encountered. The value will be SIZE_MAX if the error doesn't pertain to any specific copy. """ cdef cyruntime.cudaStream_t cystream if stream is None: pstream = 0 elif isinstance(stream, (cudaStream_t,driver.CUstream)): pstream = int(stream) else: pstream = int(cudaStream_t(stream)) cystream = pstream if not all(isinstance(_x, (int)) for _x in attrsIdxs): raise TypeError("Argument 'attrsIdxs' is not instance of type (expected tuple[int] or list[int]") attrs = [] if attrs is None else attrs if not all(isinstance(_x, (cudaMemcpyAttributes,)) for _x in attrs): raise TypeError("Argument 'attrs' is not instance of type (expected tuple[cyruntime.cudaMemcpyAttributes,] or list[cyruntime.cudaMemcpyAttributes,]") if not all(isinstance(_x, (int)) for _x in sizes): raise TypeError("Argument 'sizes' is not instance of type (expected tuple[int] or list[int]") srcs = [] if srcs is None else srcs dsts = [] if dsts is None else dsts pylist = [_HelperInputVoidPtr(pydsts) for pydsts in dsts] cdef _InputVoidPtrPtrHelper voidStarHelperdsts = _InputVoidPtrPtrHelper(pylist) cdef void** cydsts_ptr = voidStarHelperdsts.cptr pylist = [_HelperInputVoidPtr(pysrcs) for pysrcs in srcs] cdef _InputVoidPtrPtrHelper voidStarHelpersrcs = _InputVoidPtrPtrHelper(pylist) cdef void** cysrcs_ptr = voidStarHelpersrcs.cptr cdef vector[size_t] cysizes = sizes if count > len(dsts): raise RuntimeError("List is too small: " + str(len(dsts)) + " < " + str(count)) if count > len(srcs): raise RuntimeError("List is too small: " + str(len(srcs)) + " < " + str(count)) if count > len(sizes): raise RuntimeError("List is too small: " + str(len(sizes)) + " < " + str(count)) cdef cyruntime.cudaMemcpyAttributes* cyattrs = NULL if len(attrs) > 1: cyattrs = calloc(len(attrs), sizeof(cyruntime.cudaMemcpyAttributes)) if cyattrs is NULL: raise MemoryError('Failed to allocate length x size memory: ' + str(len(attrs)) + 'x' + str(sizeof(cyruntime.cudaMemcpyAttributes))) for idx in range(len(attrs)): string.memcpy(&cyattrs[idx], (attrs[idx])._pvt_ptr, sizeof(cyruntime.cudaMemcpyAttributes)) elif len(attrs) == 1: cyattrs = (attrs[0])._pvt_ptr cdef vector[size_t] cyattrsIdxs = attrsIdxs if numAttrs > len(attrs): raise RuntimeError("List is too small: " + str(len(attrs)) + " < " + str(numAttrs)) if numAttrs > len(attrsIdxs): raise RuntimeError("List is too small: " + str(len(attrsIdxs)) + " < " + str(numAttrs)) cdef size_t failIdx = 0 with nogil: err = cyruntime.cudaMemcpyBatchAsync(cydsts_ptr, cysrcs_ptr, cysizes.data(), count, cyattrs, cyattrsIdxs.data(), numAttrs, &failIdx, cystream) if len(attrs) > 1 and cyattrs is not NULL: free(cyattrs) if err != cyruntime.cudaSuccess: return (_dict_cudaError_t[err], None) return (_dict_cudaError_t[err], failIdx) @cython.embedsignature(True) def cudaMemcpy3DBatchAsync(size_t numOps, opList : Optional[tuple[cudaMemcpy3DBatchOp] | list[cudaMemcpy3DBatchOp]], unsigned long long flags, stream): """ Performs a batch of 3D memory copies asynchronously. Performs a batch of memory copies. The batch as a whole executes in stream order but copies within a batch are not guaranteed to execute in any specific order. Note that this means specifying any dependent copies within a batch will result in undefined behavior. Performs memory copies as specified in the `opList` array. The length of this array is specified in `numOps`. Each entry in this array describes a copy operation. This includes among other things, the source and destination operands for the copy as specified in :py:obj:`~.cudaMemcpy3DBatchOp.src` and :py:obj:`~.cudaMemcpy3DBatchOp.dst` respectively. The source and destination operands of a copy can either be a pointer or a CUDA array. The width, height and depth of a copy is specified in :py:obj:`~.cudaMemcpy3DBatchOp.extent`. The width, height and depth of a copy are specified in elements and must not be zero. For pointer-to- pointer copies, the element size is considered to be 1. For pointer to CUDA array or vice versa copies, the element size is determined by the CUDA array. For CUDA array to CUDA array copies, the element size of the two CUDA arrays must match. For a given operand, if :py:obj:`~.cudaMemcpy3DOperand`::type is specified as :py:obj:`~.cudaMemcpyOperandTypePointer`, then :py:obj:`~.cudaMemcpy3DOperand`::op::ptr will be used. The :py:obj:`~.cudaMemcpy3DOperand`::op::ptr::ptr field must contain the pointer where the copy should begin. The :py:obj:`~.cudaMemcpy3DOperand`::op::ptr::rowLength field specifies the length of each row in elements and must either be zero or be greater than or equal to the width of the copy specified in :py:obj:`~.cudaMemcpy3DBatchOp`::extent::width. The :py:obj:`~.cudaMemcpy3DOperand`::op::ptr::layerHeight field specifies the height of each layer and must either be zero or be greater than or equal to the height of the copy specified in :py:obj:`~.cudaMemcpy3DBatchOp`::extent::height. When either of these values is zero, that aspect of the operand is considered to be tightly packed according to the copy extent. For managed memory pointers on devices where :py:obj:`~.cudaDevAttrConcurrentManagedAccess` is true or system-allocated pageable memory on devices where :py:obj:`~.cudaDevAttrPageableMemoryAccess` is true, the :py:obj:`~.cudaMemcpy3DOperand`::op::ptr::locHint field can be used to hint the location of the operand. If an operand's type is specified as :py:obj:`~.cudaMemcpyOperandTypeArray`, then :py:obj:`~.cudaMemcpy3DOperand`::op::array will be used. The :py:obj:`~.cudaMemcpy3DOperand`::op::array::array field specifies the CUDA array and :py:obj:`~.cudaMemcpy3DOperand`::op::array::offset specifies the 3D offset into that array where the copy begins. The :py:obj:`~.cudaMemcpyAttributes.srcAccessOrder` indicates the source access ordering to be observed for copies associated with the attribute. If the source access order is set to :py:obj:`~.cudaMemcpySrcAccessOrderStream`, then the source will be accessed in stream order. If the source access order is set to :py:obj:`~.cudaMemcpySrcAccessOrderDuringApiCall` then it indicates that access to the source pointer can be out of stream order and all accesses must be complete before the API call returns. This flag is suited for ephemeral sources (ex., stack variables) when it's known that no prior operations in the stream can be accessing the memory and also that the lifetime of the memory is limited to the scope that the source variable was declared in. Specifying this flag allows the driver to optimize the copy and removes the need for the user to synchronize the stream after the API call. If the source access order is set to :py:obj:`~.cudaMemcpySrcAccessOrderAny` then it indicates that access to the source pointer can be out of stream order and the accesses can happen even after the API call returns. This flag is suited for host pointers allocated outside CUDA (ex., via malloc) when it's known that no prior operations in the stream can be accessing the memory. Specifying this flag allows the driver to optimize the copy on certain platforms. Each memcopy operation in `opList` must have a valid srcAccessOrder setting, otherwise this API will return :py:obj:`~.cudaErrorInvalidValue`. The :py:obj:`~.cudaMemcpyAttributes.flags` field can be used to specify certain flags for copies. Setting the :py:obj:`~.cudaMemcpyFlagPreferOverlapWithCompute` flag indicates that the associated copies should preferably overlap with any compute work. Note that this flag is a hint and can be ignored depending on the platform and other parameters of the copy. If any error is encountered while parsing the batch, the index within the batch where the error was encountered will be returned in `failIdx`. Parameters ---------- numOps : size_t Total number of memcpy operations. opList : list[:py:obj:`~.cudaMemcpy3DBatchOp`] Array of size `numOps` containing the actual memcpy operations. flags : unsigned long long Flags for future use, must be zero now. hStream : :py:obj:`~.CUstream` or :py:obj:`~.cudaStream_t` The stream to enqueue the operations in. Must not be default NULL stream. Returns ------- cudaError_t :py:obj:`~.cudaSuccess` :py:obj:`~.cudaErrorInvalidValue` failIdx : int Pointer to a location to return the index of the copy where a failure was encountered. The value will be SIZE_MAX if the error doesn't pertain to any specific copy. """ cdef cyruntime.cudaStream_t cystream if stream is None: pstream = 0 elif isinstance(stream, (cudaStream_t,driver.CUstream)): pstream = int(stream) else: pstream = int(cudaStream_t(stream)) cystream = pstream opList = [] if opList is None else opList if not all(isinstance(_x, (cudaMemcpy3DBatchOp,)) for _x in opList): raise TypeError("Argument 'opList' is not instance of type (expected tuple[cyruntime.cudaMemcpy3DBatchOp,] or list[cyruntime.cudaMemcpy3DBatchOp,]") if numOps > len(opList): raise RuntimeError("List is too small: " + str(len(opList)) + " < " + str(numOps)) cdef cyruntime.cudaMemcpy3DBatchOp* cyopList = NULL if len(opList) > 1: cyopList = calloc(len(opList), sizeof(cyruntime.cudaMemcpy3DBatchOp)) if cyopList is NULL: raise MemoryError('Failed to allocate length x size memory: ' + str(len(opList)) + 'x' + str(sizeof(cyruntime.cudaMemcpy3DBatchOp))) for idx in range(len(opList)): string.memcpy(&cyopList[idx], (opList[idx])._pvt_ptr, sizeof(cyruntime.cudaMemcpy3DBatchOp)) elif len(opList) == 1: cyopList = (opList[0])._pvt_ptr cdef size_t failIdx = 0 with nogil: err = cyruntime.cudaMemcpy3DBatchAsync(numOps, cyopList, &failIdx, flags, cystream) if len(opList) > 1 and cyopList is not NULL: free(cyopList) if err != cyruntime.cudaSuccess: return (_dict_cudaError_t[err], None) return (_dict_cudaError_t[err], failIdx) @cython.embedsignature(True) def cudaMemcpy2DAsync(dst, size_t dpitch, src, size_t spitch, size_t width, size_t height, kind not None : cudaMemcpyKind, stream): """ Copies data between host and device. Copies a matrix (`height` rows of `width` bytes each) from the memory area pointed to by `src` to the memory area pointed to by `dst`, where `kind` specifies the direction of the copy, and must be one of :py:obj:`~.cudaMemcpyHostToHost`, :py:obj:`~.cudaMemcpyHostToDevice`, :py:obj:`~.cudaMemcpyDeviceToHost`, :py:obj:`~.cudaMemcpyDeviceToDevice`, or :py:obj:`~.cudaMemcpyDefault`. Passing :py:obj:`~.cudaMemcpyDefault` is recommended, in which case the type of transfer is inferred from the pointer values. However, :py:obj:`~.cudaMemcpyDefault` is only allowed on systems that support unified virtual addressing. `dpitch` and `spitch` are the widths in memory in bytes of the 2D arrays pointed to by `dst` and `src`, including any padding added to the end of each row. The memory areas may not overlap. `width` must not exceed either `dpitch` or `spitch`. Calling :py:obj:`~.cudaMemcpy2DAsync()` with `dst` and `src` pointers that do not match the direction of the copy results in an undefined behavior. :py:obj:`~.cudaMemcpy2DAsync()` returns an error if `dpitch` or `spitch` is greater than the maximum allowed. :py:obj:`~.cudaMemcpy2DAsync()` is asynchronous with respect to the host, so the call may return before the copy is complete. The copy can optionally be associated to a stream by passing a non-zero `stream` argument. If `kind` is :py:obj:`~.cudaMemcpyHostToDevice` or :py:obj:`~.cudaMemcpyDeviceToHost` and `stream` is non-zero, the copy may overlap with operations in other streams. The device version of this function only handles device to device copies and cannot be given local or shared pointers. Parameters ---------- dst : Any Destination memory address dpitch : size_t Pitch of destination memory src : Any Source memory address spitch : size_t Pitch of source memory width : size_t Width of matrix transfer (columns in bytes) height : size_t Height of matrix transfer (rows) kind : :py:obj:`~.cudaMemcpyKind` Type of transfer stream : :py:obj:`~.CUstream` or :py:obj:`~.cudaStream_t` Stream identifier Returns ------- cudaError_t :py:obj:`~.cudaSuccess`, :py:obj:`~.cudaErrorInvalidValue`, :py:obj:`~.cudaErrorInvalidPitchValue`, :py:obj:`~.cudaErrorInvalidMemcpyDirection` See Also -------- :py:obj:`~.cudaMemcpy`, :py:obj:`~.cudaMemcpy2D`, :py:obj:`~.cudaMemcpy2DToArray`, :py:obj:`~.cudaMemcpy2DFromArray`, :py:obj:`~.cudaMemcpy2DArrayToArray`, :py:obj:`~.cudaMemcpyToSymbol`, :py:obj:`~.cudaMemcpyFromSymbol`, :py:obj:`~.cudaMemcpyAsync`, :py:obj:`~.cudaMemcpy2DToArrayAsync`, :py:obj:`~.cudaMemcpy2DFromArrayAsync`, :py:obj:`~.cudaMemcpyToSymbolAsync`, :py:obj:`~.cudaMemcpyFromSymbolAsync`, :py:obj:`~.cuMemcpy2DAsync` """ cdef cyruntime.cudaStream_t cystream if stream is None: pstream = 0 elif isinstance(stream, (cudaStream_t,driver.CUstream)): pstream = int(stream) else: pstream = int(cudaStream_t(stream)) cystream = pstream cydst = _HelperInputVoidPtr(dst) cdef void* cydst_ptr = cydst.cptr cysrc = _HelperInputVoidPtr(src) cdef void* cysrc_ptr = cysrc.cptr cdef cyruntime.cudaMemcpyKind cykind = kind.value with nogil: err = cyruntime.cudaMemcpy2DAsync(cydst_ptr, dpitch, cysrc_ptr, spitch, width, height, cykind, cystream) return (_dict_cudaError_t[err],) @cython.embedsignature(True) def cudaMemcpy2DToArrayAsync(dst, size_t wOffset, size_t hOffset, src, size_t spitch, size_t width, size_t height, kind not None : cudaMemcpyKind, stream): """ Copies data between host and device. Copies a matrix (`height` rows of `width` bytes each) from the memory area pointed to by `src` to the CUDA array `dst` starting at `hOffset` rows and `wOffset` bytes from the upper left corner, where `kind` specifies the direction of the copy, and must be one of :py:obj:`~.cudaMemcpyHostToHost`, :py:obj:`~.cudaMemcpyHostToDevice`, :py:obj:`~.cudaMemcpyDeviceToHost`, :py:obj:`~.cudaMemcpyDeviceToDevice`, or :py:obj:`~.cudaMemcpyDefault`. Passing :py:obj:`~.cudaMemcpyDefault` is recommended, in which case the type of transfer is inferred from the pointer values. However, :py:obj:`~.cudaMemcpyDefault` is only allowed on systems that support unified virtual addressing. `spitch` is the width in memory in bytes of the 2D array pointed to by `src`, including any padding added to the end of each row. `wOffset` + `width` must not exceed the width of the CUDA array `dst`. `width` must not exceed `spitch`. :py:obj:`~.cudaMemcpy2DToArrayAsync()` returns an error if `spitch` exceeds the maximum allowed. :py:obj:`~.cudaMemcpy2DToArrayAsync()` is asynchronous with respect to the host, so the call may return before the copy is complete. The copy can optionally be associated to a stream by passing a non-zero `stream` argument. If `kind` is :py:obj:`~.cudaMemcpyHostToDevice` or :py:obj:`~.cudaMemcpyDeviceToHost` and `stream` is non-zero, the copy may overlap with operations in other streams. :py:obj:`~.cudaMemcpy2DFromArrayAsync`, :py:obj:`~.cudaMemcpyToSymbolAsync`, :py:obj:`~.cudaMemcpyFromSymbolAsync`, :py:obj:`~.cuMemcpy2DAsync` Parameters ---------- dst : :py:obj:`~.cudaArray_t` Destination memory address wOffset : size_t Destination starting X offset (columns in bytes) hOffset : size_t Destination starting Y offset (rows) src : Any Source memory address spitch : size_t Pitch of source memory width : size_t Width of matrix transfer (columns in bytes) height : size_t Height of matrix transfer (rows) kind : :py:obj:`~.cudaMemcpyKind` Type of transfer stream : :py:obj:`~.CUstream` or :py:obj:`~.cudaStream_t` Stream identifier Returns ------- cudaError_t :py:obj:`~.cudaSuccess`, :py:obj:`~.cudaErrorInvalidValue`, :py:obj:`~.cudaErrorInvalidPitchValue`, :py:obj:`~.cudaErrorInvalidMemcpyDirection` See Also -------- :py:obj:`~.cudaMemcpy`, :py:obj:`~.cudaMemcpy2D`, :py:obj:`~.cudaMemcpy2DToArray`, :py:obj:`~.cudaMemcpy2DFromArray`, :py:obj:`~.cudaMemcpy2DArrayToArray`, :py:obj:`~.cudaMemcpyToSymbol`, :py:obj:`~.cudaMemcpyFromSymbol`, :py:obj:`~.cudaMemcpyAsync`, :py:obj:`~.cudaMemcpy2DAsync`, """ cdef cyruntime.cudaStream_t cystream if stream is None: pstream = 0 elif isinstance(stream, (cudaStream_t,driver.CUstream)): pstream = int(stream) else: pstream = int(cudaStream_t(stream)) cystream = pstream cdef cyruntime.cudaArray_t cydst if dst is None: pdst = 0 elif isinstance(dst, (cudaArray_t,)): pdst = int(dst) else: pdst = int(cudaArray_t(dst)) cydst = pdst cysrc = _HelperInputVoidPtr(src) cdef void* cysrc_ptr = cysrc.cptr cdef cyruntime.cudaMemcpyKind cykind = kind.value with nogil: err = cyruntime.cudaMemcpy2DToArrayAsync(cydst, wOffset, hOffset, cysrc_ptr, spitch, width, height, cykind, cystream) return (_dict_cudaError_t[err],) @cython.embedsignature(True) def cudaMemcpy2DFromArrayAsync(dst, size_t dpitch, src, size_t wOffset, size_t hOffset, size_t width, size_t height, kind not None : cudaMemcpyKind, stream): """ Copies data between host and device. Copies a matrix (`height` rows of `width` bytes each) from the CUDA array `src` starting at `hOffset` rows and `wOffset` bytes from the upper left corner to the memory area pointed to by `dst`, where `kind` specifies the direction of the copy, and must be one of :py:obj:`~.cudaMemcpyHostToHost`, :py:obj:`~.cudaMemcpyHostToDevice`, :py:obj:`~.cudaMemcpyDeviceToHost`, :py:obj:`~.cudaMemcpyDeviceToDevice`, or :py:obj:`~.cudaMemcpyDefault`. Passing :py:obj:`~.cudaMemcpyDefault` is recommended, in which case the type of transfer is inferred from the pointer values. However, :py:obj:`~.cudaMemcpyDefault` is only allowed on systems that support unified virtual addressing. `dpitch` is the width in memory in bytes of the 2D array pointed to by `dst`, including any padding added to the end of each row. `wOffset` + `width` must not exceed the width of the CUDA array `src`. `width` must not exceed `dpitch`. :py:obj:`~.cudaMemcpy2DFromArrayAsync()` returns an error if `dpitch` exceeds the maximum allowed. :py:obj:`~.cudaMemcpy2DFromArrayAsync()` is asynchronous with respect to the host, so the call may return before the copy is complete. The copy can optionally be associated to a stream by passing a non-zero `stream` argument. If `kind` is :py:obj:`~.cudaMemcpyHostToDevice` or :py:obj:`~.cudaMemcpyDeviceToHost` and `stream` is non-zero, the copy may overlap with operations in other streams. :py:obj:`~.cudaMemcpyToSymbolAsync`, :py:obj:`~.cudaMemcpyFromSymbolAsync`, :py:obj:`~.cuMemcpy2DAsync` Parameters ---------- dst : Any Destination memory address dpitch : size_t Pitch of destination memory src : :py:obj:`~.cudaArray_const_t` Source memory address wOffset : size_t Source starting X offset (columns in bytes) hOffset : size_t Source starting Y offset (rows) width : size_t Width of matrix transfer (columns in bytes) height : size_t Height of matrix transfer (rows) kind : :py:obj:`~.cudaMemcpyKind` Type of transfer stream : :py:obj:`~.CUstream` or :py:obj:`~.cudaStream_t` Stream identifier Returns ------- cudaError_t :py:obj:`~.cudaSuccess`, :py:obj:`~.cudaErrorInvalidValue`, :py:obj:`~.cudaErrorInvalidPitchValue`, :py:obj:`~.cudaErrorInvalidMemcpyDirection` See Also -------- :py:obj:`~.cudaMemcpy`, :py:obj:`~.cudaMemcpy2D`, :py:obj:`~.cudaMemcpy2DToArray`, :py:obj:`~.cudaMemcpy2DFromArray`, :py:obj:`~.cudaMemcpy2DArrayToArray`, :py:obj:`~.cudaMemcpyToSymbol`, :py:obj:`~.cudaMemcpyFromSymbol`, :py:obj:`~.cudaMemcpyAsync`, :py:obj:`~.cudaMemcpy2DAsync`, :py:obj:`~.cudaMemcpy2DToArrayAsync`, """ cdef cyruntime.cudaStream_t cystream if stream is None: pstream = 0 elif isinstance(stream, (cudaStream_t,driver.CUstream)): pstream = int(stream) else: pstream = int(cudaStream_t(stream)) cystream = pstream cdef cyruntime.cudaArray_const_t cysrc if src is None: psrc = 0 elif isinstance(src, (cudaArray_const_t,)): psrc = int(src) else: psrc = int(cudaArray_const_t(src)) cysrc = psrc cydst = _HelperInputVoidPtr(dst) cdef void* cydst_ptr = cydst.cptr cdef cyruntime.cudaMemcpyKind cykind = kind.value with nogil: err = cyruntime.cudaMemcpy2DFromArrayAsync(cydst_ptr, dpitch, cysrc, wOffset, hOffset, width, height, cykind, cystream) return (_dict_cudaError_t[err],) @cython.embedsignature(True) def cudaMemset(devPtr, int value, size_t count): """ Initializes or sets device memory to a value. Fills the first `count` bytes of the memory area pointed to by `devPtr` with the constant byte value `value`. Note that this function is asynchronous with respect to the host unless `devPtr` refers to pinned host memory. Parameters ---------- devPtr : Any Pointer to device memory value : int Value to set for each byte of specified memory count : size_t Size in bytes to set Returns ------- cudaError_t :py:obj:`~.cudaSuccess`, :py:obj:`~.cudaErrorInvalidValue`, See Also -------- :py:obj:`~.cuMemsetD8`, :py:obj:`~.cuMemsetD16`, :py:obj:`~.cuMemsetD32` """ cydevPtr = _HelperInputVoidPtr(devPtr) cdef void* cydevPtr_ptr = cydevPtr.cptr with nogil: err = cyruntime.cudaMemset(cydevPtr_ptr, value, count) return (_dict_cudaError_t[err],) @cython.embedsignature(True) def cudaMemset2D(devPtr, size_t pitch, int value, size_t width, size_t height): """ Initializes or sets device memory to a value. Sets to the specified value `value` a matrix (`height` rows of `width` bytes each) pointed to by `dstPtr`. `pitch` is the width in bytes of the 2D array pointed to by `dstPtr`, including any padding added to the end of each row. This function performs fastest when the pitch is one that has been passed back by :py:obj:`~.cudaMallocPitch()`. Note that this function is asynchronous with respect to the host unless `devPtr` refers to pinned host memory. Parameters ---------- devPtr : Any Pointer to 2D device memory pitch : size_t Pitch in bytes of 2D device memory(Unused if `height` is 1) value : int Value to set for each byte of specified memory width : size_t Width of matrix set (columns in bytes) height : size_t Height of matrix set (rows) Returns ------- cudaError_t :py:obj:`~.cudaSuccess`, :py:obj:`~.cudaErrorInvalidValue`, See Also -------- :py:obj:`~.cudaMemset`, :py:obj:`~.cudaMemset3D`, :py:obj:`~.cudaMemsetAsync`, :py:obj:`~.cudaMemset2DAsync`, :py:obj:`~.cudaMemset3DAsync`, :py:obj:`~.cuMemsetD2D8`, :py:obj:`~.cuMemsetD2D16`, :py:obj:`~.cuMemsetD2D32` """ cydevPtr = _HelperInputVoidPtr(devPtr) cdef void* cydevPtr_ptr = cydevPtr.cptr with nogil: err = cyruntime.cudaMemset2D(cydevPtr_ptr, pitch, value, width, height) return (_dict_cudaError_t[err],) @cython.embedsignature(True) def cudaMemset3D(pitchedDevPtr not None : cudaPitchedPtr, int value, extent not None : cudaExtent): """ Initializes or sets device memory to a value. Initializes each element of a 3D array to the specified value `value`. The object to initialize is defined by `pitchedDevPtr`. The `pitch` field of `pitchedDevPtr` is the width in memory in bytes of the 3D array pointed to by `pitchedDevPtr`, including any padding added to the end of each row. The `xsize` field specifies the logical width of each row in bytes, while the `ysize` field specifies the height of each 2D slice in rows. The `pitch` field of `pitchedDevPtr` is ignored when `height` and `depth` are both equal to 1. The extents of the initialized region are specified as a `width` in bytes, a `height` in rows, and a `depth` in slices. Extents with `width` greater than or equal to the `xsize` of `pitchedDevPtr` may perform significantly faster than extents narrower than the `xsize`. Secondarily, extents with `height` equal to the `ysize` of `pitchedDevPtr` will perform faster than when the `height` is shorter than the `ysize`. This function performs fastest when the `pitchedDevPtr` has been allocated by :py:obj:`~.cudaMalloc3D()`. Note that this function is asynchronous with respect to the host unless `pitchedDevPtr` refers to pinned host memory. Parameters ---------- pitchedDevPtr : :py:obj:`~.cudaPitchedPtr` Pointer to pitched device memory value : int Value to set for each byte of specified memory extent : :py:obj:`~.cudaExtent` Size parameters for where to set device memory (`width` field in bytes) Returns ------- cudaError_t :py:obj:`~.cudaSuccess`, :py:obj:`~.cudaErrorInvalidValue`, See Also -------- :py:obj:`~.cudaMemset`, :py:obj:`~.cudaMemset2D`, :py:obj:`~.cudaMemsetAsync`, :py:obj:`~.cudaMemset2DAsync`, :py:obj:`~.cudaMemset3DAsync`, :py:obj:`~.cudaMalloc3D`, :py:obj:`~.make_cudaPitchedPtr`, :py:obj:`~.make_cudaExtent` """ with nogil: err = cyruntime.cudaMemset3D(pitchedDevPtr._pvt_ptr[0], value, extent._pvt_ptr[0]) return (_dict_cudaError_t[err],) @cython.embedsignature(True) def cudaMemsetAsync(devPtr, int value, size_t count, stream): """ Initializes or sets device memory to a value. Fills the first `count` bytes of the memory area pointed to by `devPtr` with the constant byte value `value`. :py:obj:`~.cudaMemsetAsync()` is asynchronous with respect to the host, so the call may return before the memset is complete. The operation can optionally be associated to a stream by passing a non-zero `stream` argument. If `stream` is non-zero, the operation may overlap with operations in other streams. The device version of this function only handles device to device copies and cannot be given local or shared pointers. Parameters ---------- devPtr : Any Pointer to device memory value : int Value to set for each byte of specified memory count : size_t Size in bytes to set stream : :py:obj:`~.CUstream` or :py:obj:`~.cudaStream_t` Stream identifier Returns ------- cudaError_t :py:obj:`~.cudaSuccess`, :py:obj:`~.cudaErrorInvalidValue`, See Also -------- :py:obj:`~.cudaMemset`, :py:obj:`~.cudaMemset2D`, :py:obj:`~.cudaMemset3D`, :py:obj:`~.cudaMemset2DAsync`, :py:obj:`~.cudaMemset3DAsync`, :py:obj:`~.cuMemsetD8Async`, :py:obj:`~.cuMemsetD16Async`, :py:obj:`~.cuMemsetD32Async` """ cdef cyruntime.cudaStream_t cystream if stream is None: pstream = 0 elif isinstance(stream, (cudaStream_t,driver.CUstream)): pstream = int(stream) else: pstream = int(cudaStream_t(stream)) cystream = pstream cydevPtr = _HelperInputVoidPtr(devPtr) cdef void* cydevPtr_ptr = cydevPtr.cptr with nogil: err = cyruntime.cudaMemsetAsync(cydevPtr_ptr, value, count, cystream) return (_dict_cudaError_t[err],) @cython.embedsignature(True) def cudaMemset2DAsync(devPtr, size_t pitch, int value, size_t width, size_t height, stream): """ Initializes or sets device memory to a value. Sets to the specified value `value` a matrix (`height` rows of `width` bytes each) pointed to by `dstPtr`. `pitch` is the width in bytes of the 2D array pointed to by `dstPtr`, including any padding added to the end of each row. This function performs fastest when the pitch is one that has been passed back by :py:obj:`~.cudaMallocPitch()`. :py:obj:`~.cudaMemset2DAsync()` is asynchronous with respect to the host, so the call may return before the memset is complete. The operation can optionally be associated to a stream by passing a non- zero `stream` argument. If `stream` is non-zero, the operation may overlap with operations in other streams. The device version of this function only handles device to device copies and cannot be given local or shared pointers. Parameters ---------- devPtr : Any Pointer to 2D device memory pitch : size_t Pitch in bytes of 2D device memory(Unused if `height` is 1) value : int Value to set for each byte of specified memory width : size_t Width of matrix set (columns in bytes) height : size_t Height of matrix set (rows) stream : :py:obj:`~.CUstream` or :py:obj:`~.cudaStream_t` Stream identifier Returns ------- cudaError_t :py:obj:`~.cudaSuccess`, :py:obj:`~.cudaErrorInvalidValue`, See Also -------- :py:obj:`~.cudaMemset`, :py:obj:`~.cudaMemset2D`, :py:obj:`~.cudaMemset3D`, :py:obj:`~.cudaMemsetAsync`, :py:obj:`~.cudaMemset3DAsync`, :py:obj:`~.cuMemsetD2D8Async`, :py:obj:`~.cuMemsetD2D16Async`, :py:obj:`~.cuMemsetD2D32Async` """ cdef cyruntime.cudaStream_t cystream if stream is None: pstream = 0 elif isinstance(stream, (cudaStream_t,driver.CUstream)): pstream = int(stream) else: pstream = int(cudaStream_t(stream)) cystream = pstream cydevPtr = _HelperInputVoidPtr(devPtr) cdef void* cydevPtr_ptr = cydevPtr.cptr with nogil: err = cyruntime.cudaMemset2DAsync(cydevPtr_ptr, pitch, value, width, height, cystream) return (_dict_cudaError_t[err],) @cython.embedsignature(True) def cudaMemset3DAsync(pitchedDevPtr not None : cudaPitchedPtr, int value, extent not None : cudaExtent, stream): """ Initializes or sets device memory to a value. Initializes each element of a 3D array to the specified value `value`. The object to initialize is defined by `pitchedDevPtr`. The `pitch` field of `pitchedDevPtr` is the width in memory in bytes of the 3D array pointed to by `pitchedDevPtr`, including any padding added to the end of each row. The `xsize` field specifies the logical width of each row in bytes, while the `ysize` field specifies the height of each 2D slice in rows. The `pitch` field of `pitchedDevPtr` is ignored when `height` and `depth` are both equal to 1. The extents of the initialized region are specified as a `width` in bytes, a `height` in rows, and a `depth` in slices. Extents with `width` greater than or equal to the `xsize` of `pitchedDevPtr` may perform significantly faster than extents narrower than the `xsize`. Secondarily, extents with `height` equal to the `ysize` of `pitchedDevPtr` will perform faster than when the `height` is shorter than the `ysize`. This function performs fastest when the `pitchedDevPtr` has been allocated by :py:obj:`~.cudaMalloc3D()`. :py:obj:`~.cudaMemset3DAsync()` is asynchronous with respect to the host, so the call may return before the memset is complete. The operation can optionally be associated to a stream by passing a non- zero `stream` argument. If `stream` is non-zero, the operation may overlap with operations in other streams. The device version of this function only handles device to device copies and cannot be given local or shared pointers. Parameters ---------- pitchedDevPtr : :py:obj:`~.cudaPitchedPtr` Pointer to pitched device memory value : int Value to set for each byte of specified memory extent : :py:obj:`~.cudaExtent` Size parameters for where to set device memory (`width` field in bytes) stream : :py:obj:`~.CUstream` or :py:obj:`~.cudaStream_t` Stream identifier Returns ------- cudaError_t :py:obj:`~.cudaSuccess`, :py:obj:`~.cudaErrorInvalidValue`, See Also -------- :py:obj:`~.cudaMemset`, :py:obj:`~.cudaMemset2D`, :py:obj:`~.cudaMemset3D`, :py:obj:`~.cudaMemsetAsync`, :py:obj:`~.cudaMemset2DAsync`, :py:obj:`~.cudaMalloc3D`, :py:obj:`~.make_cudaPitchedPtr`, :py:obj:`~.make_cudaExtent` """ cdef cyruntime.cudaStream_t cystream if stream is None: pstream = 0 elif isinstance(stream, (cudaStream_t,driver.CUstream)): pstream = int(stream) else: pstream = int(cudaStream_t(stream)) cystream = pstream with nogil: err = cyruntime.cudaMemset3DAsync(pitchedDevPtr._pvt_ptr[0], value, extent._pvt_ptr[0], cystream) return (_dict_cudaError_t[err],) @cython.embedsignature(True) def cudaMemPrefetchAsync(devPtr, size_t count, int dstDevice, stream): """ Prefetches memory to the specified destination device. Prefetches memory to the specified destination device. `devPtr` is the base device pointer of the memory to be prefetched and `dstDevice` is the destination device. `count` specifies the number of bytes to copy. `stream` is the stream in which the operation is enqueued. The memory range must refer to managed memory allocated via :py:obj:`~.cudaMallocManaged` or declared via managed variables, or it may also refer to system-allocated memory on systems with non-zero cudaDevAttrPageableMemoryAccess. Passing in cudaCpuDeviceId for `dstDevice` will prefetch the data to host memory. If `dstDevice` is a GPU, then the device attribute :py:obj:`~.cudaDevAttrConcurrentManagedAccess` must be non-zero. Additionally, `stream` must be associated with a device that has a non- zero value for the device attribute :py:obj:`~.cudaDevAttrConcurrentManagedAccess`. The start address and end address of the memory range will be rounded down and rounded up respectively to be aligned to CPU page size before the prefetch operation is enqueued in the stream. If no physical memory has been allocated for this region, then this memory region will be populated and mapped on the destination device. If there's insufficient memory to prefetch the desired region, the Unified Memory driver may evict pages from other :py:obj:`~.cudaMallocManaged` allocations to host memory in order to make room. Device memory allocated using :py:obj:`~.cudaMalloc` or :py:obj:`~.cudaMallocArray` will not be evicted. By default, any mappings to the previous location of the migrated pages are removed and mappings for the new location are only setup on `dstDevice`. The exact behavior however also depends on the settings applied to this memory range via :py:obj:`~.cudaMemAdvise` as described below: If :py:obj:`~.cudaMemAdviseSetReadMostly` was set on any subset of this memory range, then that subset will create a read-only copy of the pages on `dstDevice`. If :py:obj:`~.cudaMemAdviseSetPreferredLocation` was called on any subset of this memory range, then the pages will be migrated to `dstDevice` even if `dstDevice` is not the preferred location of any pages in the memory range. If :py:obj:`~.cudaMemAdviseSetAccessedBy` was called on any subset of this memory range, then mappings to those pages from all the appropriate processors are updated to refer to the new location if establishing such a mapping is possible. Otherwise, those mappings are cleared. Note that this API is not required for functionality and only serves to improve performance by allowing the application to migrate data to a suitable location before it is accessed. Memory accesses to this range are always coherent and are allowed even when the data is actively being migrated. Note that this function is asynchronous with respect to the host and all work on other devices. Parameters ---------- devPtr : Any Pointer to be prefetched count : size_t Size in bytes dstDevice : int Destination device to prefetch to stream : :py:obj:`~.CUstream` or :py:obj:`~.cudaStream_t` Stream to enqueue prefetch operation Returns ------- cudaError_t :py:obj:`~.cudaSuccess`, :py:obj:`~.cudaErrorInvalidValue`, :py:obj:`~.cudaErrorInvalidDevice` See Also -------- :py:obj:`~.cudaMemcpy`, :py:obj:`~.cudaMemcpyPeer`, :py:obj:`~.cudaMemcpyAsync`, :py:obj:`~.cudaMemcpy3DPeerAsync`, :py:obj:`~.cudaMemAdvise`, :py:obj:`~.cudaMemAdvise_v2` :py:obj:`~.cuMemPrefetchAsync` """ cdef cyruntime.cudaStream_t cystream if stream is None: pstream = 0 elif isinstance(stream, (cudaStream_t,driver.CUstream)): pstream = int(stream) else: pstream = int(cudaStream_t(stream)) cystream = pstream cydevPtr = _HelperInputVoidPtr(devPtr) cdef void* cydevPtr_ptr = cydevPtr.cptr with nogil: err = cyruntime.cudaMemPrefetchAsync(cydevPtr_ptr, count, dstDevice, cystream) return (_dict_cudaError_t[err],) @cython.embedsignature(True) def cudaMemPrefetchAsync_v2(devPtr, size_t count, location not None : cudaMemLocation, unsigned int flags, stream): """ Prefetches memory to the specified destination location. Prefetches memory to the specified destination location. `devPtr` is the base device pointer of the memory to be prefetched and `location` specifies the destination location. `count` specifies the number of bytes to copy. `stream` is the stream in which the operation is enqueued. The memory range must refer to managed memory allocated via :py:obj:`~.cudaMallocManaged` or declared via managed variables, or it may also refer to system-allocated memory on systems with non-zero cudaDevAttrPageableMemoryAccess. Specifying :py:obj:`~.cudaMemLocationTypeDevice` for :py:obj:`~.cudaMemLocation.type` will prefetch memory to GPU specified by device ordinal :py:obj:`~.cudaMemLocation.id` which must have non- zero value for the device attribute :py:obj:`~.concurrentManagedAccess`. Additionally, `stream` must be associated with a device that has a non-zero value for the device attribute :py:obj:`~.concurrentManagedAccess`. Specifying :py:obj:`~.cudaMemLocationTypeHost` as :py:obj:`~.cudaMemLocation.type` will prefetch data to host memory. Applications can request prefetching memory to a specific host NUMA node by specifying :py:obj:`~.cudaMemLocationTypeHostNuma` for :py:obj:`~.cudaMemLocation.type` and a valid host NUMA node id in :py:obj:`~.cudaMemLocation.id` Users can also request prefetching memory to the host NUMA node closest to the current thread's CPU by specifying :py:obj:`~.cudaMemLocationTypeHostNumaCurrent` for :py:obj:`~.cudaMemLocation.type`. Note when :py:obj:`~.cudaMemLocation.type` is etiher :py:obj:`~.cudaMemLocationTypeHost` OR :py:obj:`~.cudaMemLocationTypeHostNumaCurrent`, :py:obj:`~.cudaMemLocation.id` will be ignored. The start address and end address of the memory range will be rounded down and rounded up respectively to be aligned to CPU page size before the prefetch operation is enqueued in the stream. If no physical memory has been allocated for this region, then this memory region will be populated and mapped on the destination device. If there's insufficient memory to prefetch the desired region, the Unified Memory driver may evict pages from other :py:obj:`~.cudaMallocManaged` allocations to host memory in order to make room. Device memory allocated using :py:obj:`~.cudaMalloc` or :py:obj:`~.cudaMallocArray` will not be evicted. By default, any mappings to the previous location of the migrated pages are removed and mappings for the new location are only setup on the destination location. The exact behavior however also depends on the settings applied to this memory range via :py:obj:`~.cuMemAdvise` as described below: If :py:obj:`~.cudaMemAdviseSetReadMostly` was set on any subset of this memory range, then that subset will create a read-only copy of the pages on destination location. If however the destination location is a host NUMA node, then any pages of that subset that are already in another host NUMA node will be transferred to the destination. If :py:obj:`~.cudaMemAdviseSetPreferredLocation` was called on any subset of this memory range, then the pages will be migrated to `location` even if `location` is not the preferred location of any pages in the memory range. If :py:obj:`~.cudaMemAdviseSetAccessedBy` was called on any subset of this memory range, then mappings to those pages from all the appropriate processors are updated to refer to the new location if establishing such a mapping is possible. Otherwise, those mappings are cleared. Note that this API is not required for functionality and only serves to improve performance by allowing the application to migrate data to a suitable location before it is accessed. Memory accesses to this range are always coherent and are allowed even when the data is actively being migrated. Note that this function is asynchronous with respect to the host and all work on other devices. Parameters ---------- devPtr : Any Pointer to be prefetched count : size_t Size in bytes location : :py:obj:`~.cudaMemLocation` location to prefetch to flags : unsigned int flags for future use, must be zero now. stream : :py:obj:`~.CUstream` or :py:obj:`~.cudaStream_t` Stream to enqueue prefetch operation Returns ------- cudaError_t :py:obj:`~.cudaSuccess`, :py:obj:`~.cudaErrorInvalidValue`, :py:obj:`~.cudaErrorInvalidDevice` See Also -------- :py:obj:`~.cudaMemcpy`, :py:obj:`~.cudaMemcpyPeer`, :py:obj:`~.cudaMemcpyAsync`, :py:obj:`~.cudaMemcpy3DPeerAsync`, :py:obj:`~.cudaMemAdvise`, :py:obj:`~.cudaMemAdvise_v2` :py:obj:`~.cuMemPrefetchAsync` """ cdef cyruntime.cudaStream_t cystream if stream is None: pstream = 0 elif isinstance(stream, (cudaStream_t,driver.CUstream)): pstream = int(stream) else: pstream = int(cudaStream_t(stream)) cystream = pstream cydevPtr = _HelperInputVoidPtr(devPtr) cdef void* cydevPtr_ptr = cydevPtr.cptr with nogil: err = cyruntime.cudaMemPrefetchAsync_v2(cydevPtr_ptr, count, location._pvt_ptr[0], flags, cystream) return (_dict_cudaError_t[err],) @cython.embedsignature(True) def cudaMemAdvise(devPtr, size_t count, advice not None : cudaMemoryAdvise, int device): """ Advise about the usage of a given memory range. Advise the Unified Memory subsystem about the usage pattern for the memory range starting at `devPtr` with a size of `count` bytes. The start address and end address of the memory range will be rounded down and rounded up respectively to be aligned to CPU page size before the advice is applied. The memory range must refer to managed memory allocated via :py:obj:`~.cudaMallocManaged` or declared via managed variables. The memory range could also refer to system-allocated pageable memory provided it represents a valid, host-accessible region of memory and all additional constraints imposed by `advice` as outlined below are also satisfied. Specifying an invalid system- allocated pageable memory range results in an error being returned. The `advice` parameter can take the following values: - :py:obj:`~.cudaMemAdviseSetReadMostly`: This implies that the data is mostly going to be read from and only occasionally written to. Any read accesses from any processor to this region will create a read- only copy of at least the accessed pages in that processor's memory. Additionally, if :py:obj:`~.cudaMemPrefetchAsync` is called on this region, it will create a read-only copy of the data on the destination processor. If any processor writes to this region, all copies of the corresponding page will be invalidated except for the one where the write occurred. The `device` argument is ignored for this advice. Note that for a page to be read-duplicated, the accessing processor must either be the CPU or a GPU that has a non- zero value for the device attribute :py:obj:`~.cudaDevAttrConcurrentManagedAccess`. Also, if a context is created on a device that does not have the device attribute :py:obj:`~.cudaDevAttrConcurrentManagedAccess` set, then read- duplication will not occur until all such contexts are destroyed. If the memory region refers to valid system-allocated pageable memory, then the accessing device must have a non-zero value for the device attribute :py:obj:`~.cudaDevAttrPageableMemoryAccess` for a read-only copy to be created on that device. Note however that if the accessing device also has a non-zero value for the device attribute :py:obj:`~.cudaDevAttrPageableMemoryAccessUsesHostPageTables`, then setting this advice will not create a read-only copy when that device accesses this memory region. - :py:obj:`~.cudaMemAdviceUnsetReadMostly`: Undoes the effect of :py:obj:`~.cudaMemAdviceReadMostly` and also prevents the Unified Memory driver from attempting heuristic read-duplication on the memory range. Any read-duplicated copies of the data will be collapsed into a single copy. The location for the collapsed copy will be the preferred location if the page has a preferred location and one of the read-duplicated copies was resident at that location. Otherwise, the location chosen is arbitrary. - :py:obj:`~.cudaMemAdviseSetPreferredLocation`: This advice sets the preferred location for the data to be the memory belonging to `device`. Passing in cudaCpuDeviceId for `device` sets the preferred location as host memory. If `device` is a GPU, then it must have a non-zero value for the device attribute :py:obj:`~.cudaDevAttrConcurrentManagedAccess`. Setting the preferred location does not cause data to migrate to that location immediately. Instead, it guides the migration policy when a fault occurs on that memory region. If the data is already in its preferred location and the faulting processor can establish a mapping without requiring the data to be migrated, then data migration will be avoided. On the other hand, if the data is not in its preferred location or if a direct mapping cannot be established, then it will be migrated to the processor accessing it. It is important to note that setting the preferred location does not prevent data prefetching done using :py:obj:`~.cudaMemPrefetchAsync`. Having a preferred location can override the page thrash detection and resolution logic in the Unified Memory driver. Normally, if a page is detected to be constantly thrashing between for example host and device memory, the page may eventually be pinned to host memory by the Unified Memory driver. But if the preferred location is set as device memory, then the page will continue to thrash indefinitely. If :py:obj:`~.cudaMemAdviseSetReadMostly` is also set on this memory region or any subset of it, then the policies associated with that advice will override the policies of this advice, unless read accesses from `device` will not result in a read-only copy being created on that device as outlined in description for the advice :py:obj:`~.cudaMemAdviseSetReadMostly`. If the memory region refers to valid system-allocated pageable memory, then `device` must have a non-zero value for the device attribute :py:obj:`~.cudaDevAttrPageableMemoryAccess`. - :py:obj:`~.cudaMemAdviseUnsetPreferredLocation`: Undoes the effect of :py:obj:`~.cudaMemAdviseSetPreferredLocation` and changes the preferred location to none. - :py:obj:`~.cudaMemAdviseSetAccessedBy`: This advice implies that the data will be accessed by `device`. Passing in :py:obj:`~.cudaCpuDeviceId` for `device` will set the advice for the CPU. If `device` is a GPU, then the device attribute :py:obj:`~.cudaDevAttrConcurrentManagedAccess` must be non-zero. This advice does not cause data migration and has no impact on the location of the data per se. Instead, it causes the data to always be mapped in the specified processor's page tables, as long as the location of the data permits a mapping to be established. If the data gets migrated for any reason, the mappings are updated accordingly. This advice is recommended in scenarios where data locality is not important, but avoiding faults is. Consider for example a system containing multiple GPUs with peer-to-peer access enabled, where the data located on one GPU is occasionally accessed by peer GPUs. In such scenarios, migrating data over to the other GPUs is not as important because the accesses are infrequent and the overhead of migration may be too high. But preventing faults can still help improve performance, and so having a mapping set up in advance is useful. Note that on CPU access of this data, the data may be migrated to host memory because the CPU typically cannot access device memory directly. Any GPU that had the :py:obj:`~.cudaMemAdviceSetAccessedBy` flag set for this data will now have its mapping updated to point to the page in host memory. If :py:obj:`~.cudaMemAdviseSetReadMostly` is also set on this memory region or any subset of it, then the policies associated with that advice will override the policies of this advice. Additionally, if the preferred location of this memory region or any subset of it is also `device`, then the policies associated with :py:obj:`~.cudaMemAdviseSetPreferredLocation` will override the policies of this advice. If the memory region refers to valid system- allocated pageable memory, then `device` must have a non-zero value for the device attribute :py:obj:`~.cudaDevAttrPageableMemoryAccess`. Additionally, if `device` has a non-zero value for the device attribute :py:obj:`~.cudaDevAttrPageableMemoryAccessUsesHostPageTables`, then this call has no effect. - :py:obj:`~.cudaMemAdviseUnsetAccessedBy`: Undoes the effect of :py:obj:`~.cudaMemAdviseSetAccessedBy`. Any mappings to the data from `device` may be removed at any time causing accesses to result in non-fatal page faults. If the memory region refers to valid system- allocated pageable memory, then `device` must have a non-zero value for the device attribute :py:obj:`~.cudaDevAttrPageableMemoryAccess`. Additionally, if `device` has a non-zero value for the device attribute :py:obj:`~.cudaDevAttrPageableMemoryAccessUsesHostPageTables`, then this call has no effect. Parameters ---------- devPtr : Any Pointer to memory to set the advice for count : size_t Size in bytes of the memory range advice : :py:obj:`~.cudaMemoryAdvise` Advice to be applied for the specified memory range device : int Device to apply the advice for Returns ------- cudaError_t :py:obj:`~.cudaSuccess`, :py:obj:`~.cudaErrorInvalidValue`, :py:obj:`~.cudaErrorInvalidDevice` See Also -------- :py:obj:`~.cudaMemcpy`, :py:obj:`~.cudaMemcpyPeer`, :py:obj:`~.cudaMemcpyAsync`, :py:obj:`~.cudaMemcpy3DPeerAsync`, :py:obj:`~.cudaMemPrefetchAsync`, :py:obj:`~.cuMemAdvise` """ cydevPtr = _HelperInputVoidPtr(devPtr) cdef void* cydevPtr_ptr = cydevPtr.cptr cdef cyruntime.cudaMemoryAdvise cyadvice = advice.value with nogil: err = cyruntime.cudaMemAdvise(cydevPtr_ptr, count, cyadvice, device) return (_dict_cudaError_t[err],) @cython.embedsignature(True) def cudaMemAdvise_v2(devPtr, size_t count, advice not None : cudaMemoryAdvise, location not None : cudaMemLocation): """ Advise about the usage of a given memory range. Advise the Unified Memory subsystem about the usage pattern for the memory range starting at `devPtr` with a size of `count` bytes. The start address and end address of the memory range will be rounded down and rounded up respectively to be aligned to CPU page size before the advice is applied. The memory range must refer to managed memory allocated via :py:obj:`~.cudaMallocManaged` or declared via managed variables. The memory range could also refer to system-allocated pageable memory provided it represents a valid, host-accessible region of memory and all additional constraints imposed by `advice` as outlined below are also satisfied. Specifying an invalid system- allocated pageable memory range results in an error being returned. The `advice` parameter can take the following values: - :py:obj:`~.cudaMemAdviseSetReadMostly`: This implies that the data is mostly going to be read from and only occasionally written to. Any read accesses from any processor to this region will create a read- only copy of at least the accessed pages in that processor's memory. Additionally, if :py:obj:`~.cudaMemPrefetchAsync` or :py:obj:`~.cudaMemPrefetchAsync_v2` is called on this region, it will create a read-only copy of the data on the destination processor. If the target location for :py:obj:`~.cudaMemPrefetchAsync_v2` is a host NUMA node and a read-only copy already exists on another host NUMA node, that copy will be migrated to the targeted host NUMA node. If any processor writes to this region, all copies of the corresponding page will be invalidated except for the one where the write occurred. If the writing processor is the CPU and the preferred location of the page is a host NUMA node, then the page will also be migrated to that host NUMA node. The `location` argument is ignored for this advice. Note that for a page to be read-duplicated, the accessing processor must either be the CPU or a GPU that has a non-zero value for the device attribute :py:obj:`~.cudaDevAttrConcurrentManagedAccess`. Also, if a context is created on a device that does not have the device attribute :py:obj:`~.cudaDevAttrConcurrentManagedAccess` set, then read-duplication will not occur until all such contexts are destroyed. If the memory region refers to valid system-allocated pageable memory, then the accessing device must have a non-zero value for the device attribute :py:obj:`~.cudaDevAttrPageableMemoryAccess` for a read-only copy to be created on that device. Note however that if the accessing device also has a non-zero value for the device attribute :py:obj:`~.cudaDevAttrPageableMemoryAccessUsesHostPageTables`, then setting this advice will not create a read-only copy when that device accesses this memory region. - :py:obj:`~.cudaMemAdviceUnsetReadMostly`: Undoes the effect of :py:obj:`~.cudaMemAdviseSetReadMostly` and also prevents the Unified Memory driver from attempting heuristic read-duplication on the memory range. Any read-duplicated copies of the data will be collapsed into a single copy. The location for the collapsed copy will be the preferred location if the page has a preferred location and one of the read-duplicated copies was resident at that location. Otherwise, the location chosen is arbitrary. Note: The `location` argument is ignored for this advice. - :py:obj:`~.cudaMemAdviseSetPreferredLocation`: This advice sets the preferred location for the data to be the memory belonging to `location`. When :py:obj:`~.cudaMemLocation.type` is :py:obj:`~.cudaMemLocationTypeHost`, :py:obj:`~.cudaMemLocation.id` is ignored and the preferred location is set to be host memory. To set the preferred location to a specific host NUMA node, applications must set :py:obj:`~.cudaMemLocation.type` to :py:obj:`~.cudaMemLocationTypeHostNuma` and :py:obj:`~.cudaMemLocation.id` must specify the NUMA ID of the host NUMA node. If :py:obj:`~.cudaMemLocation.type` is set to :py:obj:`~.cudaMemLocationTypeHostNumaCurrent`, :py:obj:`~.cudaMemLocation.id` will be ignored and the host NUMA node closest to the calling thread's CPU will be used as the preferred location. If :py:obj:`~.cudaMemLocation.type` is a :py:obj:`~.cudaMemLocationTypeDevice`, then :py:obj:`~.cudaMemLocation.id` must be a valid device ordinal and the device must have a non-zero value for the device attribute :py:obj:`~.cudaDevAttrConcurrentManagedAccess`. Setting the preferred location does not cause data to migrate to that location immediately. Instead, it guides the migration policy when a fault occurs on that memory region. If the data is already in its preferred location and the faulting processor can establish a mapping without requiring the data to be migrated, then data migration will be avoided. On the other hand, if the data is not in its preferred location or if a direct mapping cannot be established, then it will be migrated to the processor accessing it. It is important to note that setting the preferred location does not prevent data prefetching done using :py:obj:`~.cudaMemPrefetchAsync`. Having a preferred location can override the page thrash detection and resolution logic in the Unified Memory driver. Normally, if a page is detected to be constantly thrashing between for example host and device memory, the page may eventually be pinned to host memory by the Unified Memory driver. But if the preferred location is set as device memory, then the page will continue to thrash indefinitely. If :py:obj:`~.cudaMemAdviseSetReadMostly` is also set on this memory region or any subset of it, then the policies associated with that advice will override the policies of this advice, unless read accesses from `location` will not result in a read-only copy being created on that procesor as outlined in description for the advice :py:obj:`~.cudaMemAdviseSetReadMostly`. If the memory region refers to valid system-allocated pageable memory, and :py:obj:`~.cudaMemLocation.type` is :py:obj:`~.cudaMemLocationTypeDevice` then :py:obj:`~.cudaMemLocation.id` must be a valid device that has a non- zero alue for the device attribute :py:obj:`~.cudaDevAttrPageableMemoryAccess`. - :py:obj:`~.cudaMemAdviseUnsetPreferredLocation`: Undoes the effect of :py:obj:`~.cudaMemAdviseSetPreferredLocation` and changes the preferred location to none. The `location` argument is ignored for this advice. - :py:obj:`~.cudaMemAdviseSetAccessedBy`: This advice implies that the data will be accessed by processor `location`. The :py:obj:`~.cudaMemLocation.type` must be either :py:obj:`~.cudaMemLocationTypeDevice` with :py:obj:`~.cudaMemLocation.id` representing a valid device ordinal or :py:obj:`~.cudaMemLocationTypeHost` and :py:obj:`~.cudaMemLocation.id` will be ignored. All other location types are invalid. If :py:obj:`~.cudaMemLocation.id` is a GPU, then the device attribute :py:obj:`~.cudaDevAttrConcurrentManagedAccess` must be non-zero. This advice does not cause data migration and has no impact on the location of the data per se. Instead, it causes the data to always be mapped in the specified processor's page tables, as long as the location of the data permits a mapping to be established. If the data gets migrated for any reason, the mappings are updated accordingly. This advice is recommended in scenarios where data locality is not important, but avoiding faults is. Consider for example a system containing multiple GPUs with peer-to-peer access enabled, where the data located on one GPU is occasionally accessed by peer GPUs. In such scenarios, migrating data over to the other GPUs is not as important because the accesses are infrequent and the overhead of migration may be too high. But preventing faults can still help improve performance, and so having a mapping set up in advance is useful. Note that on CPU access of this data, the data may be migrated to host memory because the CPU typically cannot access device memory directly. Any GPU that had the :py:obj:`~.cudaMemAdviseSetAccessedBy` flag set for this data will now have its mapping updated to point to the page in host memory. If :py:obj:`~.cudaMemAdviseSetReadMostly` is also set on this memory region or any subset of it, then the policies associated with that advice will override the policies of this advice. Additionally, if the preferred location of this memory region or any subset of it is also `location`, then the policies associated with :py:obj:`~.CU_MEM_ADVISE_SET_PREFERRED_LOCATION` will override the policies of this advice. If the memory region refers to valid system- allocated pageable memory, and :py:obj:`~.cudaMemLocation.type` is :py:obj:`~.cudaMemLocationTypeDevice` then device in :py:obj:`~.cudaMemLocation.id` must have a non-zero value for the device attribute :py:obj:`~.cudaDevAttrPageableMemoryAccess`. Additionally, if :py:obj:`~.cudaMemLocation.id` has a non-zero value for the device attribute :py:obj:`~.cudaDevAttrPageableMemoryAccessUsesHostPageTables`, then this call has no effect. - :py:obj:`~.CU_MEM_ADVISE_UNSET_ACCESSED_BY`: Undoes the effect of :py:obj:`~.cudaMemAdviseSetAccessedBy`. Any mappings to the data from `location` may be removed at any time causing accesses to result in non-fatal page faults. If the memory region refers to valid system- allocated pageable memory, and :py:obj:`~.cudaMemLocation.type` is :py:obj:`~.cudaMemLocationTypeDevice` then device in :py:obj:`~.cudaMemLocation.id` must have a non-zero value for the device attribute :py:obj:`~.cudaDevAttrPageableMemoryAccess`. Additionally, if :py:obj:`~.cudaMemLocation.id` has a non-zero value for the device attribute :py:obj:`~.cudaDevAttrPageableMemoryAccessUsesHostPageTables`, then this call has no effect. Parameters ---------- devPtr : Any Pointer to memory to set the advice for count : size_t Size in bytes of the memory range advice : :py:obj:`~.cudaMemoryAdvise` Advice to be applied for the specified memory range location : :py:obj:`~.cudaMemLocation` location to apply the advice for Returns ------- cudaError_t :py:obj:`~.cudaSuccess`, :py:obj:`~.cudaErrorInvalidValue`, :py:obj:`~.cudaErrorInvalidDevice` See Also -------- :py:obj:`~.cudaMemcpy`, :py:obj:`~.cudaMemcpyPeer`, :py:obj:`~.cudaMemcpyAsync`, :py:obj:`~.cudaMemcpy3DPeerAsync`, :py:obj:`~.cudaMemPrefetchAsync`, :py:obj:`~.cuMemAdvise`, :py:obj:`~.cuMemAdvise_v2` """ cydevPtr = _HelperInputVoidPtr(devPtr) cdef void* cydevPtr_ptr = cydevPtr.cptr cdef cyruntime.cudaMemoryAdvise cyadvice = advice.value with nogil: err = cyruntime.cudaMemAdvise_v2(cydevPtr_ptr, count, cyadvice, location._pvt_ptr[0]) return (_dict_cudaError_t[err],) @cython.embedsignature(True) def cudaMemRangeGetAttribute(size_t dataSize, attribute not None : cudaMemRangeAttribute, devPtr, size_t count): """ Query an attribute of a given memory range. Query an attribute about the memory range starting at `devPtr` with a size of `count` bytes. The memory range must refer to managed memory allocated via :py:obj:`~.cudaMallocManaged` or declared via managed variables. The `attribute` parameter can take the following values: - :py:obj:`~.cudaMemRangeAttributeReadMostly`: If this attribute is specified, `data` will be interpreted as a 32-bit integer, and `dataSize` must be 4. The result returned will be 1 if all pages in the given memory range have read-duplication enabled, or 0 otherwise. - :py:obj:`~.cudaMemRangeAttributePreferredLocation`: If this attribute is specified, `data` will be interpreted as a 32-bit integer, and `dataSize` must be 4. The result returned will be a GPU device id if all pages in the memory range have that GPU as their preferred location, or it will be cudaCpuDeviceId if all pages in the memory range have the CPU as their preferred location, or it will be cudaInvalidDeviceId if either all the pages don't have the same preferred location or some of the pages don't have a preferred location at all. Note that the actual location of the pages in the memory range at the time of the query may be different from the preferred location. - :py:obj:`~.cudaMemRangeAttributeAccessedBy`: If this attribute is specified, `data` will be interpreted as an array of 32-bit integers, and `dataSize` must be a non-zero multiple of 4. The result returned will be a list of device ids that had :py:obj:`~.cudaMemAdviceSetAccessedBy` set for that entire memory range. If any device does not have that advice set for the entire memory range, that device will not be included. If `data` is larger than the number of devices that have that advice set for that memory range, cudaInvalidDeviceId will be returned in all the extra space provided. For ex., if `dataSize` is 12 (i.e. `data` has 3 elements) and only device 0 has the advice set, then the result returned will be { 0, cudaInvalidDeviceId, cudaInvalidDeviceId }. If `data` is smaller than the number of devices that have that advice set, then only as many devices will be returned as can fit in the array. There is no guarantee on which specific devices will be returned, however. - :py:obj:`~.cudaMemRangeAttributeLastPrefetchLocation`: If this attribute is specified, `data` will be interpreted as a 32-bit integer, and `dataSize` must be 4. The result returned will be the last location to which all pages in the memory range were prefetched explicitly via :py:obj:`~.cudaMemPrefetchAsync`. This will either be a GPU id or cudaCpuDeviceId depending on whether the last location for prefetch was a GPU or the CPU respectively. If any page in the memory range was never explicitly prefetched or if all pages were not prefetched to the same location, cudaInvalidDeviceId will be returned. Note that this simply returns the last location that the applicaton requested to prefetch the memory range to. It gives no indication as to whether the prefetch operation to that location has completed or even begun. - :py:obj:`~.cudaMemRangeAttributePreferredLocationType`: If this attribute is specified, `data` will be interpreted as a :py:obj:`~.cudaMemLocationType`, and `dataSize` must be sizeof(cudaMemLocationType). The :py:obj:`~.cudaMemLocationType` returned will be :py:obj:`~.cudaMemLocationTypeDevice` if all pages in the memory range have the same GPU as their preferred location, or :py:obj:`~.cudaMemLocationType` will be :py:obj:`~.cudaMemLocationTypeHost` if all pages in the memory range have the CPU as their preferred location, or or it will be :py:obj:`~.cudaMemLocationTypeHostNuma` if all the pages in the memory range have the same host NUMA node ID as their preferred location or it will be :py:obj:`~.cudaMemLocationTypeInvalid` if either all the pages don't have the same preferred location or some of the pages don't have a preferred location at all. Note that the actual location type of the pages in the memory range at the time of the query may be different from the preferred location type. - :py:obj:`~.cudaMemRangeAttributePreferredLocationId`: If this attribute is specified, `data` will be interpreted as a 32-bit integer, and `dataSize` must be 4. If the :py:obj:`~.cudaMemRangeAttributePreferredLocationType` query for the same address range returns :py:obj:`~.cudaMemLocationTypeDevice`, it will be a valid device ordinal or if it returns :py:obj:`~.cudaMemLocationTypeHostNuma`, it will be a valid host NUMA node ID or if it returns any other location type, the id should be ignored. - :py:obj:`~.cudaMemRangeAttributeLastPrefetchLocationType`: If this attribute is specified, `data` will be interpreted as a :py:obj:`~.cudaMemLocationType`, and `dataSize` must be sizeof(cudaMemLocationType). The result returned will be the last location type to which all pages in the memory range were prefetched explicitly via :py:obj:`~.cuMemPrefetchAsync`. The :py:obj:`~.cudaMemLocationType` returned will be :py:obj:`~.cudaMemLocationTypeDevice` if the last prefetch location was the GPU or :py:obj:`~.cudaMemLocationTypeHost` if it was the CPU or :py:obj:`~.cudaMemLocationTypeHostNuma` if the last prefetch location was a specific host NUMA node. If any page in the memory range was never explicitly prefetched or if all pages were not prefetched to the same location, :py:obj:`~.CUmemLocationType` will be :py:obj:`~.cudaMemLocationTypeInvalid`. Note that this simply returns the last location type that the application requested to prefetch the memory range to. It gives no indication as to whether the prefetch operation to that location has completed or even begun. - :py:obj:`~.cudaMemRangeAttributeLastPrefetchLocationId`: If this attribute is specified, `data` will be interpreted as a 32-bit integer, and `dataSize` must be 4. If the :py:obj:`~.cudaMemRangeAttributeLastPrefetchLocationType` query for the same address range returns :py:obj:`~.cudaMemLocationTypeDevice`, it will be a valid device ordinal or if it returns :py:obj:`~.cudaMemLocationTypeHostNuma`, it will be a valid host NUMA node ID or if it returns any other location type, the id should be ignored. Parameters ---------- dataSize : size_t Array containing the size of data attribute : :py:obj:`~.cudaMemRangeAttribute` The attribute to query devPtr : Any Start of the range to query count : size_t Size of the range to query Returns ------- cudaError_t :py:obj:`~.cudaSuccess`, :py:obj:`~.cudaErrorInvalidValue` data : Any A pointers to a memory location where the result of each attribute query will be written to. See Also -------- :py:obj:`~.cudaMemRangeGetAttributes`, :py:obj:`~.cudaMemPrefetchAsync`, :py:obj:`~.cudaMemAdvise`, :py:obj:`~.cuMemRangeGetAttribute` """ cdef _HelperCUmem_range_attribute cydata = _HelperCUmem_range_attribute(attribute, dataSize) cdef void* cydata_ptr = cydata.cptr cdef cyruntime.cudaMemRangeAttribute cyattribute = attribute.value cydevPtr = _HelperInputVoidPtr(devPtr) cdef void* cydevPtr_ptr = cydevPtr.cptr with nogil: err = cyruntime.cudaMemRangeGetAttribute(cydata_ptr, dataSize, cyattribute, cydevPtr_ptr, count) if err != cyruntime.cudaSuccess: return (_dict_cudaError_t[err], None) return (_dict_cudaError_t[err], cydata.pyObj()) @cython.embedsignature(True) def cudaMemRangeGetAttributes(dataSizes : tuple[int] | list[int], attributes : Optional[tuple[cudaMemRangeAttribute] | list[cudaMemRangeAttribute]], size_t numAttributes, devPtr, size_t count): """ Query attributes of a given memory range. Query attributes of the memory range starting at `devPtr` with a size of `count` bytes. The memory range must refer to managed memory allocated via :py:obj:`~.cudaMallocManaged` or declared via managed variables. The `attributes` array will be interpreted to have `numAttributes` entries. The `dataSizes` array will also be interpreted to have `numAttributes` entries. The results of the query will be stored in `data`. The list of supported attributes are given below. Please refer to :py:obj:`~.cudaMemRangeGetAttribute` for attribute descriptions and restrictions. - :py:obj:`~.cudaMemRangeAttributeReadMostly` - :py:obj:`~.cudaMemRangeAttributePreferredLocation` - :py:obj:`~.cudaMemRangeAttributeAccessedBy` - :py:obj:`~.cudaMemRangeAttributeLastPrefetchLocation` - :: cudaMemRangeAttributePreferredLocationType - :: cudaMemRangeAttributePreferredLocationId - :: cudaMemRangeAttributeLastPrefetchLocationType - :: cudaMemRangeAttributeLastPrefetchLocationId Parameters ---------- dataSizes : list[int] Array containing the sizes of each result attributes : list[:py:obj:`~.cudaMemRangeAttribute`] An array of attributes to query (numAttributes and the number of attributes in this array should match) numAttributes : size_t Number of attributes to query devPtr : Any Start of the range to query count : size_t Size of the range to query Returns ------- cudaError_t :py:obj:`~.cudaSuccess`, :py:obj:`~.cudaErrorInvalidValue` data : list[Any] A two-dimensional array containing pointers to memory locations where the result of each attribute query will be written to. See Also -------- :py:obj:`~.cudaMemRangeGetAttribute`, :py:obj:`~.cudaMemAdvise`, :py:obj:`~.cudaMemPrefetchAsync`, :py:obj:`~.cuMemRangeGetAttributes` """ attributes = [] if attributes is None else attributes if not all(isinstance(_x, (cudaMemRangeAttribute)) for _x in attributes): raise TypeError("Argument 'attributes' is not instance of type (expected tuple[cyruntime.cudaMemRangeAttribute] or list[cyruntime.cudaMemRangeAttribute]") if not all(isinstance(_x, (int)) for _x in dataSizes): raise TypeError("Argument 'dataSizes' is not instance of type (expected tuple[int] or list[int]") pylist = [_HelperCUmem_range_attribute(pyattributes, pydataSizes) for (pyattributes, pydataSizes) in zip(attributes, dataSizes)] cdef _InputVoidPtrPtrHelper voidStarHelperdata = _InputVoidPtrPtrHelper(pylist) cdef void** cyvoidStarHelper_ptr = voidStarHelperdata.cptr cdef vector[size_t] cydataSizes = dataSizes cdef vector[cyruntime.cudaMemRangeAttribute] cyattributes = [pyattributes.value for pyattributes in (attributes)] if numAttributes > len(dataSizes): raise RuntimeError("List is too small: " + str(len(dataSizes)) + " < " + str(numAttributes)) if numAttributes > len(attributes): raise RuntimeError("List is too small: " + str(len(attributes)) + " < " + str(numAttributes)) cydevPtr = _HelperInputVoidPtr(devPtr) cdef void* cydevPtr_ptr = cydevPtr.cptr with nogil: err = cyruntime.cudaMemRangeGetAttributes(cyvoidStarHelper_ptr, cydataSizes.data(), cyattributes.data(), numAttributes, cydevPtr_ptr, count) if err != cyruntime.cudaSuccess: return (_dict_cudaError_t[err], None) return (_dict_cudaError_t[err], [obj.pyObj() for obj in pylist]) @cython.embedsignature(True) def cudaMemcpyToArray(dst, size_t wOffset, size_t hOffset, src, size_t count, kind not None : cudaMemcpyKind): """ Copies data between host and device. [Deprecated] Copies `count` bytes from the memory area pointed to by `src` to the CUDA array `dst` starting at `hOffset` rows and `wOffset` bytes from the upper left corner, where `kind` specifies the direction of the copy, and must be one of :py:obj:`~.cudaMemcpyHostToHost`, :py:obj:`~.cudaMemcpyHostToDevice`, :py:obj:`~.cudaMemcpyDeviceToHost`, :py:obj:`~.cudaMemcpyDeviceToDevice`, or :py:obj:`~.cudaMemcpyDefault`. Passing :py:obj:`~.cudaMemcpyDefault` is recommended, in which case the type of transfer is inferred from the pointer values. However, :py:obj:`~.cudaMemcpyDefault` is only allowed on systems that support unified virtual addressing. Parameters ---------- dst : :py:obj:`~.cudaArray_t` Destination memory address wOffset : size_t Destination starting X offset (columns in bytes) hOffset : size_t Destination starting Y offset (rows) src : Any Source memory address count : size_t Size in bytes to copy kind : :py:obj:`~.cudaMemcpyKind` Type of transfer Returns ------- cudaError_t :py:obj:`~.cudaSuccess`, :py:obj:`~.cudaErrorInvalidValue`, :py:obj:`~.cudaErrorInvalidMemcpyDirection` See Also -------- :py:obj:`~.cudaMemcpy`, :py:obj:`~.cudaMemcpy2D`, :py:obj:`~.cudaMemcpy2DToArray`, :py:obj:`~.cudaMemcpyFromArray`, :py:obj:`~.cudaMemcpy2DFromArray`, :py:obj:`~.cudaMemcpyArrayToArray`, :py:obj:`~.cudaMemcpy2DArrayToArray`, :py:obj:`~.cudaMemcpyToSymbol`, :py:obj:`~.cudaMemcpyFromSymbol`, :py:obj:`~.cudaMemcpyAsync`, :py:obj:`~.cudaMemcpy2DAsync`, :py:obj:`~.cudaMemcpyToArrayAsync`, :py:obj:`~.cudaMemcpy2DToArrayAsync`, :py:obj:`~.cudaMemcpyFromArrayAsync`, :py:obj:`~.cudaMemcpy2DFromArrayAsync`, :py:obj:`~.cudaMemcpyToSymbolAsync`, :py:obj:`~.cudaMemcpyFromSymbolAsync`, :py:obj:`~.cuMemcpyHtoA`, :py:obj:`~.cuMemcpyDtoA` """ cdef cyruntime.cudaArray_t cydst if dst is None: pdst = 0 elif isinstance(dst, (cudaArray_t,)): pdst = int(dst) else: pdst = int(cudaArray_t(dst)) cydst = pdst cysrc = _HelperInputVoidPtr(src) cdef void* cysrc_ptr = cysrc.cptr cdef cyruntime.cudaMemcpyKind cykind = kind.value with nogil: err = cyruntime.cudaMemcpyToArray(cydst, wOffset, hOffset, cysrc_ptr, count, cykind) return (_dict_cudaError_t[err],) @cython.embedsignature(True) def cudaMemcpyFromArray(dst, src, size_t wOffset, size_t hOffset, size_t count, kind not None : cudaMemcpyKind): """ Copies data between host and device. [Deprecated] Copies `count` bytes from the CUDA array `src` starting at `hOffset` rows and `wOffset` bytes from the upper left corner to the memory area pointed to by `dst`, where `kind` specifies the direction of the copy, and must be one of :py:obj:`~.cudaMemcpyHostToHost`, :py:obj:`~.cudaMemcpyHostToDevice`, :py:obj:`~.cudaMemcpyDeviceToHost`, :py:obj:`~.cudaMemcpyDeviceToDevice`, or :py:obj:`~.cudaMemcpyDefault`. Passing :py:obj:`~.cudaMemcpyDefault` is recommended, in which case the type of transfer is inferred from the pointer values. However, :py:obj:`~.cudaMemcpyDefault` is only allowed on systems that support unified virtual addressing. Parameters ---------- dst : Any Destination memory address src : :py:obj:`~.cudaArray_const_t` Source memory address wOffset : size_t Source starting X offset (columns in bytes) hOffset : size_t Source starting Y offset (rows) count : size_t Size in bytes to copy kind : :py:obj:`~.cudaMemcpyKind` Type of transfer Returns ------- cudaError_t :py:obj:`~.cudaSuccess`, :py:obj:`~.cudaErrorInvalidValue`, :py:obj:`~.cudaErrorInvalidMemcpyDirection` See Also -------- :py:obj:`~.cudaMemcpy`, :py:obj:`~.cudaMemcpy2D`, :py:obj:`~.cudaMemcpyToArray`, :py:obj:`~.cudaMemcpy2DToArray`, :py:obj:`~.cudaMemcpy2DFromArray`, :py:obj:`~.cudaMemcpyArrayToArray`, :py:obj:`~.cudaMemcpy2DArrayToArray`, :py:obj:`~.cudaMemcpyToSymbol`, :py:obj:`~.cudaMemcpyFromSymbol`, :py:obj:`~.cudaMemcpyAsync`, :py:obj:`~.cudaMemcpy2DAsync`, :py:obj:`~.cudaMemcpyToArrayAsync`, :py:obj:`~.cudaMemcpy2DToArrayAsync`, :py:obj:`~.cudaMemcpyFromArrayAsync`, :py:obj:`~.cudaMemcpy2DFromArrayAsync`, :py:obj:`~.cudaMemcpyToSymbolAsync`, :py:obj:`~.cudaMemcpyFromSymbolAsync`, :py:obj:`~.cuMemcpyAtoH`, :py:obj:`~.cuMemcpyAtoD` """ cdef cyruntime.cudaArray_const_t cysrc if src is None: psrc = 0 elif isinstance(src, (cudaArray_const_t,)): psrc = int(src) else: psrc = int(cudaArray_const_t(src)) cysrc = psrc cydst = _HelperInputVoidPtr(dst) cdef void* cydst_ptr = cydst.cptr cdef cyruntime.cudaMemcpyKind cykind = kind.value with nogil: err = cyruntime.cudaMemcpyFromArray(cydst_ptr, cysrc, wOffset, hOffset, count, cykind) return (_dict_cudaError_t[err],) @cython.embedsignature(True) def cudaMemcpyArrayToArray(dst, size_t wOffsetDst, size_t hOffsetDst, src, size_t wOffsetSrc, size_t hOffsetSrc, size_t count, kind not None : cudaMemcpyKind): """ Copies data between host and device. [Deprecated] Copies `count` bytes from the CUDA array `src` starting at `hOffsetSrc` rows and `wOffsetSrc` bytes from the upper left corner to the CUDA array `dst` starting at `hOffsetDst` rows and `wOffsetDst` bytes from the upper left corner, where `kind` specifies the direction of the copy, and must be one of :py:obj:`~.cudaMemcpyHostToHost`, :py:obj:`~.cudaMemcpyHostToDevice`, :py:obj:`~.cudaMemcpyDeviceToHost`, :py:obj:`~.cudaMemcpyDeviceToDevice`, or :py:obj:`~.cudaMemcpyDefault`. Passing :py:obj:`~.cudaMemcpyDefault` is recommended, in which case the type of transfer is inferred from the pointer values. However, :py:obj:`~.cudaMemcpyDefault` is only allowed on systems that support unified virtual addressing. Parameters ---------- dst : :py:obj:`~.cudaArray_t` Destination memory address wOffsetDst : size_t Destination starting X offset (columns in bytes) hOffsetDst : size_t Destination starting Y offset (rows) src : :py:obj:`~.cudaArray_const_t` Source memory address wOffsetSrc : size_t Source starting X offset (columns in bytes) hOffsetSrc : size_t Source starting Y offset (rows) count : size_t Size in bytes to copy kind : :py:obj:`~.cudaMemcpyKind` Type of transfer Returns ------- cudaError_t :py:obj:`~.cudaSuccess`, :py:obj:`~.cudaErrorInvalidValue`, :py:obj:`~.cudaErrorInvalidMemcpyDirection` See Also -------- :py:obj:`~.cudaMemcpy`, :py:obj:`~.cudaMemcpy2D`, :py:obj:`~.cudaMemcpyToArray`, :py:obj:`~.cudaMemcpy2DToArray`, :py:obj:`~.cudaMemcpyFromArray`, :py:obj:`~.cudaMemcpy2DFromArray`, :py:obj:`~.cudaMemcpy2DArrayToArray`, :py:obj:`~.cudaMemcpyToSymbol`, :py:obj:`~.cudaMemcpyFromSymbol`, :py:obj:`~.cudaMemcpyAsync`, :py:obj:`~.cudaMemcpy2DAsync`, :py:obj:`~.cudaMemcpyToArrayAsync`, :py:obj:`~.cudaMemcpy2DToArrayAsync`, :py:obj:`~.cudaMemcpyFromArrayAsync`, :py:obj:`~.cudaMemcpy2DFromArrayAsync`, :py:obj:`~.cudaMemcpyToSymbolAsync`, :py:obj:`~.cudaMemcpyFromSymbolAsync`, :py:obj:`~.cuMemcpyAtoA` """ cdef cyruntime.cudaArray_const_t cysrc if src is None: psrc = 0 elif isinstance(src, (cudaArray_const_t,)): psrc = int(src) else: psrc = int(cudaArray_const_t(src)) cysrc = psrc cdef cyruntime.cudaArray_t cydst if dst is None: pdst = 0 elif isinstance(dst, (cudaArray_t,)): pdst = int(dst) else: pdst = int(cudaArray_t(dst)) cydst = pdst cdef cyruntime.cudaMemcpyKind cykind = kind.value with nogil: err = cyruntime.cudaMemcpyArrayToArray(cydst, wOffsetDst, hOffsetDst, cysrc, wOffsetSrc, hOffsetSrc, count, cykind) return (_dict_cudaError_t[err],) @cython.embedsignature(True) def cudaMemcpyToArrayAsync(dst, size_t wOffset, size_t hOffset, src, size_t count, kind not None : cudaMemcpyKind, stream): """ Copies data between host and device. [Deprecated] Copies `count` bytes from the memory area pointed to by `src` to the CUDA array `dst` starting at `hOffset` rows and `wOffset` bytes from the upper left corner, where `kind` specifies the direction of the copy, and must be one of :py:obj:`~.cudaMemcpyHostToHost`, :py:obj:`~.cudaMemcpyHostToDevice`, :py:obj:`~.cudaMemcpyDeviceToHost`, :py:obj:`~.cudaMemcpyDeviceToDevice`, or :py:obj:`~.cudaMemcpyDefault`. Passing :py:obj:`~.cudaMemcpyDefault` is recommended, in which case the type of transfer is inferred from the pointer values. However, :py:obj:`~.cudaMemcpyDefault` is only allowed on systems that support unified virtual addressing. :py:obj:`~.cudaMemcpyToArrayAsync()` is asynchronous with respect to the host, so the call may return before the copy is complete. The copy can optionally be associated to a stream by passing a non-zero `stream` argument. If `kind` is :py:obj:`~.cudaMemcpyHostToDevice` or :py:obj:`~.cudaMemcpyDeviceToHost` and `stream` is non-zero, the copy may overlap with operations in other streams. Parameters ---------- dst : :py:obj:`~.cudaArray_t` Destination memory address wOffset : size_t Destination starting X offset (columns in bytes) hOffset : size_t Destination starting Y offset (rows) src : Any Source memory address count : size_t Size in bytes to copy kind : :py:obj:`~.cudaMemcpyKind` Type of transfer stream : :py:obj:`~.CUstream` or :py:obj:`~.cudaStream_t` Stream identifier Returns ------- cudaError_t :py:obj:`~.cudaSuccess`, :py:obj:`~.cudaErrorInvalidValue`, :py:obj:`~.cudaErrorInvalidMemcpyDirection` See Also -------- :py:obj:`~.cudaMemcpy`, :py:obj:`~.cudaMemcpy2D`, :py:obj:`~.cudaMemcpyToArray`, :py:obj:`~.cudaMemcpy2DToArray`, :py:obj:`~.cudaMemcpyFromArray`, :py:obj:`~.cudaMemcpy2DFromArray`, :py:obj:`~.cudaMemcpyArrayToArray`, :py:obj:`~.cudaMemcpy2DArrayToArray`, :py:obj:`~.cudaMemcpyToSymbol`, :py:obj:`~.cudaMemcpyFromSymbol`, :py:obj:`~.cudaMemcpyAsync`, :py:obj:`~.cudaMemcpy2DAsync`, :py:obj:`~.cudaMemcpy2DToArrayAsync`, :py:obj:`~.cudaMemcpyFromArrayAsync`, :py:obj:`~.cudaMemcpy2DFromArrayAsync`, :py:obj:`~.cudaMemcpyToSymbolAsync`, :py:obj:`~.cudaMemcpyFromSymbolAsync`, :py:obj:`~.cuMemcpyHtoAAsync`, :py:obj:`~.cuMemcpy2DAsync` """ cdef cyruntime.cudaStream_t cystream if stream is None: pstream = 0 elif isinstance(stream, (cudaStream_t,driver.CUstream)): pstream = int(stream) else: pstream = int(cudaStream_t(stream)) cystream = pstream cdef cyruntime.cudaArray_t cydst if dst is None: pdst = 0 elif isinstance(dst, (cudaArray_t,)): pdst = int(dst) else: pdst = int(cudaArray_t(dst)) cydst = pdst cysrc = _HelperInputVoidPtr(src) cdef void* cysrc_ptr = cysrc.cptr cdef cyruntime.cudaMemcpyKind cykind = kind.value with nogil: err = cyruntime.cudaMemcpyToArrayAsync(cydst, wOffset, hOffset, cysrc_ptr, count, cykind, cystream) return (_dict_cudaError_t[err],) @cython.embedsignature(True) def cudaMemcpyFromArrayAsync(dst, src, size_t wOffset, size_t hOffset, size_t count, kind not None : cudaMemcpyKind, stream): """ Copies data between host and device. [Deprecated] Copies `count` bytes from the CUDA array `src` starting at `hOffset` rows and `wOffset` bytes from the upper left corner to the memory area pointed to by `dst`, where `kind` specifies the direction of the copy, and must be one of :py:obj:`~.cudaMemcpyHostToHost`, :py:obj:`~.cudaMemcpyHostToDevice`, :py:obj:`~.cudaMemcpyDeviceToHost`, :py:obj:`~.cudaMemcpyDeviceToDevice`, or :py:obj:`~.cudaMemcpyDefault`. Passing :py:obj:`~.cudaMemcpyDefault` is recommended, in which case the type of transfer is inferred from the pointer values. However, :py:obj:`~.cudaMemcpyDefault` is only allowed on systems that support unified virtual addressing. :py:obj:`~.cudaMemcpyFromArrayAsync()` is asynchronous with respect to the host, so the call may return before the copy is complete. The copy can optionally be associated to a stream by passing a non-zero `stream` argument. If `kind` is :py:obj:`~.cudaMemcpyHostToDevice` or :py:obj:`~.cudaMemcpyDeviceToHost` and `stream` is non-zero, the copy may overlap with operations in other streams. Parameters ---------- dst : Any Destination memory address src : :py:obj:`~.cudaArray_const_t` Source memory address wOffset : size_t Source starting X offset (columns in bytes) hOffset : size_t Source starting Y offset (rows) count : size_t Size in bytes to copy kind : :py:obj:`~.cudaMemcpyKind` Type of transfer stream : :py:obj:`~.CUstream` or :py:obj:`~.cudaStream_t` Stream identifier Returns ------- cudaError_t :py:obj:`~.cudaSuccess`, :py:obj:`~.cudaErrorInvalidValue`, :py:obj:`~.cudaErrorInvalidMemcpyDirection` See Also -------- :py:obj:`~.cudaMemcpy`, :py:obj:`~.cudaMemcpy2D`, :py:obj:`~.cudaMemcpyToArray`, :py:obj:`~.cudaMemcpy2DToArray`, :py:obj:`~.cudaMemcpyFromArray`, :py:obj:`~.cudaMemcpy2DFromArray`, :py:obj:`~.cudaMemcpyArrayToArray`, :py:obj:`~.cudaMemcpy2DArrayToArray`, :py:obj:`~.cudaMemcpyToSymbol`, :py:obj:`~.cudaMemcpyFromSymbol`, :py:obj:`~.cudaMemcpyAsync`, :py:obj:`~.cudaMemcpy2DAsync`, :py:obj:`~.cudaMemcpyToArrayAsync`, :py:obj:`~.cudaMemcpy2DToArrayAsync`, :py:obj:`~.cudaMemcpy2DFromArrayAsync`, :py:obj:`~.cudaMemcpyToSymbolAsync`, :py:obj:`~.cudaMemcpyFromSymbolAsync`, :py:obj:`~.cuMemcpyAtoHAsync`, :py:obj:`~.cuMemcpy2DAsync` """ cdef cyruntime.cudaStream_t cystream if stream is None: pstream = 0 elif isinstance(stream, (cudaStream_t,driver.CUstream)): pstream = int(stream) else: pstream = int(cudaStream_t(stream)) cystream = pstream cdef cyruntime.cudaArray_const_t cysrc if src is None: psrc = 0 elif isinstance(src, (cudaArray_const_t,)): psrc = int(src) else: psrc = int(cudaArray_const_t(src)) cysrc = psrc cydst = _HelperInputVoidPtr(dst) cdef void* cydst_ptr = cydst.cptr cdef cyruntime.cudaMemcpyKind cykind = kind.value with nogil: err = cyruntime.cudaMemcpyFromArrayAsync(cydst_ptr, cysrc, wOffset, hOffset, count, cykind, cystream) return (_dict_cudaError_t[err],) @cython.embedsignature(True) def cudaMallocAsync(size_t size, hStream): """ Allocates memory with stream ordered semantics. Inserts an allocation operation into `hStream`. A pointer to the allocated memory is returned immediately in *dptr. The allocation must not be accessed until the the allocation operation completes. The allocation comes from the memory pool associated with the stream's device. Parameters ---------- size : size_t Number of bytes to allocate hStream : :py:obj:`~.CUstream` or :py:obj:`~.cudaStream_t` The stream establishing the stream ordering contract and the memory pool to allocate from Returns ------- cudaError_t :py:obj:`~.cudaSuccess`, :py:obj:`~.cudaErrorInvalidValue`, :py:obj:`~.cudaErrorNotSupported`, :py:obj:`~.cudaErrorOutOfMemory`, devPtr : Any Returned device pointer See Also -------- :py:obj:`~.cuMemAllocAsync`, cudaMallocAsync (C++ API), :py:obj:`~.cudaMallocFromPoolAsync`, :py:obj:`~.cudaFreeAsync`, :py:obj:`~.cudaDeviceSetMemPool`, :py:obj:`~.cudaDeviceGetDefaultMemPool`, :py:obj:`~.cudaDeviceGetMemPool`, :py:obj:`~.cudaMemPoolSetAccess`, :py:obj:`~.cudaMemPoolSetAttribute`, :py:obj:`~.cudaMemPoolGetAttribute` Notes ----- The default memory pool of a device contains device memory from that device. Basic stream ordering allows future work submitted into the same stream to use the allocation. Stream query, stream synchronize, and CUDA events can be used to guarantee that the allocation operation completes before work submitted in a separate stream runs. During stream capture, this function results in the creation of an allocation node. In this case, the allocation is owned by the graph instead of the memory pool. The memory pool's properties are used to set the node's creation parameters. """ cdef cyruntime.cudaStream_t cyhStream if hStream is None: phStream = 0 elif isinstance(hStream, (cudaStream_t,driver.CUstream)): phStream = int(hStream) else: phStream = int(cudaStream_t(hStream)) cyhStream = phStream cdef void_ptr devPtr = 0 with nogil: err = cyruntime.cudaMallocAsync(&devPtr, size, cyhStream) if err != cyruntime.cudaSuccess: return (_dict_cudaError_t[err], None) return (_dict_cudaError_t[err], devPtr) @cython.embedsignature(True) def cudaFreeAsync(devPtr, hStream): """ Frees memory with stream ordered semantics. Inserts a free operation into `hStream`. The allocation must not be accessed after stream execution reaches the free. After this API returns, accessing the memory from any subsequent work launched on the GPU or querying its pointer attributes results in undefined behavior. Parameters ---------- dptr : Any memory to free hStream : :py:obj:`~.CUstream` or :py:obj:`~.cudaStream_t` The stream establishing the stream ordering promise Returns ------- cudaError_t :py:obj:`~.cudaSuccess`, :py:obj:`~.cudaErrorInvalidValue`, :py:obj:`~.cudaErrorNotSupported` See Also -------- :py:obj:`~.cuMemFreeAsync`, :py:obj:`~.cudaMallocAsync` Notes ----- During stream capture, this function results in the creation of a free node and must therefore be passed the address of a graph allocation. """ cdef cyruntime.cudaStream_t cyhStream if hStream is None: phStream = 0 elif isinstance(hStream, (cudaStream_t,driver.CUstream)): phStream = int(hStream) else: phStream = int(cudaStream_t(hStream)) cyhStream = phStream cydevPtr = _HelperInputVoidPtr(devPtr) cdef void* cydevPtr_ptr = cydevPtr.cptr with nogil: err = cyruntime.cudaFreeAsync(cydevPtr_ptr, cyhStream) return (_dict_cudaError_t[err],) @cython.embedsignature(True) def cudaMemPoolTrimTo(memPool, size_t minBytesToKeep): """ Tries to release memory back to the OS. Releases memory back to the OS until the pool contains fewer than minBytesToKeep reserved bytes, or there is no more memory that the allocator can safely release. The allocator cannot release OS allocations that back outstanding asynchronous allocations. The OS allocations may happen at different granularity from the user allocations. Parameters ---------- pool : :py:obj:`~.CUmemoryPool` or :py:obj:`~.cudaMemPool_t` The memory pool to trim minBytesToKeep : size_t If the pool has less than minBytesToKeep reserved, the TrimTo operation is a no-op. Otherwise the pool will be guaranteed to have at least minBytesToKeep bytes reserved after the operation. Returns ------- cudaError_t :py:obj:`~.cudaSuccess`, :py:obj:`~.cudaErrorInvalidValue` See Also -------- :py:obj:`~.cuMemPoolTrimTo`, :py:obj:`~.cudaMallocAsync`, :py:obj:`~.cudaFreeAsync`, :py:obj:`~.cudaDeviceGetDefaultMemPool`, :py:obj:`~.cudaDeviceGetMemPool`, :py:obj:`~.cudaMemPoolCreate` Notes ----- : Allocations that have not been freed count as outstanding. : Allocations that have been asynchronously freed but whose completion has not been observed on the host (eg. by a synchronize) can count as outstanding. """ cdef cyruntime.cudaMemPool_t cymemPool if memPool is None: pmemPool = 0 elif isinstance(memPool, (cudaMemPool_t,driver.CUmemoryPool)): pmemPool = int(memPool) else: pmemPool = int(cudaMemPool_t(memPool)) cymemPool = pmemPool with nogil: err = cyruntime.cudaMemPoolTrimTo(cymemPool, minBytesToKeep) return (_dict_cudaError_t[err],) @cython.embedsignature(True) def cudaMemPoolSetAttribute(memPool, attr not None : cudaMemPoolAttr, value): """ Sets attributes of a memory pool. Supported attributes are: - :py:obj:`~.cudaMemPoolAttrReleaseThreshold`: (value type = cuuint64_t) Amount of reserved memory in bytes to hold onto before trying to release memory back to the OS. When more than the release threshold bytes of memory are held by the memory pool, the allocator will try to release memory back to the OS on the next call to stream, event or context synchronize. (default 0) - :py:obj:`~.cudaMemPoolReuseFollowEventDependencies`: (value type = int) Allow :py:obj:`~.cudaMallocAsync` to use memory asynchronously freed in another stream as long as a stream ordering dependency of the allocating stream on the free action exists. Cuda events and null stream interactions can create the required stream ordered dependencies. (default enabled) - :py:obj:`~.cudaMemPoolReuseAllowOpportunistic`: (value type = int) Allow reuse of already completed frees when there is no dependency between the free and allocation. (default enabled) - :py:obj:`~.cudaMemPoolReuseAllowInternalDependencies`: (value type = int) Allow :py:obj:`~.cudaMallocAsync` to insert new stream dependencies in order to establish the stream ordering required to reuse a piece of memory released by :py:obj:`~.cudaFreeAsync` (default enabled). - :py:obj:`~.cudaMemPoolAttrReservedMemHigh`: (value type = cuuint64_t) Reset the high watermark that tracks the amount of backing memory that was allocated for the memory pool. It is illegal to set this attribute to a non-zero value. - :py:obj:`~.cudaMemPoolAttrUsedMemHigh`: (value type = cuuint64_t) Reset the high watermark that tracks the amount of used memory that was allocated for the memory pool. It is illegal to set this attribute to a non-zero value. Parameters ---------- pool : :py:obj:`~.CUmemoryPool` or :py:obj:`~.cudaMemPool_t` The memory pool to modify attr : :py:obj:`~.cudaMemPoolAttr` The attribute to modify value : Any Pointer to the value to assign Returns ------- cudaError_t :py:obj:`~.cudaSuccess`, :py:obj:`~.cudaErrorInvalidValue` See Also -------- :py:obj:`~.cuMemPoolSetAttribute`, :py:obj:`~.cudaMallocAsync`, :py:obj:`~.cudaFreeAsync`, :py:obj:`~.cudaDeviceGetDefaultMemPool`, :py:obj:`~.cudaDeviceGetMemPool`, :py:obj:`~.cudaMemPoolCreate` """ cdef cyruntime.cudaMemPool_t cymemPool if memPool is None: pmemPool = 0 elif isinstance(memPool, (cudaMemPool_t,driver.CUmemoryPool)): pmemPool = int(memPool) else: pmemPool = int(cudaMemPool_t(memPool)) cymemPool = pmemPool cdef cyruntime.cudaMemPoolAttr cyattr = attr.value cdef _HelperCUmemPool_attribute cyvalue = _HelperCUmemPool_attribute(attr, value, is_getter=False) cdef void* cyvalue_ptr = cyvalue.cptr with nogil: err = cyruntime.cudaMemPoolSetAttribute(cymemPool, cyattr, cyvalue_ptr) return (_dict_cudaError_t[err],) @cython.embedsignature(True) def cudaMemPoolGetAttribute(memPool, attr not None : cudaMemPoolAttr): """ Gets attributes of a memory pool. Supported attributes are: - :py:obj:`~.cudaMemPoolAttrReleaseThreshold`: (value type = cuuint64_t) Amount of reserved memory in bytes to hold onto before trying to release memory back to the OS. When more than the release threshold bytes of memory are held by the memory pool, the allocator will try to release memory back to the OS on the next call to stream, event or context synchronize. (default 0) - :py:obj:`~.cudaMemPoolReuseFollowEventDependencies`: (value type = int) Allow :py:obj:`~.cudaMallocAsync` to use memory asynchronously freed in another stream as long as a stream ordering dependency of the allocating stream on the free action exists. Cuda events and null stream interactions can create the required stream ordered dependencies. (default enabled) - :py:obj:`~.cudaMemPoolReuseAllowOpportunistic`: (value type = int) Allow reuse of already completed frees when there is no dependency between the free and allocation. (default enabled) - :py:obj:`~.cudaMemPoolReuseAllowInternalDependencies`: (value type = int) Allow :py:obj:`~.cudaMallocAsync` to insert new stream dependencies in order to establish the stream ordering required to reuse a piece of memory released by :py:obj:`~.cudaFreeAsync` (default enabled). - :py:obj:`~.cudaMemPoolAttrReservedMemCurrent`: (value type = cuuint64_t) Amount of backing memory currently allocated for the mempool. - :py:obj:`~.cudaMemPoolAttrReservedMemHigh`: (value type = cuuint64_t) High watermark of backing memory allocated for the mempool since the last time it was reset. - :py:obj:`~.cudaMemPoolAttrUsedMemCurrent`: (value type = cuuint64_t) Amount of memory from the pool that is currently in use by the application. - :py:obj:`~.cudaMemPoolAttrUsedMemHigh`: (value type = cuuint64_t) High watermark of the amount of memory from the pool that was in use by the application since the last time it was reset. Parameters ---------- pool : :py:obj:`~.CUmemoryPool` or :py:obj:`~.cudaMemPool_t` The memory pool to get attributes of attr : :py:obj:`~.cudaMemPoolAttr` The attribute to get Returns ------- cudaError_t :py:obj:`~.cudaSuccess`, :py:obj:`~.cudaErrorInvalidValue` value : Any Retrieved value See Also -------- :py:obj:`~.cuMemPoolGetAttribute`, :py:obj:`~.cudaMallocAsync`, :py:obj:`~.cudaFreeAsync`, :py:obj:`~.cudaDeviceGetDefaultMemPool`, :py:obj:`~.cudaDeviceGetMemPool`, :py:obj:`~.cudaMemPoolCreate` """ cdef cyruntime.cudaMemPool_t cymemPool if memPool is None: pmemPool = 0 elif isinstance(memPool, (cudaMemPool_t,driver.CUmemoryPool)): pmemPool = int(memPool) else: pmemPool = int(cudaMemPool_t(memPool)) cymemPool = pmemPool cdef cyruntime.cudaMemPoolAttr cyattr = attr.value cdef _HelperCUmemPool_attribute cyvalue = _HelperCUmemPool_attribute(attr, 0, is_getter=True) cdef void* cyvalue_ptr = cyvalue.cptr with nogil: err = cyruntime.cudaMemPoolGetAttribute(cymemPool, cyattr, cyvalue_ptr) if err != cyruntime.cudaSuccess: return (_dict_cudaError_t[err], None) return (_dict_cudaError_t[err], cyvalue.pyObj()) @cython.embedsignature(True) def cudaMemPoolSetAccess(memPool, descList : Optional[tuple[cudaMemAccessDesc] | list[cudaMemAccessDesc]], size_t count): """ Controls visibility of pools between devices. Parameters ---------- pool : :py:obj:`~.CUmemoryPool` or :py:obj:`~.cudaMemPool_t` The pool being modified map : list[:py:obj:`~.cudaMemAccessDesc`] Array of access descriptors. Each descriptor instructs the access to enable for a single gpu count : size_t Number of descriptors in the map array. Returns ------- cudaError_t :py:obj:`~.cudaSuccess`, :py:obj:`~.cudaErrorInvalidValue` See Also -------- :py:obj:`~.cuMemPoolSetAccess`, :py:obj:`~.cudaMemPoolGetAccess`, :py:obj:`~.cudaMallocAsync`, :py:obj:`~.cudaFreeAsync` """ descList = [] if descList is None else descList if not all(isinstance(_x, (cudaMemAccessDesc,)) for _x in descList): raise TypeError("Argument 'descList' is not instance of type (expected tuple[cyruntime.cudaMemAccessDesc,] or list[cyruntime.cudaMemAccessDesc,]") cdef cyruntime.cudaMemPool_t cymemPool if memPool is None: pmemPool = 0 elif isinstance(memPool, (cudaMemPool_t,driver.CUmemoryPool)): pmemPool = int(memPool) else: pmemPool = int(cudaMemPool_t(memPool)) cymemPool = pmemPool cdef cyruntime.cudaMemAccessDesc* cydescList = NULL if len(descList) > 1: cydescList = calloc(len(descList), sizeof(cyruntime.cudaMemAccessDesc)) if cydescList is NULL: raise MemoryError('Failed to allocate length x size memory: ' + str(len(descList)) + 'x' + str(sizeof(cyruntime.cudaMemAccessDesc))) for idx in range(len(descList)): string.memcpy(&cydescList[idx], (descList[idx])._pvt_ptr, sizeof(cyruntime.cudaMemAccessDesc)) elif len(descList) == 1: cydescList = (descList[0])._pvt_ptr if count > len(descList): raise RuntimeError("List is too small: " + str(len(descList)) + " < " + str(count)) with nogil: err = cyruntime.cudaMemPoolSetAccess(cymemPool, cydescList, count) if len(descList) > 1 and cydescList is not NULL: free(cydescList) return (_dict_cudaError_t[err],) @cython.embedsignature(True) def cudaMemPoolGetAccess(memPool, location : Optional[cudaMemLocation]): """ Returns the accessibility of a pool from a device. Returns the accessibility of the pool's memory from the specified location. Parameters ---------- memPool : :py:obj:`~.CUmemoryPool` or :py:obj:`~.cudaMemPool_t` the pool being queried location : :py:obj:`~.cudaMemLocation` the location accessing the pool Returns ------- cudaError_t flags : :py:obj:`~.cudaMemAccessFlags` the accessibility of the pool from the specified location See Also -------- :py:obj:`~.cuMemPoolGetAccess`, :py:obj:`~.cudaMemPoolSetAccess` """ cdef cyruntime.cudaMemPool_t cymemPool if memPool is None: pmemPool = 0 elif isinstance(memPool, (cudaMemPool_t,driver.CUmemoryPool)): pmemPool = int(memPool) else: pmemPool = int(cudaMemPool_t(memPool)) cymemPool = pmemPool cdef cyruntime.cudaMemAccessFlags flags cdef cyruntime.cudaMemLocation* cylocation_ptr = location._pvt_ptr if location is not None else NULL with nogil: err = cyruntime.cudaMemPoolGetAccess(&flags, cymemPool, cylocation_ptr) if err != cyruntime.cudaSuccess: return (_dict_cudaError_t[err], None) return (_dict_cudaError_t[err], cudaMemAccessFlags(flags)) @cython.embedsignature(True) def cudaMemPoolCreate(poolProps : Optional[cudaMemPoolProps]): """ Creates a memory pool. Creates a CUDA memory pool and returns the handle in `pool`. The `poolProps` determines the properties of the pool such as the backing device and IPC capabilities. To create a memory pool targeting a specific host NUMA node, applications must set :py:obj:`~.cudaMemPoolProps`::cudaMemLocation::type to :py:obj:`~.cudaMemLocationTypeHostNuma` and :py:obj:`~.cudaMemPoolProps`::cudaMemLocation::id must specify the NUMA ID of the host memory node. Specifying :py:obj:`~.cudaMemLocationTypeHostNumaCurrent` or :py:obj:`~.cudaMemLocationTypeHost` as the :py:obj:`~.cudaMemPoolProps`::cudaMemLocation::type will result in :py:obj:`~.cudaErrorInvalidValue`. By default, the pool's memory will be accessible from the device it is allocated on. In the case of pools created with :py:obj:`~.cudaMemLocationTypeHostNuma`, their default accessibility will be from the host CPU. Applications can control the maximum size of the pool by specifying a non-zero value for :py:obj:`~.cudaMemPoolProps.maxSize`. If set to 0, the maximum size of the pool will default to a system dependent value. Applications that intend to use :py:obj:`~.CU_MEM_HANDLE_TYPE_FABRIC` based memory sharing must ensure: (1) `nvidia-caps-imex-channels` character device is created by the driver and is listed under /proc/devices (2) have at least one IMEX channel file accessible by the user launching the application. When exporter and importer CUDA processes have been granted access to the same IMEX channel, they can securely share memory. The IMEX channel security model works on a per user basis. Which means all processes under a user can share memory if the user has access to a valid IMEX channel. When multi-user isolation is desired, a separate IMEX channel is required for each user. These channel files exist in /dev/nvidia-caps-imex-channels/channel* and can be created using standard OS native calls like mknod on Linux. For example: To create channel0 with the major number from /proc/devices users can execute the following command: `mknod /dev/nvidia-caps-imex-channels/channel0 c 0` Parameters ---------- poolProps : :py:obj:`~.cudaMemPoolProps` None Returns ------- cudaError_t :py:obj:`~.cudaSuccess`, :py:obj:`~.cudaErrorInvalidValue`, :py:obj:`~.cudaErrorNotSupported` memPool : :py:obj:`~.cudaMemPool_t` None See Also -------- :py:obj:`~.cuMemPoolCreate`, :py:obj:`~.cudaDeviceSetMemPool`, :py:obj:`~.cudaMallocFromPoolAsync`, :py:obj:`~.cudaMemPoolExportToShareableHandle`, :py:obj:`~.cudaDeviceGetDefaultMemPool`, :py:obj:`~.cudaDeviceGetMemPool` Notes ----- Specifying cudaMemHandleTypeNone creates a memory pool that will not support IPC. """ cdef cudaMemPool_t memPool = cudaMemPool_t() cdef cyruntime.cudaMemPoolProps* cypoolProps_ptr = poolProps._pvt_ptr if poolProps is not None else NULL with nogil: err = cyruntime.cudaMemPoolCreate(memPool._pvt_ptr, cypoolProps_ptr) if err != cyruntime.cudaSuccess: return (_dict_cudaError_t[err], None) return (_dict_cudaError_t[err], memPool) @cython.embedsignature(True) def cudaMemPoolDestroy(memPool): """ Destroys the specified memory pool. If any pointers obtained from this pool haven't been freed or the pool has free operations that haven't completed when :py:obj:`~.cudaMemPoolDestroy` is invoked, the function will return immediately and the resources associated with the pool will be released automatically once there are no more outstanding allocations. Destroying the current mempool of a device sets the default mempool of that device as the current mempool for that device. Parameters ---------- memPool : :py:obj:`~.CUmemoryPool` or :py:obj:`~.cudaMemPool_t` None Returns ------- cudaError_t :py:obj:`~.cudaSuccess`, :py:obj:`~.cudaErrorInvalidValue` See Also -------- cuMemPoolDestroy, :py:obj:`~.cudaFreeAsync`, :py:obj:`~.cudaDeviceSetMemPool`, :py:obj:`~.cudaDeviceGetDefaultMemPool`, :py:obj:`~.cudaDeviceGetMemPool`, :py:obj:`~.cudaMemPoolCreate` Notes ----- A device's default memory pool cannot be destroyed. """ cdef cyruntime.cudaMemPool_t cymemPool if memPool is None: pmemPool = 0 elif isinstance(memPool, (cudaMemPool_t,driver.CUmemoryPool)): pmemPool = int(memPool) else: pmemPool = int(cudaMemPool_t(memPool)) cymemPool = pmemPool with nogil: err = cyruntime.cudaMemPoolDestroy(cymemPool) return (_dict_cudaError_t[err],) @cython.embedsignature(True) def cudaMallocFromPoolAsync(size_t size, memPool, stream): """ Allocates memory from a specified pool with stream ordered semantics. Inserts an allocation operation into `hStream`. A pointer to the allocated memory is returned immediately in *dptr. The allocation must not be accessed until the the allocation operation completes. The allocation comes from the specified memory pool. Parameters ---------- bytesize : size_t Number of bytes to allocate memPool : :py:obj:`~.CUmemoryPool` or :py:obj:`~.cudaMemPool_t` The pool to allocate from stream : :py:obj:`~.CUstream` or :py:obj:`~.cudaStream_t` The stream establishing the stream ordering semantic Returns ------- cudaError_t :py:obj:`~.cudaSuccess`, :py:obj:`~.cudaErrorInvalidValue`, :py:obj:`~.cudaErrorNotSupported`, :py:obj:`~.cudaErrorOutOfMemory` ptr : Any Returned device pointer See Also -------- :py:obj:`~.cuMemAllocFromPoolAsync`, cudaMallocAsync (C++ API), :py:obj:`~.cudaMallocAsync`, :py:obj:`~.cudaFreeAsync`, :py:obj:`~.cudaDeviceGetDefaultMemPool`, :py:obj:`~.cudaMemPoolCreate`, :py:obj:`~.cudaMemPoolSetAccess`, :py:obj:`~.cudaMemPoolSetAttribute` Notes ----- During stream capture, this function results in the creation of an allocation node. In this case, the allocation is owned by the graph instead of the memory pool. The memory pool's properties are used to set the node's creation parameters. """ cdef cyruntime.cudaStream_t cystream if stream is None: pstream = 0 elif isinstance(stream, (cudaStream_t,driver.CUstream)): pstream = int(stream) else: pstream = int(cudaStream_t(stream)) cystream = pstream cdef cyruntime.cudaMemPool_t cymemPool if memPool is None: pmemPool = 0 elif isinstance(memPool, (cudaMemPool_t,driver.CUmemoryPool)): pmemPool = int(memPool) else: pmemPool = int(cudaMemPool_t(memPool)) cymemPool = pmemPool cdef void_ptr ptr = 0 with nogil: err = cyruntime.cudaMallocFromPoolAsync(&ptr, size, cymemPool, cystream) if err != cyruntime.cudaSuccess: return (_dict_cudaError_t[err], None) return (_dict_cudaError_t[err], ptr) @cython.embedsignature(True) def cudaMemPoolExportToShareableHandle(memPool, handleType not None : cudaMemAllocationHandleType, unsigned int flags): """ Exports a memory pool to the requested handle type. Given an IPC capable mempool, create an OS handle to share the pool with another process. A recipient process can convert the shareable handle into a mempool with :py:obj:`~.cudaMemPoolImportFromShareableHandle`. Individual pointers can then be shared with the :py:obj:`~.cudaMemPoolExportPointer` and :py:obj:`~.cudaMemPoolImportPointer` APIs. The implementation of what the shareable handle is and how it can be transferred is defined by the requested handle type. Parameters ---------- pool : :py:obj:`~.CUmemoryPool` or :py:obj:`~.cudaMemPool_t` pool to export handleType : :py:obj:`~.cudaMemAllocationHandleType` the type of handle to create flags : unsigned int must be 0 Returns ------- cudaError_t :py:obj:`~.cudaSuccess`, :py:obj:`~.cudaErrorInvalidValue`, :py:obj:`~.cudaErrorOutOfMemory` handle_out : Any pointer to the location in which to store the requested handle See Also -------- :py:obj:`~.cuMemPoolExportToShareableHandle`, :py:obj:`~.cudaMemPoolImportFromShareableHandle`, :py:obj:`~.cudaMemPoolExportPointer`, :py:obj:`~.cudaMemPoolImportPointer` Notes ----- : To create an IPC capable mempool, create a mempool with a CUmemAllocationHandleType other than cudaMemHandleTypeNone. """ cdef cyruntime.cudaMemPool_t cymemPool if memPool is None: pmemPool = 0 elif isinstance(memPool, (cudaMemPool_t,driver.CUmemoryPool)): pmemPool = int(memPool) else: pmemPool = int(cudaMemPool_t(memPool)) cymemPool = pmemPool cdef _HelperCUmemAllocationHandleType cyshareableHandle = _HelperCUmemAllocationHandleType(handleType) cdef void* cyshareableHandle_ptr = cyshareableHandle.cptr cdef cyruntime.cudaMemAllocationHandleType cyhandleType = handleType.value with nogil: err = cyruntime.cudaMemPoolExportToShareableHandle(cyshareableHandle_ptr, cymemPool, cyhandleType, flags) if err != cyruntime.cudaSuccess: return (_dict_cudaError_t[err], None) return (_dict_cudaError_t[err], cyshareableHandle.pyObj()) @cython.embedsignature(True) def cudaMemPoolImportFromShareableHandle(shareableHandle, handleType not None : cudaMemAllocationHandleType, unsigned int flags): """ imports a memory pool from a shared handle. Specific allocations can be imported from the imported pool with :py:obj:`~.cudaMemPoolImportPointer`. Parameters ---------- handle : Any OS handle of the pool to open handleType : :py:obj:`~.cudaMemAllocationHandleType` The type of handle being imported flags : unsigned int must be 0 Returns ------- cudaError_t :py:obj:`~.cudaSuccess`, :py:obj:`~.cudaErrorInvalidValue`, :py:obj:`~.cudaErrorOutOfMemory` pool_out : :py:obj:`~.cudaMemPool_t` Returned memory pool See Also -------- :py:obj:`~.cuMemPoolImportFromShareableHandle`, :py:obj:`~.cudaMemPoolExportToShareableHandle`, :py:obj:`~.cudaMemPoolExportPointer`, :py:obj:`~.cudaMemPoolImportPointer` Notes ----- Imported memory pools do not support creating new allocations. As such imported memory pools may not be used in :py:obj:`~.cudaDeviceSetMemPool` or :py:obj:`~.cudaMallocFromPoolAsync` calls. """ cdef cudaMemPool_t memPool = cudaMemPool_t() cyshareableHandle = _HelperInputVoidPtr(shareableHandle) cdef void* cyshareableHandle_ptr = cyshareableHandle.cptr cdef cyruntime.cudaMemAllocationHandleType cyhandleType = handleType.value with nogil: err = cyruntime.cudaMemPoolImportFromShareableHandle(memPool._pvt_ptr, cyshareableHandle_ptr, cyhandleType, flags) if err != cyruntime.cudaSuccess: return (_dict_cudaError_t[err], None) return (_dict_cudaError_t[err], memPool) @cython.embedsignature(True) def cudaMemPoolExportPointer(ptr): """ Export data to share a memory pool allocation between processes. Constructs `shareData_out` for sharing a specific allocation from an already shared memory pool. The recipient process can import the allocation with the :py:obj:`~.cudaMemPoolImportPointer` api. The data is not a handle and may be shared through any IPC mechanism. Parameters ---------- ptr : Any pointer to memory being exported Returns ------- cudaError_t :py:obj:`~.cudaSuccess`, :py:obj:`~.cudaErrorInvalidValue`, :py:obj:`~.cudaErrorOutOfMemory` shareData_out : :py:obj:`~.cudaMemPoolPtrExportData` Returned export data See Also -------- :py:obj:`~.cuMemPoolExportPointer`, :py:obj:`~.cudaMemPoolExportToShareableHandle`, :py:obj:`~.cudaMemPoolImportFromShareableHandle`, :py:obj:`~.cudaMemPoolImportPointer` """ cdef cudaMemPoolPtrExportData exportData = cudaMemPoolPtrExportData() cyptr = _HelperInputVoidPtr(ptr) cdef void* cyptr_ptr = cyptr.cptr with nogil: err = cyruntime.cudaMemPoolExportPointer(exportData._pvt_ptr, cyptr_ptr) if err != cyruntime.cudaSuccess: return (_dict_cudaError_t[err], None) return (_dict_cudaError_t[err], exportData) @cython.embedsignature(True) def cudaMemPoolImportPointer(memPool, exportData : Optional[cudaMemPoolPtrExportData]): """ Import a memory pool allocation from another process. Returns in `ptr_out` a pointer to the imported memory. The imported memory must not be accessed before the allocation operation completes in the exporting process. The imported memory must be freed from all importing processes before being freed in the exporting process. The pointer may be freed with cudaFree or cudaFreeAsync. If :py:obj:`~.cudaFreeAsync` is used, the free must be completed on the importing process before the free operation on the exporting process. Parameters ---------- pool : :py:obj:`~.CUmemoryPool` or :py:obj:`~.cudaMemPool_t` pool from which to import shareData : :py:obj:`~.cudaMemPoolPtrExportData` data specifying the memory to import Returns ------- cudaError_t :py:obj:`~.CUDA_SUCCESS`, :py:obj:`~.CUDA_ERROR_INVALID_VALUE`, :py:obj:`~.CUDA_ERROR_NOT_INITIALIZED`, :py:obj:`~.CUDA_ERROR_OUT_OF_MEMORY` ptr_out : Any pointer to imported memory See Also -------- :py:obj:`~.cuMemPoolImportPointer`, :py:obj:`~.cudaMemPoolExportToShareableHandle`, :py:obj:`~.cudaMemPoolImportFromShareableHandle`, :py:obj:`~.cudaMemPoolExportPointer` Notes ----- The :py:obj:`~.cudaFreeAsync` api may be used in the exporting process before the :py:obj:`~.cudaFreeAsync` operation completes in its stream as long as the :py:obj:`~.cudaFreeAsync` in the exporting process specifies a stream with a stream dependency on the importing process's :py:obj:`~.cudaFreeAsync`. """ cdef cyruntime.cudaMemPool_t cymemPool if memPool is None: pmemPool = 0 elif isinstance(memPool, (cudaMemPool_t,driver.CUmemoryPool)): pmemPool = int(memPool) else: pmemPool = int(cudaMemPool_t(memPool)) cymemPool = pmemPool cdef void_ptr ptr = 0 cdef cyruntime.cudaMemPoolPtrExportData* cyexportData_ptr = exportData._pvt_ptr if exportData is not None else NULL with nogil: err = cyruntime.cudaMemPoolImportPointer(&ptr, cymemPool, cyexportData_ptr) if err != cyruntime.cudaSuccess: return (_dict_cudaError_t[err], None) return (_dict_cudaError_t[err], ptr) @cython.embedsignature(True) def cudaPointerGetAttributes(ptr): """ Returns attributes about a specified pointer. Returns in `*attributes` the attributes of the pointer `ptr`. If pointer was not allocated in, mapped by or registered with context supporting unified addressing :py:obj:`~.cudaErrorInvalidValue` is returned. The :py:obj:`~.cudaPointerAttributes` structure is defined as: **View CUDA Toolkit Documentation for a C++ code example** In this structure, the individual fields mean - :py:obj:`~.cudaPointerAttributes.type` identifies type of memory. It can be :py:obj:`~.cudaMemoryTypeUnregistered` for unregistered host memory, :py:obj:`~.cudaMemoryTypeHost` for registered host memory, :py:obj:`~.cudaMemoryTypeDevice` for device memory or :py:obj:`~.cudaMemoryTypeManaged` for managed memory. - :py:obj:`~.device` is the device against which `ptr` was allocated. If `ptr` has memory type :py:obj:`~.cudaMemoryTypeDevice` then this identifies the device on which the memory referred to by `ptr` physically resides. If `ptr` has memory type :py:obj:`~.cudaMemoryTypeHost` then this identifies the device which was current when the allocation was made (and if that device is deinitialized then this allocation will vanish with that device's state). - :py:obj:`~.devicePointer` is the device pointer alias through which the memory referred to by `ptr` may be accessed on the current device. If the memory referred to by `ptr` cannot be accessed directly by the current device then this is NULL. - :py:obj:`~.hostPointer` is the host pointer alias through which the memory referred to by `ptr` may be accessed on the host. If the memory referred to by `ptr` cannot be accessed directly by the host then this is NULL. Parameters ---------- ptr : Any Pointer to get attributes for Returns ------- cudaError_t :py:obj:`~.cudaSuccess`, :py:obj:`~.cudaErrorInvalidDevice`, :py:obj:`~.cudaErrorInvalidValue` attributes : :py:obj:`~.cudaPointerAttributes` Attributes for the specified pointer See Also -------- :py:obj:`~.cudaGetDeviceCount`, :py:obj:`~.cudaGetDevice`, :py:obj:`~.cudaSetDevice`, :py:obj:`~.cudaChooseDevice`, :py:obj:`~.cudaInitDevice`, :py:obj:`~.cuPointerGetAttributes` Notes ----- In CUDA 11.0 forward passing host pointer will return :py:obj:`~.cudaMemoryTypeUnregistered` in :py:obj:`~.cudaPointerAttributes.type` and call will return :py:obj:`~.cudaSuccess`. """ cdef cudaPointerAttributes attributes = cudaPointerAttributes() cyptr = _HelperInputVoidPtr(ptr) cdef void* cyptr_ptr = cyptr.cptr with nogil: err = cyruntime.cudaPointerGetAttributes(attributes._pvt_ptr, cyptr_ptr) if err != cyruntime.cudaSuccess: return (_dict_cudaError_t[err], None) return (_dict_cudaError_t[err], attributes) @cython.embedsignature(True) def cudaDeviceCanAccessPeer(int device, int peerDevice): """ Queries if a device may directly access a peer device's memory. Returns in `*canAccessPeer` a value of 1 if device `device` is capable of directly accessing memory from `peerDevice` and 0 otherwise. If direct access of `peerDevice` from `device` is possible, then access may be enabled by calling :py:obj:`~.cudaDeviceEnablePeerAccess()`. Parameters ---------- device : int Device from which allocations on `peerDevice` are to be directly accessed. peerDevice : int Device on which the allocations to be directly accessed by `device` reside. Returns ------- cudaError_t :py:obj:`~.cudaSuccess`, :py:obj:`~.cudaErrorInvalidDevice` canAccessPeer : int Returned access capability See Also -------- :py:obj:`~.cudaDeviceEnablePeerAccess`, :py:obj:`~.cudaDeviceDisablePeerAccess`, :py:obj:`~.cuDeviceCanAccessPeer` """ cdef int canAccessPeer = 0 with nogil: err = cyruntime.cudaDeviceCanAccessPeer(&canAccessPeer, device, peerDevice) if err != cyruntime.cudaSuccess: return (_dict_cudaError_t[err], None) return (_dict_cudaError_t[err], canAccessPeer) @cython.embedsignature(True) def cudaDeviceEnablePeerAccess(int peerDevice, unsigned int flags): """ Enables direct access to memory allocations on a peer device. On success, all allocations from `peerDevice` will immediately be accessible by the current device. They will remain accessible until access is explicitly disabled using :py:obj:`~.cudaDeviceDisablePeerAccess()` or either device is reset using :py:obj:`~.cudaDeviceReset()`. Note that access granted by this call is unidirectional and that in order to access memory on the current device from `peerDevice`, a separate symmetric call to :py:obj:`~.cudaDeviceEnablePeerAccess()` is required. Note that there are both device-wide and system-wide limitations per system configuration, as noted in the CUDA Programming Guide under the section "Peer-to-Peer Memory Access". Returns :py:obj:`~.cudaErrorInvalidDevice` if :py:obj:`~.cudaDeviceCanAccessPeer()` indicates that the current device cannot directly access memory from `peerDevice`. Returns :py:obj:`~.cudaErrorPeerAccessAlreadyEnabled` if direct access of `peerDevice` from the current device has already been enabled. Returns :py:obj:`~.cudaErrorInvalidValue` if `flags` is not 0. Parameters ---------- peerDevice : int Peer device to enable direct access to from the current device flags : unsigned int Reserved for future use and must be set to 0 Returns ------- cudaError_t :py:obj:`~.cudaSuccess`, :py:obj:`~.cudaErrorInvalidDevice`, :py:obj:`~.cudaErrorPeerAccessAlreadyEnabled`, :py:obj:`~.cudaErrorInvalidValue` See Also -------- :py:obj:`~.cudaDeviceCanAccessPeer`, :py:obj:`~.cudaDeviceDisablePeerAccess`, :py:obj:`~.cuCtxEnablePeerAccess` """ with nogil: err = cyruntime.cudaDeviceEnablePeerAccess(peerDevice, flags) return (_dict_cudaError_t[err],) @cython.embedsignature(True) def cudaDeviceDisablePeerAccess(int peerDevice): """ Disables direct access to memory allocations on a peer device. Returns :py:obj:`~.cudaErrorPeerAccessNotEnabled` if direct access to memory on `peerDevice` has not yet been enabled from the current device. Parameters ---------- peerDevice : int Peer device to disable direct access to Returns ------- cudaError_t :py:obj:`~.cudaSuccess`, :py:obj:`~.cudaErrorPeerAccessNotEnabled`, :py:obj:`~.cudaErrorInvalidDevice` See Also -------- :py:obj:`~.cudaDeviceCanAccessPeer`, :py:obj:`~.cudaDeviceEnablePeerAccess`, :py:obj:`~.cuCtxDisablePeerAccess` """ with nogil: err = cyruntime.cudaDeviceDisablePeerAccess(peerDevice) return (_dict_cudaError_t[err],) @cython.embedsignature(True) def cudaGraphicsUnregisterResource(resource): """ Unregisters a graphics resource for access by CUDA. Unregisters the graphics resource `resource` so it is not accessible by CUDA unless registered again. If `resource` is invalid then :py:obj:`~.cudaErrorInvalidResourceHandle` is returned. Parameters ---------- resource : :py:obj:`~.cudaGraphicsResource_t` Resource to unregister Returns ------- cudaError_t :py:obj:`~.cudaSuccess`, :py:obj:`~.cudaErrorInvalidResourceHandle`, :py:obj:`~.cudaErrorUnknown` See Also -------- :py:obj:`~.cudaGraphicsD3D9RegisterResource`, :py:obj:`~.cudaGraphicsD3D10RegisterResource`, :py:obj:`~.cudaGraphicsD3D11RegisterResource`, :py:obj:`~.cudaGraphicsGLRegisterBuffer`, :py:obj:`~.cudaGraphicsGLRegisterImage`, :py:obj:`~.cuGraphicsUnregisterResource` """ cdef cyruntime.cudaGraphicsResource_t cyresource if resource is None: presource = 0 elif isinstance(resource, (cudaGraphicsResource_t,)): presource = int(resource) else: presource = int(cudaGraphicsResource_t(resource)) cyresource = presource with nogil: err = cyruntime.cudaGraphicsUnregisterResource(cyresource) return (_dict_cudaError_t[err],) @cython.embedsignature(True) def cudaGraphicsResourceSetMapFlags(resource, unsigned int flags): """ Set usage flags for mapping a graphics resource. Set `flags` for mapping the graphics resource `resource`. Changes to `flags` will take effect the next time `resource` is mapped. The `flags` argument may be any of the following: - :py:obj:`~.cudaGraphicsMapFlagsNone`: Specifies no hints about how `resource` will be used. It is therefore assumed that CUDA may read from or write to `resource`. - :py:obj:`~.cudaGraphicsMapFlagsReadOnly`: Specifies that CUDA will not write to `resource`. - :py:obj:`~.cudaGraphicsMapFlagsWriteDiscard`: Specifies CUDA will not read from `resource` and will write over the entire contents of `resource`, so none of the data previously stored in `resource` will be preserved. If `resource` is presently mapped for access by CUDA then :py:obj:`~.cudaErrorUnknown` is returned. If `flags` is not one of the above values then :py:obj:`~.cudaErrorInvalidValue` is returned. Parameters ---------- resource : :py:obj:`~.cudaGraphicsResource_t` Registered resource to set flags for flags : unsigned int Parameters for resource mapping Returns ------- cudaError_t :py:obj:`~.cudaSuccess`, :py:obj:`~.cudaErrorInvalidValue`, :py:obj:`~.cudaErrorInvalidResourceHandle`, :py:obj:`~.cudaErrorUnknown`, See Also -------- :py:obj:`~.cudaGraphicsMapResources`, :py:obj:`~.cuGraphicsResourceSetMapFlags` """ cdef cyruntime.cudaGraphicsResource_t cyresource if resource is None: presource = 0 elif isinstance(resource, (cudaGraphicsResource_t,)): presource = int(resource) else: presource = int(cudaGraphicsResource_t(resource)) cyresource = presource with nogil: err = cyruntime.cudaGraphicsResourceSetMapFlags(cyresource, flags) return (_dict_cudaError_t[err],) @cython.embedsignature(True) def cudaGraphicsMapResources(int count, resources, stream): """ Map graphics resources for access by CUDA. Maps the `count` graphics resources in `resources` for access by CUDA. The resources in `resources` may be accessed by CUDA until they are unmapped. The graphics API from which `resources` were registered should not access any resources while they are mapped by CUDA. If an application does so, the results are undefined. This function provides the synchronization guarantee that any graphics calls issued before :py:obj:`~.cudaGraphicsMapResources()` will complete before any subsequent CUDA work issued in `stream` begins. If `resources` contains any duplicate entries then :py:obj:`~.cudaErrorInvalidResourceHandle` is returned. If any of `resources` are presently mapped for access by CUDA then :py:obj:`~.cudaErrorUnknown` is returned. Parameters ---------- count : int Number of resources to map resources : :py:obj:`~.cudaGraphicsResource_t` Resources to map for CUDA stream : :py:obj:`~.CUstream` or :py:obj:`~.cudaStream_t` Stream for synchronization Returns ------- cudaError_t :py:obj:`~.cudaSuccess`, :py:obj:`~.cudaErrorInvalidResourceHandle`, :py:obj:`~.cudaErrorUnknown` See Also -------- :py:obj:`~.cudaGraphicsResourceGetMappedPointer`, :py:obj:`~.cudaGraphicsSubResourceGetMappedArray`, :py:obj:`~.cudaGraphicsUnmapResources`, :py:obj:`~.cuGraphicsMapResources` """ cdef cyruntime.cudaStream_t cystream if stream is None: pstream = 0 elif isinstance(stream, (cudaStream_t,driver.CUstream)): pstream = int(stream) else: pstream = int(cudaStream_t(stream)) cystream = pstream cdef cyruntime.cudaGraphicsResource_t *cyresources if resources is None: cyresources = NULL elif isinstance(resources, (cudaGraphicsResource_t,)): presources = resources.getPtr() cyresources = presources elif isinstance(resources, (int)): cyresources = resources else: raise TypeError("Argument 'resources' is not instance of type (expected , found " + str(type(resources))) with nogil: err = cyruntime.cudaGraphicsMapResources(count, cyresources, cystream) return (_dict_cudaError_t[err],) @cython.embedsignature(True) def cudaGraphicsUnmapResources(int count, resources, stream): """ Unmap graphics resources. Unmaps the `count` graphics resources in `resources`. Once unmapped, the resources in `resources` may not be accessed by CUDA until they are mapped again. This function provides the synchronization guarantee that any CUDA work issued in `stream` before :py:obj:`~.cudaGraphicsUnmapResources()` will complete before any subsequently issued graphics work begins. If `resources` contains any duplicate entries then :py:obj:`~.cudaErrorInvalidResourceHandle` is returned. If any of `resources` are not presently mapped for access by CUDA then :py:obj:`~.cudaErrorUnknown` is returned. Parameters ---------- count : int Number of resources to unmap resources : :py:obj:`~.cudaGraphicsResource_t` Resources to unmap stream : :py:obj:`~.CUstream` or :py:obj:`~.cudaStream_t` Stream for synchronization Returns ------- cudaError_t :py:obj:`~.cudaSuccess`, :py:obj:`~.cudaErrorInvalidResourceHandle`, :py:obj:`~.cudaErrorUnknown` See Also -------- :py:obj:`~.cudaGraphicsMapResources`, :py:obj:`~.cuGraphicsUnmapResources` """ cdef cyruntime.cudaStream_t cystream if stream is None: pstream = 0 elif isinstance(stream, (cudaStream_t,driver.CUstream)): pstream = int(stream) else: pstream = int(cudaStream_t(stream)) cystream = pstream cdef cyruntime.cudaGraphicsResource_t *cyresources if resources is None: cyresources = NULL elif isinstance(resources, (cudaGraphicsResource_t,)): presources = resources.getPtr() cyresources = presources elif isinstance(resources, (int)): cyresources = resources else: raise TypeError("Argument 'resources' is not instance of type (expected , found " + str(type(resources))) with nogil: err = cyruntime.cudaGraphicsUnmapResources(count, cyresources, cystream) return (_dict_cudaError_t[err],) @cython.embedsignature(True) def cudaGraphicsResourceGetMappedPointer(resource): """ Get an device pointer through which to access a mapped graphics resource. Returns in `*devPtr` a pointer through which the mapped graphics resource `resource` may be accessed. Returns in `*size` the size of the memory in bytes which may be accessed from that pointer. The value set in `devPtr` may change every time that `resource` is mapped. If `resource` is not a buffer then it cannot be accessed via a pointer and :py:obj:`~.cudaErrorUnknown` is returned. If `resource` is not mapped then :py:obj:`~.cudaErrorUnknown` is returned. Parameters ---------- resource : :py:obj:`~.cudaGraphicsResource_t` None Returns ------- cudaError_t devPtr : Any None size : int None """ cdef cyruntime.cudaGraphicsResource_t cyresource if resource is None: presource = 0 elif isinstance(resource, (cudaGraphicsResource_t,)): presource = int(resource) else: presource = int(cudaGraphicsResource_t(resource)) cyresource = presource cdef void_ptr devPtr = 0 cdef size_t size = 0 with nogil: err = cyruntime.cudaGraphicsResourceGetMappedPointer(&devPtr, &size, cyresource) if err != cyruntime.cudaSuccess: return (_dict_cudaError_t[err], None, None) return (_dict_cudaError_t[err], devPtr, size) @cython.embedsignature(True) def cudaGraphicsSubResourceGetMappedArray(resource, unsigned int arrayIndex, unsigned int mipLevel): """ Get an array through which to access a subresource of a mapped graphics resource. Returns in `*array` an array through which the subresource of the mapped graphics resource `resource` which corresponds to array index `arrayIndex` and mipmap level `mipLevel` may be accessed. The value set in `array` may change every time that `resource` is mapped. If `resource` is not a texture then it cannot be accessed via an array and :py:obj:`~.cudaErrorUnknown` is returned. If `arrayIndex` is not a valid array index for `resource` then :py:obj:`~.cudaErrorInvalidValue` is returned. If `mipLevel` is not a valid mipmap level for `resource` then :py:obj:`~.cudaErrorInvalidValue` is returned. If `resource` is not mapped then :py:obj:`~.cudaErrorUnknown` is returned. Parameters ---------- resource : :py:obj:`~.cudaGraphicsResource_t` Mapped resource to access arrayIndex : unsigned int Array index for array textures or cubemap face index as defined by :py:obj:`~.cudaGraphicsCubeFace` for cubemap textures for the subresource to access mipLevel : unsigned int Mipmap level for the subresource to access Returns ------- cudaError_t :py:obj:`~.cudaSuccess`, :py:obj:`~.cudaErrorInvalidValue`, :py:obj:`~.cudaErrorInvalidResourceHandle`, :py:obj:`~.cudaErrorUnknown` array : :py:obj:`~.cudaArray_t` Returned array through which a subresource of `resource` may be accessed See Also -------- :py:obj:`~.cudaGraphicsResourceGetMappedPointer`, :py:obj:`~.cuGraphicsSubResourceGetMappedArray` """ cdef cyruntime.cudaGraphicsResource_t cyresource if resource is None: presource = 0 elif isinstance(resource, (cudaGraphicsResource_t,)): presource = int(resource) else: presource = int(cudaGraphicsResource_t(resource)) cyresource = presource cdef cudaArray_t array = cudaArray_t() with nogil: err = cyruntime.cudaGraphicsSubResourceGetMappedArray(array._pvt_ptr, cyresource, arrayIndex, mipLevel) if err != cyruntime.cudaSuccess: return (_dict_cudaError_t[err], None) return (_dict_cudaError_t[err], array) @cython.embedsignature(True) def cudaGraphicsResourceGetMappedMipmappedArray(resource): """ Get a mipmapped array through which to access a mapped graphics resource. Returns in `*mipmappedArray` a mipmapped array through which the mapped graphics resource `resource` may be accessed. The value set in `mipmappedArray` may change every time that `resource` is mapped. If `resource` is not a texture then it cannot be accessed via an array and :py:obj:`~.cudaErrorUnknown` is returned. If `resource` is not mapped then :py:obj:`~.cudaErrorUnknown` is returned. Parameters ---------- resource : :py:obj:`~.cudaGraphicsResource_t` Mapped resource to access Returns ------- cudaError_t :py:obj:`~.cudaSuccess`, :py:obj:`~.cudaErrorInvalidValue`, :py:obj:`~.cudaErrorInvalidResourceHandle`, :py:obj:`~.cudaErrorUnknown` mipmappedArray : :py:obj:`~.cudaMipmappedArray_t` Returned mipmapped array through which `resource` may be accessed See Also -------- :py:obj:`~.cudaGraphicsResourceGetMappedPointer`, :py:obj:`~.cuGraphicsResourceGetMappedMipmappedArray` """ cdef cyruntime.cudaGraphicsResource_t cyresource if resource is None: presource = 0 elif isinstance(resource, (cudaGraphicsResource_t,)): presource = int(resource) else: presource = int(cudaGraphicsResource_t(resource)) cyresource = presource cdef cudaMipmappedArray_t mipmappedArray = cudaMipmappedArray_t() with nogil: err = cyruntime.cudaGraphicsResourceGetMappedMipmappedArray(mipmappedArray._pvt_ptr, cyresource) if err != cyruntime.cudaSuccess: return (_dict_cudaError_t[err], None) return (_dict_cudaError_t[err], mipmappedArray) @cython.embedsignature(True) def cudaGetChannelDesc(array): """ Get the channel descriptor of an array. Returns in `*desc` the channel descriptor of the CUDA array `array`. Parameters ---------- array : :py:obj:`~.cudaArray_const_t` Memory array on device Returns ------- cudaError_t :py:obj:`~.cudaSuccess`, :py:obj:`~.cudaErrorInvalidValue` desc : :py:obj:`~.cudaChannelFormatDesc` Channel format See Also -------- :py:obj:`~.cudaCreateChannelDesc (C API)`, :py:obj:`~.cudaCreateTextureObject`, :py:obj:`~.cudaCreateSurfaceObject` """ cdef cyruntime.cudaArray_const_t cyarray if array is None: parray = 0 elif isinstance(array, (cudaArray_const_t,)): parray = int(array) else: parray = int(cudaArray_const_t(array)) cyarray = parray cdef cudaChannelFormatDesc desc = cudaChannelFormatDesc() with nogil: err = cyruntime.cudaGetChannelDesc(desc._pvt_ptr, cyarray) if err != cyruntime.cudaSuccess: return (_dict_cudaError_t[err], None) return (_dict_cudaError_t[err], desc) @cython.embedsignature(True) def cudaCreateChannelDesc(int x, int y, int z, int w, f not None : cudaChannelFormatKind): """ Returns a channel descriptor using the specified format. Returns a channel descriptor with format `f` and number of bits of each component `x`, `y`, `z`, and `w`. The :py:obj:`~.cudaChannelFormatDesc` is defined as: **View CUDA Toolkit Documentation for a C++ code example** where :py:obj:`~.cudaChannelFormatKind` is one of :py:obj:`~.cudaChannelFormatKindSigned`, :py:obj:`~.cudaChannelFormatKindUnsigned`, or :py:obj:`~.cudaChannelFormatKindFloat`. Parameters ---------- x : int X component y : int Y component z : int Z component w : int W component f : :py:obj:`~.cudaChannelFormatKind` Channel format Returns ------- cudaError_t.cudaSuccess cudaError_t.cudaSuccess :py:obj:`~.cudaChannelFormatDesc` Channel descriptor with format `f` See Also -------- cudaCreateChannelDesc (C++ API), :py:obj:`~.cudaGetChannelDesc`, :py:obj:`~.cudaCreateTextureObject`, :py:obj:`~.cudaCreateSurfaceObject` """ cdef cyruntime.cudaChannelFormatKind cyf = f.value with nogil: err = cyruntime.cudaCreateChannelDesc(x, y, z, w, cyf) cdef cudaChannelFormatDesc wrapper = cudaChannelFormatDesc() wrapper._pvt_ptr[0] = err return (cudaError_t.cudaSuccess, wrapper) @cython.embedsignature(True) def cudaCreateTextureObject(pResDesc : Optional[cudaResourceDesc], pTexDesc : Optional[cudaTextureDesc], pResViewDesc : Optional[cudaResourceViewDesc]): """ Creates a texture object. Creates a texture object and returns it in `pTexObject`. `pResDesc` describes the data to texture from. `pTexDesc` describes how the data should be sampled. `pResViewDesc` is an optional argument that specifies an alternate format for the data described by `pResDesc`, and also describes the subresource region to restrict access to when texturing. `pResViewDesc` can only be specified if the type of resource is a CUDA array or a CUDA mipmapped array not in a block compressed format. Texture objects are only supported on devices of compute capability 3.0 or higher. Additionally, a texture object is an opaque value, and, as such, should only be accessed through CUDA API calls. The :py:obj:`~.cudaResourceDesc` structure is defined as: **View CUDA Toolkit Documentation for a C++ code example** where: - :py:obj:`~.cudaResourceDesc.resType` specifies the type of resource to texture from. CUresourceType is defined as: - **View CUDA Toolkit Documentation for a C++ code example** If :py:obj:`~.cudaResourceDesc.resType` is set to :py:obj:`~.cudaResourceTypeArray`, :py:obj:`~.cudaResourceDesc`::res::array::array must be set to a valid CUDA array handle. If :py:obj:`~.cudaResourceDesc.resType` is set to :py:obj:`~.cudaResourceTypeMipmappedArray`, :py:obj:`~.cudaResourceDesc`::res::mipmap::mipmap must be set to a valid CUDA mipmapped array handle and :py:obj:`~.cudaTextureDesc.normalizedCoords` must be set to true. If :py:obj:`~.cudaResourceDesc.resType` is set to :py:obj:`~.cudaResourceTypeLinear`, :py:obj:`~.cudaResourceDesc`::res::linear::devPtr must be set to a valid device pointer, that is aligned to :py:obj:`~.cudaDeviceProp.textureAlignment`. :py:obj:`~.cudaResourceDesc`::res::linear::desc describes the format and the number of components per array element. :py:obj:`~.cudaResourceDesc`::res::linear::sizeInBytes specifies the size of the array in bytes. The total number of elements in the linear address range cannot exceed :py:obj:`~.cudaDeviceProp.maxTexture1DLinear`. The number of elements is computed as (sizeInBytes / sizeof(desc)). If :py:obj:`~.cudaResourceDesc.resType` is set to :py:obj:`~.cudaResourceTypePitch2D`, :py:obj:`~.cudaResourceDesc`::res::pitch2D::devPtr must be set to a valid device pointer, that is aligned to :py:obj:`~.cudaDeviceProp.textureAlignment`. :py:obj:`~.cudaResourceDesc`::res::pitch2D::desc describes the format and the number of components per array element. :py:obj:`~.cudaResourceDesc`::res::pitch2D::width and :py:obj:`~.cudaResourceDesc`::res::pitch2D::height specify the width and height of the array in elements, and cannot exceed :py:obj:`~.cudaDeviceProp.maxTexture2DLinear`[0] and :py:obj:`~.cudaDeviceProp.maxTexture2DLinear`[1] respectively. :py:obj:`~.cudaResourceDesc`::res::pitch2D::pitchInBytes specifies the pitch between two rows in bytes and has to be aligned to :py:obj:`~.cudaDeviceProp.texturePitchAlignment`. Pitch cannot exceed :py:obj:`~.cudaDeviceProp.maxTexture2DLinear`[2]. The :py:obj:`~.cudaTextureDesc` struct is defined as **View CUDA Toolkit Documentation for a C++ code example** where - :py:obj:`~.cudaTextureDesc.addressMode` specifies the addressing mode for each dimension of the texture data. :py:obj:`~.cudaTextureAddressMode` is defined as: - **View CUDA Toolkit Documentation for a C++ code example** - This is ignored if :py:obj:`~.cudaResourceDesc.resType` is :py:obj:`~.cudaResourceTypeLinear`. Also, if :py:obj:`~.cudaTextureDesc.normalizedCoords` is set to zero, :py:obj:`~.cudaAddressModeWrap` and :py:obj:`~.cudaAddressModeMirror` won't be supported and will be switched to :py:obj:`~.cudaAddressModeClamp`. - :py:obj:`~.cudaTextureDesc.filterMode` specifies the filtering mode to be used when fetching from the texture. :py:obj:`~.cudaTextureFilterMode` is defined as: - **View CUDA Toolkit Documentation for a C++ code example** - This is ignored if :py:obj:`~.cudaResourceDesc.resType` is :py:obj:`~.cudaResourceTypeLinear`. - :py:obj:`~.cudaTextureDesc.readMode` specifies whether integer data should be converted to floating point or not. :py:obj:`~.cudaTextureReadMode` is defined as: - **View CUDA Toolkit Documentation for a C++ code example** - Note that this applies only to 8-bit and 16-bit integer formats. 32-bit integer format would not be promoted, regardless of whether or not this :py:obj:`~.cudaTextureDesc.readMode` is set :py:obj:`~.cudaReadModeNormalizedFloat` is specified. - :py:obj:`~.cudaTextureDesc.sRGB` specifies whether sRGB to linear conversion should be performed during texture fetch. - :py:obj:`~.cudaTextureDesc.borderColor` specifies the float values of color. where: :py:obj:`~.cudaTextureDesc.borderColor`[0] contains value of 'R', :py:obj:`~.cudaTextureDesc.borderColor`[1] contains value of 'G', :py:obj:`~.cudaTextureDesc.borderColor`[2] contains value of 'B', :py:obj:`~.cudaTextureDesc.borderColor`[3] contains value of 'A' Note that application using integer border color values will need to these values to float. The values are set only when the addressing mode specified by :py:obj:`~.cudaTextureDesc.addressMode` is cudaAddressModeBorder. - :py:obj:`~.cudaTextureDesc.normalizedCoords` specifies whether the texture coordinates will be normalized or not. - :py:obj:`~.cudaTextureDesc.maxAnisotropy` specifies the maximum anistropy ratio to be used when doing anisotropic filtering. This value will be clamped to the range [1,16]. - :py:obj:`~.cudaTextureDesc.mipmapFilterMode` specifies the filter mode when the calculated mipmap level lies between two defined mipmap levels. - :py:obj:`~.cudaTextureDesc.mipmapLevelBias` specifies the offset to be applied to the calculated mipmap level. - :py:obj:`~.cudaTextureDesc.minMipmapLevelClamp` specifies the lower end of the mipmap level range to clamp access to. - :py:obj:`~.cudaTextureDesc.maxMipmapLevelClamp` specifies the upper end of the mipmap level range to clamp access to. - :py:obj:`~.cudaTextureDesc.disableTrilinearOptimization` specifies whether the trilinear filtering optimizations will be disabled. - :py:obj:`~.cudaTextureDesc.seamlessCubemap` specifies whether seamless cube map filtering is enabled. This flag can only be specified if the underlying resource is a CUDA array or a CUDA mipmapped array that was created with the flag :py:obj:`~.cudaArrayCubemap`. When seamless cube map filtering is enabled, texture address modes specified by :py:obj:`~.cudaTextureDesc.addressMode` are ignored. Instead, if the :py:obj:`~.cudaTextureDesc.filterMode` is set to :py:obj:`~.cudaFilterModePoint` the address mode :py:obj:`~.cudaAddressModeClamp` will be applied for all dimensions. If the :py:obj:`~.cudaTextureDesc.filterMode` is set to :py:obj:`~.cudaFilterModeLinear` seamless cube map filtering will be performed when sampling along the cube face borders. The :py:obj:`~.cudaResourceViewDesc` struct is defined as **View CUDA Toolkit Documentation for a C++ code example** where: - :py:obj:`~.cudaResourceViewDesc.format` specifies how the data contained in the CUDA array or CUDA mipmapped array should be interpreted. Note that this can incur a change in size of the texture data. If the resource view format is a block compressed format, then the underlying CUDA array or CUDA mipmapped array has to have a 32-bit unsigned integer format with 2 or 4 channels, depending on the block compressed format. For ex., BC1 and BC4 require the underlying CUDA array to have a 32-bit unsigned int with 2 channels. The other BC formats require the underlying resource to have the same 32-bit unsigned int format but with 4 channels. - :py:obj:`~.cudaResourceViewDesc.width` specifies the new width of the texture data. If the resource view format is a block compressed format, this value has to be 4 times the original width of the resource. For non block compressed formats, this value has to be equal to that of the original resource. - :py:obj:`~.cudaResourceViewDesc.height` specifies the new height of the texture data. If the resource view format is a block compressed format, this value has to be 4 times the original height of the resource. For non block compressed formats, this value has to be equal to that of the original resource. - :py:obj:`~.cudaResourceViewDesc.depth` specifies the new depth of the texture data. This value has to be equal to that of the original resource. - :py:obj:`~.cudaResourceViewDesc.firstMipmapLevel` specifies the most detailed mipmap level. This will be the new mipmap level zero. For non-mipmapped resources, this value has to be zero.:py:obj:`~.cudaTextureDesc.minMipmapLevelClamp` and :py:obj:`~.cudaTextureDesc.maxMipmapLevelClamp` will be relative to this value. For ex., if the firstMipmapLevel is set to 2, and a minMipmapLevelClamp of 1.2 is specified, then the actual minimum mipmap level clamp will be 3.2. - :py:obj:`~.cudaResourceViewDesc.lastMipmapLevel` specifies the least detailed mipmap level. For non-mipmapped resources, this value has to be zero. - :py:obj:`~.cudaResourceViewDesc.firstLayer` specifies the first layer index for layered textures. This will be the new layer zero. For non- layered resources, this value has to be zero. - :py:obj:`~.cudaResourceViewDesc.lastLayer` specifies the last layer index for layered textures. For non-layered resources, this value has to be zero. Parameters ---------- pResDesc : :py:obj:`~.cudaResourceDesc` Resource descriptor pTexDesc : :py:obj:`~.cudaTextureDesc` Texture descriptor pResViewDesc : :py:obj:`~.cudaResourceViewDesc` Resource view descriptor Returns ------- cudaError_t :py:obj:`~.cudaSuccess`, :py:obj:`~.cudaErrorInvalidValue` pTexObject : :py:obj:`~.cudaTextureObject_t` Texture object to create See Also -------- :py:obj:`~.cudaDestroyTextureObject`, :py:obj:`~.cuTexObjectCreate` """ cdef cudaTextureObject_t pTexObject = cudaTextureObject_t() cdef cyruntime.cudaResourceDesc* cypResDesc_ptr = pResDesc._pvt_ptr if pResDesc is not None else NULL cdef cyruntime.cudaTextureDesc* cypTexDesc_ptr = pTexDesc._pvt_ptr if pTexDesc is not None else NULL cdef cyruntime.cudaResourceViewDesc* cypResViewDesc_ptr = pResViewDesc._pvt_ptr if pResViewDesc is not None else NULL with nogil: err = cyruntime.cudaCreateTextureObject(pTexObject._pvt_ptr, cypResDesc_ptr, cypTexDesc_ptr, cypResViewDesc_ptr) if err != cyruntime.cudaSuccess: return (_dict_cudaError_t[err], None) return (_dict_cudaError_t[err], pTexObject) @cython.embedsignature(True) def cudaDestroyTextureObject(texObject): """ Destroys a texture object. Destroys the texture object specified by `texObject`. Parameters ---------- texObject : :py:obj:`~.cudaTextureObject_t` Texture object to destroy Returns ------- cudaError_t :py:obj:`~.cudaSuccess`, :py:obj:`~.cudaErrorInvalidValue` See Also -------- :py:obj:`~.cudaCreateTextureObject`, :py:obj:`~.cuTexObjectDestroy` """ cdef cyruntime.cudaTextureObject_t cytexObject if texObject is None: ptexObject = 0 elif isinstance(texObject, (cudaTextureObject_t,)): ptexObject = int(texObject) else: ptexObject = int(cudaTextureObject_t(texObject)) cytexObject = ptexObject with nogil: err = cyruntime.cudaDestroyTextureObject(cytexObject) return (_dict_cudaError_t[err],) @cython.embedsignature(True) def cudaGetTextureObjectResourceDesc(texObject): """ Returns a texture object's resource descriptor. Returns the resource descriptor for the texture object specified by `texObject`. Parameters ---------- texObject : :py:obj:`~.cudaTextureObject_t` Texture object Returns ------- cudaError_t :py:obj:`~.cudaSuccess`, :py:obj:`~.cudaErrorInvalidValue` pResDesc : :py:obj:`~.cudaResourceDesc` Resource descriptor See Also -------- :py:obj:`~.cudaCreateTextureObject`, :py:obj:`~.cuTexObjectGetResourceDesc` """ cdef cyruntime.cudaTextureObject_t cytexObject if texObject is None: ptexObject = 0 elif isinstance(texObject, (cudaTextureObject_t,)): ptexObject = int(texObject) else: ptexObject = int(cudaTextureObject_t(texObject)) cytexObject = ptexObject cdef cudaResourceDesc pResDesc = cudaResourceDesc() with nogil: err = cyruntime.cudaGetTextureObjectResourceDesc(pResDesc._pvt_ptr, cytexObject) if err != cyruntime.cudaSuccess: return (_dict_cudaError_t[err], None) return (_dict_cudaError_t[err], pResDesc) @cython.embedsignature(True) def cudaGetTextureObjectTextureDesc(texObject): """ Returns a texture object's texture descriptor. Returns the texture descriptor for the texture object specified by `texObject`. Parameters ---------- texObject : :py:obj:`~.cudaTextureObject_t` Texture object Returns ------- cudaError_t :py:obj:`~.cudaSuccess`, :py:obj:`~.cudaErrorInvalidValue` pTexDesc : :py:obj:`~.cudaTextureDesc` Texture descriptor See Also -------- :py:obj:`~.cudaCreateTextureObject`, :py:obj:`~.cuTexObjectGetTextureDesc` """ cdef cyruntime.cudaTextureObject_t cytexObject if texObject is None: ptexObject = 0 elif isinstance(texObject, (cudaTextureObject_t,)): ptexObject = int(texObject) else: ptexObject = int(cudaTextureObject_t(texObject)) cytexObject = ptexObject cdef cudaTextureDesc pTexDesc = cudaTextureDesc() with nogil: err = cyruntime.cudaGetTextureObjectTextureDesc(pTexDesc._pvt_ptr, cytexObject) if err != cyruntime.cudaSuccess: return (_dict_cudaError_t[err], None) return (_dict_cudaError_t[err], pTexDesc) @cython.embedsignature(True) def cudaGetTextureObjectResourceViewDesc(texObject): """ Returns a texture object's resource view descriptor. Returns the resource view descriptor for the texture object specified by `texObject`. If no resource view was specified, :py:obj:`~.cudaErrorInvalidValue` is returned. Parameters ---------- texObject : :py:obj:`~.cudaTextureObject_t` Texture object Returns ------- cudaError_t :py:obj:`~.cudaSuccess`, :py:obj:`~.cudaErrorInvalidValue` pResViewDesc : :py:obj:`~.cudaResourceViewDesc` Resource view descriptor See Also -------- :py:obj:`~.cudaCreateTextureObject`, :py:obj:`~.cuTexObjectGetResourceViewDesc` """ cdef cyruntime.cudaTextureObject_t cytexObject if texObject is None: ptexObject = 0 elif isinstance(texObject, (cudaTextureObject_t,)): ptexObject = int(texObject) else: ptexObject = int(cudaTextureObject_t(texObject)) cytexObject = ptexObject cdef cudaResourceViewDesc pResViewDesc = cudaResourceViewDesc() with nogil: err = cyruntime.cudaGetTextureObjectResourceViewDesc(pResViewDesc._pvt_ptr, cytexObject) if err != cyruntime.cudaSuccess: return (_dict_cudaError_t[err], None) return (_dict_cudaError_t[err], pResViewDesc) @cython.embedsignature(True) def cudaCreateSurfaceObject(pResDesc : Optional[cudaResourceDesc]): """ Creates a surface object. Creates a surface object and returns it in `pSurfObject`. `pResDesc` describes the data to perform surface load/stores on. :py:obj:`~.cudaResourceDesc.resType` must be :py:obj:`~.cudaResourceTypeArray` and :py:obj:`~.cudaResourceDesc`::res::array::array must be set to a valid CUDA array handle. Surface objects are only supported on devices of compute capability 3.0 or higher. Additionally, a surface object is an opaque value, and, as such, should only be accessed through CUDA API calls. Parameters ---------- pResDesc : :py:obj:`~.cudaResourceDesc` Resource descriptor Returns ------- cudaError_t :py:obj:`~.cudaSuccess`, :py:obj:`~.cudaErrorInvalidValue`, :py:obj:`~.cudaErrorInvalidChannelDescriptor`, :py:obj:`~.cudaErrorInvalidResourceHandle` pSurfObject : :py:obj:`~.cudaSurfaceObject_t` Surface object to create See Also -------- :py:obj:`~.cudaDestroySurfaceObject`, :py:obj:`~.cuSurfObjectCreate` """ cdef cudaSurfaceObject_t pSurfObject = cudaSurfaceObject_t() cdef cyruntime.cudaResourceDesc* cypResDesc_ptr = pResDesc._pvt_ptr if pResDesc is not None else NULL with nogil: err = cyruntime.cudaCreateSurfaceObject(pSurfObject._pvt_ptr, cypResDesc_ptr) if err != cyruntime.cudaSuccess: return (_dict_cudaError_t[err], None) return (_dict_cudaError_t[err], pSurfObject) @cython.embedsignature(True) def cudaDestroySurfaceObject(surfObject): """ Destroys a surface object. Destroys the surface object specified by `surfObject`. Parameters ---------- surfObject : :py:obj:`~.cudaSurfaceObject_t` Surface object to destroy Returns ------- cudaError_t :py:obj:`~.cudaSuccess`, :py:obj:`~.cudaErrorInvalidValue` See Also -------- :py:obj:`~.cudaCreateSurfaceObject`, :py:obj:`~.cuSurfObjectDestroy` """ cdef cyruntime.cudaSurfaceObject_t cysurfObject if surfObject is None: psurfObject = 0 elif isinstance(surfObject, (cudaSurfaceObject_t,)): psurfObject = int(surfObject) else: psurfObject = int(cudaSurfaceObject_t(surfObject)) cysurfObject = psurfObject with nogil: err = cyruntime.cudaDestroySurfaceObject(cysurfObject) return (_dict_cudaError_t[err],) @cython.embedsignature(True) def cudaGetSurfaceObjectResourceDesc(surfObject): """ Returns a surface object's resource descriptor Returns the resource descriptor for the surface object specified by `surfObject`. Parameters ---------- surfObject : :py:obj:`~.cudaSurfaceObject_t` Surface object Returns ------- cudaError_t :py:obj:`~.cudaSuccess`, :py:obj:`~.cudaErrorInvalidValue` pResDesc : :py:obj:`~.cudaResourceDesc` Resource descriptor See Also -------- :py:obj:`~.cudaCreateSurfaceObject`, :py:obj:`~.cuSurfObjectGetResourceDesc` """ cdef cyruntime.cudaSurfaceObject_t cysurfObject if surfObject is None: psurfObject = 0 elif isinstance(surfObject, (cudaSurfaceObject_t,)): psurfObject = int(surfObject) else: psurfObject = int(cudaSurfaceObject_t(surfObject)) cysurfObject = psurfObject cdef cudaResourceDesc pResDesc = cudaResourceDesc() with nogil: err = cyruntime.cudaGetSurfaceObjectResourceDesc(pResDesc._pvt_ptr, cysurfObject) if err != cyruntime.cudaSuccess: return (_dict_cudaError_t[err], None) return (_dict_cudaError_t[err], pResDesc) @cython.embedsignature(True) def cudaDriverGetVersion(): """ Returns the latest version of CUDA supported by the driver. Returns in `*driverVersion` the latest version of CUDA supported by the driver. The version is returned as (1000 * major + 10 * minor). For example, CUDA 9.2 would be represented by 9020. If no driver is installed, then 0 is returned as the driver version. This function automatically returns :py:obj:`~.cudaErrorInvalidValue` if `driverVersion` is NULL. Returns ------- cudaError_t :py:obj:`~.cudaSuccess`, :py:obj:`~.cudaErrorInvalidValue` driverVersion : int Returns the CUDA driver version. See Also -------- :py:obj:`~.cudaRuntimeGetVersion`, :py:obj:`~.cuDriverGetVersion` """ cdef int driverVersion = 0 with nogil: err = cyruntime.cudaDriverGetVersion(&driverVersion) if err != cyruntime.cudaSuccess: return (_dict_cudaError_t[err], None) return (_dict_cudaError_t[err], driverVersion) @cython.embedsignature(True) def cudaRuntimeGetVersion(): """ Returns the CUDA Runtime version. Returns in `*runtimeVersion` the version number of the current CUDA Runtime instance. The version is returned as (1000 * major + 10 * minor). For example, CUDA 9.2 would be represented by 9020. As of CUDA 12.0, this function no longer initializes CUDA. The purpose of this API is solely to return a compile-time constant stating the CUDA Toolkit version in the above format. This function automatically returns :py:obj:`~.cudaErrorInvalidValue` if the `runtimeVersion` argument is NULL. Returns ------- cudaError_t :py:obj:`~.cudaSuccess`, :py:obj:`~.cudaErrorInvalidValue` runtimeVersion : int Returns the CUDA Runtime version. See Also -------- :py:obj:`~.cudaDriverGetVersion`, :py:obj:`~.cuDriverGetVersion` """ cdef int runtimeVersion = 0 with nogil: err = cyruntime.cudaRuntimeGetVersion(&runtimeVersion) if err != cyruntime.cudaSuccess: return (_dict_cudaError_t[err], None) return (_dict_cudaError_t[err], runtimeVersion) @cython.embedsignature(True) def cudaGraphCreate(unsigned int flags): """ Creates a graph. Creates an empty graph, which is returned via `pGraph`. Parameters ---------- flags : unsigned int Graph creation flags, must be 0 Returns ------- cudaError_t :py:obj:`~.cudaSuccess`, :py:obj:`~.cudaErrorInvalidValue`, :py:obj:`~.cudaErrorMemoryAllocation` pGraph : :py:obj:`~.cudaGraph_t` Returns newly created graph See Also -------- :py:obj:`~.cudaGraphAddChildGraphNode`, :py:obj:`~.cudaGraphAddEmptyNode`, :py:obj:`~.cudaGraphAddKernelNode`, :py:obj:`~.cudaGraphAddHostNode`, :py:obj:`~.cudaGraphAddMemcpyNode`, :py:obj:`~.cudaGraphAddMemsetNode`, :py:obj:`~.cudaGraphInstantiate`, :py:obj:`~.cudaGraphDestroy`, :py:obj:`~.cudaGraphGetNodes`, :py:obj:`~.cudaGraphGetRootNodes`, :py:obj:`~.cudaGraphGetEdges`, :py:obj:`~.cudaGraphClone` """ cdef cudaGraph_t pGraph = cudaGraph_t() with nogil: err = cyruntime.cudaGraphCreate(pGraph._pvt_ptr, flags) if err != cyruntime.cudaSuccess: return (_dict_cudaError_t[err], None) return (_dict_cudaError_t[err], pGraph) @cython.embedsignature(True) def cudaGraphAddKernelNode(graph, pDependencies : Optional[tuple[cudaGraphNode_t] | list[cudaGraphNode_t]], size_t numDependencies, pNodeParams : Optional[cudaKernelNodeParams]): """ Creates a kernel execution node and adds it to a graph. Creates a new kernel execution node and adds it to `graph` with `numDependencies` dependencies specified via `pDependencies` and arguments specified in `pNodeParams`. It is possible for `numDependencies` to be 0, in which case the node will be placed at the root of the graph. `pDependencies` may not have any duplicate entries. A handle to the new node will be returned in `pGraphNode`. The :py:obj:`~.cudaKernelNodeParams` structure is defined as: **View CUDA Toolkit Documentation for a C++ code example** When the graph is launched, the node will invoke kernel `func` on a (`gridDim.x` x `gridDim.y` x `gridDim.z`) grid of blocks. Each block contains (`blockDim.x` x `blockDim.y` x `blockDim.z`) threads. `sharedMem` sets the amount of dynamic shared memory that will be available to each thread block. Kernel parameters to `func` can be specified in one of two ways: 1) Kernel parameters can be specified via `kernelParams`. If the kernel has N parameters, then `kernelParams` needs to be an array of N pointers. Each pointer, from `kernelParams`[0] to `kernelParams`[N-1], points to the region of memory from which the actual parameter will be copied. The number of kernel parameters and their offsets and sizes do not need to be specified as that information is retrieved directly from the kernel's image. 2) Kernel parameters can also be packaged by the application into a single buffer that is passed in via `extra`. This places the burden on the application of knowing each kernel parameter's size and alignment/padding within the buffer. The `extra` parameter exists to allow this function to take additional less commonly used arguments. `extra` specifies a list of names of extra settings and their corresponding values. Each extra setting name is immediately followed by the corresponding value. The list must be terminated with either NULL or CU_LAUNCH_PARAM_END. - :py:obj:`~.CU_LAUNCH_PARAM_END`, which indicates the end of the `extra` array; - :py:obj:`~.CU_LAUNCH_PARAM_BUFFER_POINTER`, which specifies that the next value in `extra` will be a pointer to a buffer containing all the kernel parameters for launching kernel `func`; - :py:obj:`~.CU_LAUNCH_PARAM_BUFFER_SIZE`, which specifies that the next value in `extra` will be a pointer to a size_t containing the size of the buffer specified with :py:obj:`~.CU_LAUNCH_PARAM_BUFFER_POINTER`; The error :py:obj:`~.cudaErrorInvalidValue` will be returned if kernel parameters are specified with both `kernelParams` and `extra` (i.e. both `kernelParams` and `extra` are non-NULL). The `kernelParams` or `extra` array, as well as the argument values it points to, are copied during this call. Parameters ---------- graph : :py:obj:`~.CUgraph` or :py:obj:`~.cudaGraph_t` Graph to which to add the node pDependencies : list[:py:obj:`~.cudaGraphNode_t`] Dependencies of the node numDependencies : size_t Number of dependencies pNodeParams : :py:obj:`~.cudaKernelNodeParams` Parameters for the GPU execution node Returns ------- cudaError_t :py:obj:`~.cudaSuccess`, :py:obj:`~.cudaErrorInvalidValue`, :py:obj:`~.cudaErrorInvalidDeviceFunction` pGraphNode : :py:obj:`~.cudaGraphNode_t` Returns newly created node See Also -------- :py:obj:`~.cudaGraphAddNode`, :py:obj:`~.cudaLaunchKernel`, :py:obj:`~.cudaGraphKernelNodeGetParams`, :py:obj:`~.cudaGraphKernelNodeSetParams`, :py:obj:`~.cudaGraphCreate`, :py:obj:`~.cudaGraphDestroyNode`, :py:obj:`~.cudaGraphAddChildGraphNode`, :py:obj:`~.cudaGraphAddEmptyNode`, :py:obj:`~.cudaGraphAddHostNode`, :py:obj:`~.cudaGraphAddMemcpyNode`, :py:obj:`~.cudaGraphAddMemsetNode` Notes ----- Kernels launched using graphs must not use texture and surface references. Reading or writing through any texture or surface reference is undefined behavior. This restriction does not apply to texture and surface objects. """ pDependencies = [] if pDependencies is None else pDependencies if not all(isinstance(_x, (cudaGraphNode_t,driver.CUgraphNode)) for _x in pDependencies): raise TypeError("Argument 'pDependencies' is not instance of type (expected tuple[cyruntime.cudaGraphNode_t,driver.CUgraphNode] or list[cyruntime.cudaGraphNode_t,driver.CUgraphNode]") cdef cyruntime.cudaGraph_t cygraph if graph is None: pgraph = 0 elif isinstance(graph, (cudaGraph_t,driver.CUgraph)): pgraph = int(graph) else: pgraph = int(cudaGraph_t(graph)) cygraph = pgraph cdef cudaGraphNode_t pGraphNode = cudaGraphNode_t() cdef cyruntime.cudaGraphNode_t* cypDependencies = NULL if len(pDependencies) > 1: cypDependencies = calloc(len(pDependencies), sizeof(cyruntime.cudaGraphNode_t)) if cypDependencies is NULL: raise MemoryError('Failed to allocate length x size memory: ' + str(len(pDependencies)) + 'x' + str(sizeof(cyruntime.cudaGraphNode_t))) else: for idx in range(len(pDependencies)): cypDependencies[idx] = (pDependencies[idx])._pvt_ptr[0] elif len(pDependencies) == 1: cypDependencies = (pDependencies[0])._pvt_ptr if numDependencies > len(pDependencies): raise RuntimeError("List is too small: " + str(len(pDependencies)) + " < " + str(numDependencies)) cdef cyruntime.cudaKernelNodeParams* cypNodeParams_ptr = pNodeParams._pvt_ptr if pNodeParams is not None else NULL with nogil: err = cyruntime.cudaGraphAddKernelNode(pGraphNode._pvt_ptr, cygraph, cypDependencies, numDependencies, cypNodeParams_ptr) if len(pDependencies) > 1 and cypDependencies is not NULL: free(cypDependencies) if err != cyruntime.cudaSuccess: return (_dict_cudaError_t[err], None) return (_dict_cudaError_t[err], pGraphNode) @cython.embedsignature(True) def cudaGraphKernelNodeGetParams(node): """ Returns a kernel node's parameters. Returns the parameters of kernel node `node` in `pNodeParams`. The `kernelParams` or `extra` array returned in `pNodeParams`, as well as the argument values it points to, are owned by the node. This memory remains valid until the node is destroyed or its parameters are modified, and should not be modified directly. Use :py:obj:`~.cudaGraphKernelNodeSetParams` to update the parameters of this node. The params will contain either `kernelParams` or `extra`, according to which of these was most recently set on the node. Parameters ---------- node : :py:obj:`~.CUgraphNode` or :py:obj:`~.cudaGraphNode_t` Node to get the parameters for Returns ------- cudaError_t :py:obj:`~.cudaSuccess`, :py:obj:`~.cudaErrorInvalidValue`, :py:obj:`~.cudaErrorInvalidDeviceFunction` pNodeParams : :py:obj:`~.cudaKernelNodeParams` Pointer to return the parameters See Also -------- :py:obj:`~.cudaLaunchKernel`, :py:obj:`~.cudaGraphAddKernelNode`, :py:obj:`~.cudaGraphKernelNodeSetParams` """ cdef cyruntime.cudaGraphNode_t cynode if node is None: pnode = 0 elif isinstance(node, (cudaGraphNode_t,driver.CUgraphNode)): pnode = int(node) else: pnode = int(cudaGraphNode_t(node)) cynode = pnode cdef cudaKernelNodeParams pNodeParams = cudaKernelNodeParams() with nogil: err = cyruntime.cudaGraphKernelNodeGetParams(cynode, pNodeParams._pvt_ptr) if err != cyruntime.cudaSuccess: return (_dict_cudaError_t[err], None) return (_dict_cudaError_t[err], pNodeParams) @cython.embedsignature(True) def cudaGraphKernelNodeSetParams(node, pNodeParams : Optional[cudaKernelNodeParams]): """ Sets a kernel node's parameters. Sets the parameters of kernel node `node` to `pNodeParams`. Parameters ---------- node : :py:obj:`~.CUgraphNode` or :py:obj:`~.cudaGraphNode_t` Node to set the parameters for pNodeParams : :py:obj:`~.cudaKernelNodeParams` Parameters to copy Returns ------- cudaError_t :py:obj:`~.cudaSuccess`, :py:obj:`~.cudaErrorInvalidValue`, :py:obj:`~.cudaErrorInvalidResourceHandle`, :py:obj:`~.cudaErrorMemoryAllocation` See Also -------- :py:obj:`~.cudaGraphNodeSetParams`, :py:obj:`~.cudaLaunchKernel`, :py:obj:`~.cudaGraphAddKernelNode`, :py:obj:`~.cudaGraphKernelNodeGetParams` """ cdef cyruntime.cudaGraphNode_t cynode if node is None: pnode = 0 elif isinstance(node, (cudaGraphNode_t,driver.CUgraphNode)): pnode = int(node) else: pnode = int(cudaGraphNode_t(node)) cynode = pnode cdef cyruntime.cudaKernelNodeParams* cypNodeParams_ptr = pNodeParams._pvt_ptr if pNodeParams is not None else NULL with nogil: err = cyruntime.cudaGraphKernelNodeSetParams(cynode, cypNodeParams_ptr) return (_dict_cudaError_t[err],) @cython.embedsignature(True) def cudaGraphKernelNodeCopyAttributes(hSrc, hDst): """ Copies attributes from source node to destination node. Copies attributes from source node `src` to destination node `dst`. Both node must have the same context. Parameters ---------- dst : :py:obj:`~.CUgraphNode` or :py:obj:`~.cudaGraphNode_t` Destination node src : :py:obj:`~.CUgraphNode` or :py:obj:`~.cudaGraphNode_t` Source node For list of attributes see :py:obj:`~.cudaKernelNodeAttrID` Returns ------- cudaError_t :py:obj:`~.cudaSuccess`, :py:obj:`~.cudaErrorInvalidContext` See Also -------- :py:obj:`~.cudaAccessPolicyWindow` """ cdef cyruntime.cudaGraphNode_t cyhDst if hDst is None: phDst = 0 elif isinstance(hDst, (cudaGraphNode_t,driver.CUgraphNode)): phDst = int(hDst) else: phDst = int(cudaGraphNode_t(hDst)) cyhDst = phDst cdef cyruntime.cudaGraphNode_t cyhSrc if hSrc is None: phSrc = 0 elif isinstance(hSrc, (cudaGraphNode_t,driver.CUgraphNode)): phSrc = int(hSrc) else: phSrc = int(cudaGraphNode_t(hSrc)) cyhSrc = phSrc with nogil: err = cyruntime.cudaGraphKernelNodeCopyAttributes(cyhSrc, cyhDst) return (_dict_cudaError_t[err],) @cython.embedsignature(True) def cudaGraphKernelNodeGetAttribute(hNode, attr not None : cudaKernelNodeAttrID): """ Queries node attribute. Queries attribute `attr` from node `hNode` and stores it in corresponding member of `value_out`. Parameters ---------- hNode : :py:obj:`~.CUgraphNode` or :py:obj:`~.cudaGraphNode_t` attr : :py:obj:`~.cudaKernelNodeAttrID` Returns ------- cudaError_t :py:obj:`~.cudaSuccess`, :py:obj:`~.cudaErrorInvalidValue`, :py:obj:`~.cudaErrorInvalidResourceHandle` value_out : :py:obj:`~.cudaKernelNodeAttrValue` See Also -------- :py:obj:`~.cudaAccessPolicyWindow` """ cdef cyruntime.cudaGraphNode_t cyhNode if hNode is None: phNode = 0 elif isinstance(hNode, (cudaGraphNode_t,driver.CUgraphNode)): phNode = int(hNode) else: phNode = int(cudaGraphNode_t(hNode)) cyhNode = phNode cdef cyruntime.cudaKernelNodeAttrID cyattr = attr.value cdef cudaKernelNodeAttrValue value_out = cudaKernelNodeAttrValue() with nogil: err = cyruntime.cudaGraphKernelNodeGetAttribute(cyhNode, cyattr, value_out._pvt_ptr) if err != cyruntime.cudaSuccess: return (_dict_cudaError_t[err], None) return (_dict_cudaError_t[err], value_out) @cython.embedsignature(True) def cudaGraphKernelNodeSetAttribute(hNode, attr not None : cudaKernelNodeAttrID, value : Optional[cudaKernelNodeAttrValue]): """ Sets node attribute. Sets attribute `attr` on node `hNode` from corresponding attribute of `value`. Parameters ---------- hNode : :py:obj:`~.CUgraphNode` or :py:obj:`~.cudaGraphNode_t` attr : :py:obj:`~.cudaKernelNodeAttrID` value : :py:obj:`~.cudaKernelNodeAttrValue` Returns ------- cudaError_t :py:obj:`~.cudaSuccess`, :py:obj:`~.cudaErrorInvalidValue`, :py:obj:`~.cudaErrorInvalidResourceHandle` See Also -------- :py:obj:`~.cudaAccessPolicyWindow` """ cdef cyruntime.cudaGraphNode_t cyhNode if hNode is None: phNode = 0 elif isinstance(hNode, (cudaGraphNode_t,driver.CUgraphNode)): phNode = int(hNode) else: phNode = int(cudaGraphNode_t(hNode)) cyhNode = phNode cdef cyruntime.cudaKernelNodeAttrID cyattr = attr.value cdef cyruntime.cudaKernelNodeAttrValue* cyvalue_ptr = value._pvt_ptr if value is not None else NULL with nogil: err = cyruntime.cudaGraphKernelNodeSetAttribute(cyhNode, cyattr, cyvalue_ptr) return (_dict_cudaError_t[err],) @cython.embedsignature(True) def cudaGraphAddMemcpyNode(graph, pDependencies : Optional[tuple[cudaGraphNode_t] | list[cudaGraphNode_t]], size_t numDependencies, pCopyParams : Optional[cudaMemcpy3DParms]): """ Creates a memcpy node and adds it to a graph. Creates a new memcpy node and adds it to `graph` with `numDependencies` dependencies specified via `pDependencies`. It is possible for `numDependencies` to be 0, in which case the node will be placed at the root of the graph. `pDependencies` may not have any duplicate entries. A handle to the new node will be returned in `pGraphNode`. When the graph is launched, the node will perform the memcpy described by `pCopyParams`. See :py:obj:`~.cudaMemcpy3D()` for a description of the structure and its restrictions. Memcpy nodes have some additional restrictions with regards to managed memory, if the system contains at least one device which has a zero value for the device attribute :py:obj:`~.cudaDevAttrConcurrentManagedAccess`. Parameters ---------- graph : :py:obj:`~.CUgraph` or :py:obj:`~.cudaGraph_t` Graph to which to add the node pDependencies : list[:py:obj:`~.cudaGraphNode_t`] Dependencies of the node numDependencies : size_t Number of dependencies pCopyParams : :py:obj:`~.cudaMemcpy3DParms` Parameters for the memory copy Returns ------- cudaError_t :py:obj:`~.cudaSuccess`, :py:obj:`~.cudaErrorInvalidValue` pGraphNode : :py:obj:`~.cudaGraphNode_t` Returns newly created node See Also -------- :py:obj:`~.cudaGraphAddNode`, :py:obj:`~.cudaMemcpy3D`, :py:obj:`~.cudaGraphAddMemcpyNodeToSymbol`, :py:obj:`~.cudaGraphAddMemcpyNodeFromSymbol`, :py:obj:`~.cudaGraphAddMemcpyNode1D`, :py:obj:`~.cudaGraphMemcpyNodeGetParams`, :py:obj:`~.cudaGraphMemcpyNodeSetParams`, :py:obj:`~.cudaGraphCreate`, :py:obj:`~.cudaGraphDestroyNode`, :py:obj:`~.cudaGraphAddChildGraphNode`, :py:obj:`~.cudaGraphAddEmptyNode`, :py:obj:`~.cudaGraphAddKernelNode`, :py:obj:`~.cudaGraphAddHostNode`, :py:obj:`~.cudaGraphAddMemsetNode` """ pDependencies = [] if pDependencies is None else pDependencies if not all(isinstance(_x, (cudaGraphNode_t,driver.CUgraphNode)) for _x in pDependencies): raise TypeError("Argument 'pDependencies' is not instance of type (expected tuple[cyruntime.cudaGraphNode_t,driver.CUgraphNode] or list[cyruntime.cudaGraphNode_t,driver.CUgraphNode]") cdef cyruntime.cudaGraph_t cygraph if graph is None: pgraph = 0 elif isinstance(graph, (cudaGraph_t,driver.CUgraph)): pgraph = int(graph) else: pgraph = int(cudaGraph_t(graph)) cygraph = pgraph cdef cudaGraphNode_t pGraphNode = cudaGraphNode_t() cdef cyruntime.cudaGraphNode_t* cypDependencies = NULL if len(pDependencies) > 1: cypDependencies = calloc(len(pDependencies), sizeof(cyruntime.cudaGraphNode_t)) if cypDependencies is NULL: raise MemoryError('Failed to allocate length x size memory: ' + str(len(pDependencies)) + 'x' + str(sizeof(cyruntime.cudaGraphNode_t))) else: for idx in range(len(pDependencies)): cypDependencies[idx] = (pDependencies[idx])._pvt_ptr[0] elif len(pDependencies) == 1: cypDependencies = (pDependencies[0])._pvt_ptr if numDependencies > len(pDependencies): raise RuntimeError("List is too small: " + str(len(pDependencies)) + " < " + str(numDependencies)) cdef cyruntime.cudaMemcpy3DParms* cypCopyParams_ptr = pCopyParams._pvt_ptr if pCopyParams is not None else NULL with nogil: err = cyruntime.cudaGraphAddMemcpyNode(pGraphNode._pvt_ptr, cygraph, cypDependencies, numDependencies, cypCopyParams_ptr) if len(pDependencies) > 1 and cypDependencies is not NULL: free(cypDependencies) if err != cyruntime.cudaSuccess: return (_dict_cudaError_t[err], None) return (_dict_cudaError_t[err], pGraphNode) @cython.embedsignature(True) def cudaGraphAddMemcpyNode1D(graph, pDependencies : Optional[tuple[cudaGraphNode_t] | list[cudaGraphNode_t]], size_t numDependencies, dst, src, size_t count, kind not None : cudaMemcpyKind): """ Creates a 1D memcpy node and adds it to a graph. Creates a new 1D memcpy node and adds it to `graph` with `numDependencies` dependencies specified via `pDependencies`. It is possible for `numDependencies` to be 0, in which case the node will be placed at the root of the graph. `pDependencies` may not have any duplicate entries. A handle to the new node will be returned in `pGraphNode`. When the graph is launched, the node will copy `count` bytes from the memory area pointed to by `src` to the memory area pointed to by `dst`, where `kind` specifies the direction of the copy, and must be one of :py:obj:`~.cudaMemcpyHostToHost`, :py:obj:`~.cudaMemcpyHostToDevice`, :py:obj:`~.cudaMemcpyDeviceToHost`, :py:obj:`~.cudaMemcpyDeviceToDevice`, or :py:obj:`~.cudaMemcpyDefault`. Passing :py:obj:`~.cudaMemcpyDefault` is recommended, in which case the type of transfer is inferred from the pointer values. However, :py:obj:`~.cudaMemcpyDefault` is only allowed on systems that support unified virtual addressing. Launching a memcpy node with dst and src pointers that do not match the direction of the copy results in an undefined behavior. Memcpy nodes have some additional restrictions with regards to managed memory, if the system contains at least one device which has a zero value for the device attribute :py:obj:`~.cudaDevAttrConcurrentManagedAccess`. Parameters ---------- graph : :py:obj:`~.CUgraph` or :py:obj:`~.cudaGraph_t` Graph to which to add the node pDependencies : list[:py:obj:`~.cudaGraphNode_t`] Dependencies of the node numDependencies : size_t Number of dependencies dst : Any Destination memory address src : Any Source memory address count : size_t Size in bytes to copy kind : :py:obj:`~.cudaMemcpyKind` Type of transfer Returns ------- cudaError_t :py:obj:`~.cudaSuccess`, :py:obj:`~.cudaErrorInvalidValue` pGraphNode : :py:obj:`~.cudaGraphNode_t` Returns newly created node See Also -------- :py:obj:`~.cudaMemcpy`, :py:obj:`~.cudaGraphAddMemcpyNode`, :py:obj:`~.cudaGraphMemcpyNodeGetParams`, :py:obj:`~.cudaGraphMemcpyNodeSetParams`, :py:obj:`~.cudaGraphMemcpyNodeSetParams1D`, :py:obj:`~.cudaGraphCreate`, :py:obj:`~.cudaGraphDestroyNode`, :py:obj:`~.cudaGraphAddChildGraphNode`, :py:obj:`~.cudaGraphAddEmptyNode`, :py:obj:`~.cudaGraphAddKernelNode`, :py:obj:`~.cudaGraphAddHostNode`, :py:obj:`~.cudaGraphAddMemsetNode` """ pDependencies = [] if pDependencies is None else pDependencies if not all(isinstance(_x, (cudaGraphNode_t,driver.CUgraphNode)) for _x in pDependencies): raise TypeError("Argument 'pDependencies' is not instance of type (expected tuple[cyruntime.cudaGraphNode_t,driver.CUgraphNode] or list[cyruntime.cudaGraphNode_t,driver.CUgraphNode]") cdef cyruntime.cudaGraph_t cygraph if graph is None: pgraph = 0 elif isinstance(graph, (cudaGraph_t,driver.CUgraph)): pgraph = int(graph) else: pgraph = int(cudaGraph_t(graph)) cygraph = pgraph cdef cudaGraphNode_t pGraphNode = cudaGraphNode_t() cdef cyruntime.cudaGraphNode_t* cypDependencies = NULL if len(pDependencies) > 1: cypDependencies = calloc(len(pDependencies), sizeof(cyruntime.cudaGraphNode_t)) if cypDependencies is NULL: raise MemoryError('Failed to allocate length x size memory: ' + str(len(pDependencies)) + 'x' + str(sizeof(cyruntime.cudaGraphNode_t))) else: for idx in range(len(pDependencies)): cypDependencies[idx] = (pDependencies[idx])._pvt_ptr[0] elif len(pDependencies) == 1: cypDependencies = (pDependencies[0])._pvt_ptr cydst = _HelperInputVoidPtr(dst) cdef void* cydst_ptr = cydst.cptr cysrc = _HelperInputVoidPtr(src) cdef void* cysrc_ptr = cysrc.cptr cdef cyruntime.cudaMemcpyKind cykind = kind.value with nogil: err = cyruntime.cudaGraphAddMemcpyNode1D(pGraphNode._pvt_ptr, cygraph, cypDependencies, numDependencies, cydst_ptr, cysrc_ptr, count, cykind) if len(pDependencies) > 1 and cypDependencies is not NULL: free(cypDependencies) if err != cyruntime.cudaSuccess: return (_dict_cudaError_t[err], None) return (_dict_cudaError_t[err], pGraphNode) @cython.embedsignature(True) def cudaGraphMemcpyNodeGetParams(node): """ Returns a memcpy node's parameters. Returns the parameters of memcpy node `node` in `pNodeParams`. Parameters ---------- node : :py:obj:`~.CUgraphNode` or :py:obj:`~.cudaGraphNode_t` Node to get the parameters for Returns ------- cudaError_t :py:obj:`~.cudaSuccess`, :py:obj:`~.cudaErrorInvalidValue` pNodeParams : :py:obj:`~.cudaMemcpy3DParms` Pointer to return the parameters See Also -------- :py:obj:`~.cudaMemcpy3D`, :py:obj:`~.cudaGraphAddMemcpyNode`, :py:obj:`~.cudaGraphMemcpyNodeSetParams` """ cdef cyruntime.cudaGraphNode_t cynode if node is None: pnode = 0 elif isinstance(node, (cudaGraphNode_t,driver.CUgraphNode)): pnode = int(node) else: pnode = int(cudaGraphNode_t(node)) cynode = pnode cdef cudaMemcpy3DParms pNodeParams = cudaMemcpy3DParms() with nogil: err = cyruntime.cudaGraphMemcpyNodeGetParams(cynode, pNodeParams._pvt_ptr) if err != cyruntime.cudaSuccess: return (_dict_cudaError_t[err], None) return (_dict_cudaError_t[err], pNodeParams) @cython.embedsignature(True) def cudaGraphMemcpyNodeSetParams(node, pNodeParams : Optional[cudaMemcpy3DParms]): """ Sets a memcpy node's parameters. Sets the parameters of memcpy node `node` to `pNodeParams`. Parameters ---------- node : :py:obj:`~.CUgraphNode` or :py:obj:`~.cudaGraphNode_t` Node to set the parameters for pNodeParams : :py:obj:`~.cudaMemcpy3DParms` Parameters to copy Returns ------- cudaError_t :py:obj:`~.cudaSuccess`, :py:obj:`~.cudaErrorInvalidValue`, See Also -------- :py:obj:`~.cudaGraphNodeSetParams`, :py:obj:`~.cudaMemcpy3D`, :py:obj:`~.cudaGraphMemcpyNodeSetParamsToSymbol`, :py:obj:`~.cudaGraphMemcpyNodeSetParamsFromSymbol`, :py:obj:`~.cudaGraphMemcpyNodeSetParams1D`, :py:obj:`~.cudaGraphAddMemcpyNode`, :py:obj:`~.cudaGraphMemcpyNodeGetParams` """ cdef cyruntime.cudaGraphNode_t cynode if node is None: pnode = 0 elif isinstance(node, (cudaGraphNode_t,driver.CUgraphNode)): pnode = int(node) else: pnode = int(cudaGraphNode_t(node)) cynode = pnode cdef cyruntime.cudaMemcpy3DParms* cypNodeParams_ptr = pNodeParams._pvt_ptr if pNodeParams is not None else NULL with nogil: err = cyruntime.cudaGraphMemcpyNodeSetParams(cynode, cypNodeParams_ptr) return (_dict_cudaError_t[err],) @cython.embedsignature(True) def cudaGraphMemcpyNodeSetParams1D(node, dst, src, size_t count, kind not None : cudaMemcpyKind): """ Sets a memcpy node's parameters to perform a 1-dimensional copy. Sets the parameters of memcpy node `node` to the copy described by the provided parameters. When the graph is launched, the node will copy `count` bytes from the memory area pointed to by `src` to the memory area pointed to by `dst`, where `kind` specifies the direction of the copy, and must be one of :py:obj:`~.cudaMemcpyHostToHost`, :py:obj:`~.cudaMemcpyHostToDevice`, :py:obj:`~.cudaMemcpyDeviceToHost`, :py:obj:`~.cudaMemcpyDeviceToDevice`, or :py:obj:`~.cudaMemcpyDefault`. Passing :py:obj:`~.cudaMemcpyDefault` is recommended, in which case the type of transfer is inferred from the pointer values. However, :py:obj:`~.cudaMemcpyDefault` is only allowed on systems that support unified virtual addressing. Launching a memcpy node with dst and src pointers that do not match the direction of the copy results in an undefined behavior. Parameters ---------- node : :py:obj:`~.CUgraphNode` or :py:obj:`~.cudaGraphNode_t` Node to set the parameters for dst : Any Destination memory address src : Any Source memory address count : size_t Size in bytes to copy kind : :py:obj:`~.cudaMemcpyKind` Type of transfer Returns ------- cudaError_t :py:obj:`~.cudaSuccess`, :py:obj:`~.cudaErrorInvalidValue` See Also -------- :py:obj:`~.cudaMemcpy`, :py:obj:`~.cudaGraphMemcpyNodeSetParams`, :py:obj:`~.cudaGraphAddMemcpyNode`, :py:obj:`~.cudaGraphMemcpyNodeGetParams` """ cdef cyruntime.cudaGraphNode_t cynode if node is None: pnode = 0 elif isinstance(node, (cudaGraphNode_t,driver.CUgraphNode)): pnode = int(node) else: pnode = int(cudaGraphNode_t(node)) cynode = pnode cydst = _HelperInputVoidPtr(dst) cdef void* cydst_ptr = cydst.cptr cysrc = _HelperInputVoidPtr(src) cdef void* cysrc_ptr = cysrc.cptr cdef cyruntime.cudaMemcpyKind cykind = kind.value with nogil: err = cyruntime.cudaGraphMemcpyNodeSetParams1D(cynode, cydst_ptr, cysrc_ptr, count, cykind) return (_dict_cudaError_t[err],) @cython.embedsignature(True) def cudaGraphAddMemsetNode(graph, pDependencies : Optional[tuple[cudaGraphNode_t] | list[cudaGraphNode_t]], size_t numDependencies, pMemsetParams : Optional[cudaMemsetParams]): """ Creates a memset node and adds it to a graph. Creates a new memset node and adds it to `graph` with `numDependencies` dependencies specified via `pDependencies`. It is possible for `numDependencies` to be 0, in which case the node will be placed at the root of the graph. `pDependencies` may not have any duplicate entries. A handle to the new node will be returned in `pGraphNode`. The element size must be 1, 2, or 4 bytes. When the graph is launched, the node will perform the memset described by `pMemsetParams`. Parameters ---------- graph : :py:obj:`~.CUgraph` or :py:obj:`~.cudaGraph_t` Graph to which to add the node pDependencies : list[:py:obj:`~.cudaGraphNode_t`] Dependencies of the node numDependencies : size_t Number of dependencies pMemsetParams : :py:obj:`~.cudaMemsetParams` Parameters for the memory set Returns ------- cudaError_t :py:obj:`~.cudaSuccess`, :py:obj:`~.cudaErrorInvalidValue`, :py:obj:`~.cudaErrorInvalidDevice` pGraphNode : :py:obj:`~.cudaGraphNode_t` Returns newly created node See Also -------- :py:obj:`~.cudaGraphAddNode`, :py:obj:`~.cudaMemset2D`, :py:obj:`~.cudaGraphMemsetNodeGetParams`, :py:obj:`~.cudaGraphMemsetNodeSetParams`, :py:obj:`~.cudaGraphCreate`, :py:obj:`~.cudaGraphDestroyNode`, :py:obj:`~.cudaGraphAddChildGraphNode`, :py:obj:`~.cudaGraphAddEmptyNode`, :py:obj:`~.cudaGraphAddKernelNode`, :py:obj:`~.cudaGraphAddHostNode`, :py:obj:`~.cudaGraphAddMemcpyNode` """ pDependencies = [] if pDependencies is None else pDependencies if not all(isinstance(_x, (cudaGraphNode_t,driver.CUgraphNode)) for _x in pDependencies): raise TypeError("Argument 'pDependencies' is not instance of type (expected tuple[cyruntime.cudaGraphNode_t,driver.CUgraphNode] or list[cyruntime.cudaGraphNode_t,driver.CUgraphNode]") cdef cyruntime.cudaGraph_t cygraph if graph is None: pgraph = 0 elif isinstance(graph, (cudaGraph_t,driver.CUgraph)): pgraph = int(graph) else: pgraph = int(cudaGraph_t(graph)) cygraph = pgraph cdef cudaGraphNode_t pGraphNode = cudaGraphNode_t() cdef cyruntime.cudaGraphNode_t* cypDependencies = NULL if len(pDependencies) > 1: cypDependencies = calloc(len(pDependencies), sizeof(cyruntime.cudaGraphNode_t)) if cypDependencies is NULL: raise MemoryError('Failed to allocate length x size memory: ' + str(len(pDependencies)) + 'x' + str(sizeof(cyruntime.cudaGraphNode_t))) else: for idx in range(len(pDependencies)): cypDependencies[idx] = (pDependencies[idx])._pvt_ptr[0] elif len(pDependencies) == 1: cypDependencies = (pDependencies[0])._pvt_ptr if numDependencies > len(pDependencies): raise RuntimeError("List is too small: " + str(len(pDependencies)) + " < " + str(numDependencies)) cdef cyruntime.cudaMemsetParams* cypMemsetParams_ptr = pMemsetParams._pvt_ptr if pMemsetParams is not None else NULL with nogil: err = cyruntime.cudaGraphAddMemsetNode(pGraphNode._pvt_ptr, cygraph, cypDependencies, numDependencies, cypMemsetParams_ptr) if len(pDependencies) > 1 and cypDependencies is not NULL: free(cypDependencies) if err != cyruntime.cudaSuccess: return (_dict_cudaError_t[err], None) return (_dict_cudaError_t[err], pGraphNode) @cython.embedsignature(True) def cudaGraphMemsetNodeGetParams(node): """ Returns a memset node's parameters. Returns the parameters of memset node `node` in `pNodeParams`. Parameters ---------- node : :py:obj:`~.CUgraphNode` or :py:obj:`~.cudaGraphNode_t` Node to get the parameters for Returns ------- cudaError_t :py:obj:`~.cudaSuccess`, :py:obj:`~.cudaErrorInvalidValue` pNodeParams : :py:obj:`~.cudaMemsetParams` Pointer to return the parameters See Also -------- :py:obj:`~.cudaMemset2D`, :py:obj:`~.cudaGraphAddMemsetNode`, :py:obj:`~.cudaGraphMemsetNodeSetParams` """ cdef cyruntime.cudaGraphNode_t cynode if node is None: pnode = 0 elif isinstance(node, (cudaGraphNode_t,driver.CUgraphNode)): pnode = int(node) else: pnode = int(cudaGraphNode_t(node)) cynode = pnode cdef cudaMemsetParams pNodeParams = cudaMemsetParams() with nogil: err = cyruntime.cudaGraphMemsetNodeGetParams(cynode, pNodeParams._pvt_ptr) if err != cyruntime.cudaSuccess: return (_dict_cudaError_t[err], None) return (_dict_cudaError_t[err], pNodeParams) @cython.embedsignature(True) def cudaGraphMemsetNodeSetParams(node, pNodeParams : Optional[cudaMemsetParams]): """ Sets a memset node's parameters. Sets the parameters of memset node `node` to `pNodeParams`. Parameters ---------- node : :py:obj:`~.CUgraphNode` or :py:obj:`~.cudaGraphNode_t` Node to set the parameters for pNodeParams : :py:obj:`~.cudaMemsetParams` Parameters to copy Returns ------- cudaError_t :py:obj:`~.cudaSuccess`, :py:obj:`~.cudaErrorInvalidValue` See Also -------- :py:obj:`~.cudaGraphNodeSetParams`, :py:obj:`~.cudaMemset2D`, :py:obj:`~.cudaGraphAddMemsetNode`, :py:obj:`~.cudaGraphMemsetNodeGetParams` """ cdef cyruntime.cudaGraphNode_t cynode if node is None: pnode = 0 elif isinstance(node, (cudaGraphNode_t,driver.CUgraphNode)): pnode = int(node) else: pnode = int(cudaGraphNode_t(node)) cynode = pnode cdef cyruntime.cudaMemsetParams* cypNodeParams_ptr = pNodeParams._pvt_ptr if pNodeParams is not None else NULL with nogil: err = cyruntime.cudaGraphMemsetNodeSetParams(cynode, cypNodeParams_ptr) return (_dict_cudaError_t[err],) @cython.embedsignature(True) def cudaGraphAddHostNode(graph, pDependencies : Optional[tuple[cudaGraphNode_t] | list[cudaGraphNode_t]], size_t numDependencies, pNodeParams : Optional[cudaHostNodeParams]): """ Creates a host execution node and adds it to a graph. Creates a new CPU execution node and adds it to `graph` with `numDependencies` dependencies specified via `pDependencies` and arguments specified in `pNodeParams`. It is possible for `numDependencies` to be 0, in which case the node will be placed at the root of the graph. `pDependencies` may not have any duplicate entries. A handle to the new node will be returned in `pGraphNode`. When the graph is launched, the node will invoke the specified CPU function. Host nodes are not supported under MPS with pre-Volta GPUs. Parameters ---------- graph : :py:obj:`~.CUgraph` or :py:obj:`~.cudaGraph_t` Graph to which to add the node pDependencies : list[:py:obj:`~.cudaGraphNode_t`] Dependencies of the node numDependencies : size_t Number of dependencies pNodeParams : :py:obj:`~.cudaHostNodeParams` Parameters for the host node Returns ------- cudaError_t :py:obj:`~.cudaSuccess`, :py:obj:`~.cudaErrorNotSupported`, :py:obj:`~.cudaErrorInvalidValue` pGraphNode : :py:obj:`~.cudaGraphNode_t` Returns newly created node See Also -------- :py:obj:`~.cudaGraphAddNode`, :py:obj:`~.cudaLaunchHostFunc`, :py:obj:`~.cudaGraphHostNodeGetParams`, :py:obj:`~.cudaGraphHostNodeSetParams`, :py:obj:`~.cudaGraphCreate`, :py:obj:`~.cudaGraphDestroyNode`, :py:obj:`~.cudaGraphAddChildGraphNode`, :py:obj:`~.cudaGraphAddEmptyNode`, :py:obj:`~.cudaGraphAddKernelNode`, :py:obj:`~.cudaGraphAddMemcpyNode`, :py:obj:`~.cudaGraphAddMemsetNode` """ pDependencies = [] if pDependencies is None else pDependencies if not all(isinstance(_x, (cudaGraphNode_t,driver.CUgraphNode)) for _x in pDependencies): raise TypeError("Argument 'pDependencies' is not instance of type (expected tuple[cyruntime.cudaGraphNode_t,driver.CUgraphNode] or list[cyruntime.cudaGraphNode_t,driver.CUgraphNode]") cdef cyruntime.cudaGraph_t cygraph if graph is None: pgraph = 0 elif isinstance(graph, (cudaGraph_t,driver.CUgraph)): pgraph = int(graph) else: pgraph = int(cudaGraph_t(graph)) cygraph = pgraph cdef cudaGraphNode_t pGraphNode = cudaGraphNode_t() cdef cyruntime.cudaGraphNode_t* cypDependencies = NULL if len(pDependencies) > 1: cypDependencies = calloc(len(pDependencies), sizeof(cyruntime.cudaGraphNode_t)) if cypDependencies is NULL: raise MemoryError('Failed to allocate length x size memory: ' + str(len(pDependencies)) + 'x' + str(sizeof(cyruntime.cudaGraphNode_t))) else: for idx in range(len(pDependencies)): cypDependencies[idx] = (pDependencies[idx])._pvt_ptr[0] elif len(pDependencies) == 1: cypDependencies = (pDependencies[0])._pvt_ptr if numDependencies > len(pDependencies): raise RuntimeError("List is too small: " + str(len(pDependencies)) + " < " + str(numDependencies)) cdef cyruntime.cudaHostNodeParams* cypNodeParams_ptr = pNodeParams._pvt_ptr if pNodeParams is not None else NULL with nogil: err = cyruntime.cudaGraphAddHostNode(pGraphNode._pvt_ptr, cygraph, cypDependencies, numDependencies, cypNodeParams_ptr) if len(pDependencies) > 1 and cypDependencies is not NULL: free(cypDependencies) if err != cyruntime.cudaSuccess: return (_dict_cudaError_t[err], None) return (_dict_cudaError_t[err], pGraphNode) @cython.embedsignature(True) def cudaGraphHostNodeGetParams(node): """ Returns a host node's parameters. Returns the parameters of host node `node` in `pNodeParams`. Parameters ---------- node : :py:obj:`~.CUgraphNode` or :py:obj:`~.cudaGraphNode_t` Node to get the parameters for Returns ------- cudaError_t :py:obj:`~.cudaSuccess`, :py:obj:`~.cudaErrorInvalidValue` pNodeParams : :py:obj:`~.cudaHostNodeParams` Pointer to return the parameters See Also -------- :py:obj:`~.cudaLaunchHostFunc`, :py:obj:`~.cudaGraphAddHostNode`, :py:obj:`~.cudaGraphHostNodeSetParams` """ cdef cyruntime.cudaGraphNode_t cynode if node is None: pnode = 0 elif isinstance(node, (cudaGraphNode_t,driver.CUgraphNode)): pnode = int(node) else: pnode = int(cudaGraphNode_t(node)) cynode = pnode cdef cudaHostNodeParams pNodeParams = cudaHostNodeParams() with nogil: err = cyruntime.cudaGraphHostNodeGetParams(cynode, pNodeParams._pvt_ptr) if err != cyruntime.cudaSuccess: return (_dict_cudaError_t[err], None) return (_dict_cudaError_t[err], pNodeParams) @cython.embedsignature(True) def cudaGraphHostNodeSetParams(node, pNodeParams : Optional[cudaHostNodeParams]): """ Sets a host node's parameters. Sets the parameters of host node `node` to `nodeParams`. Parameters ---------- node : :py:obj:`~.CUgraphNode` or :py:obj:`~.cudaGraphNode_t` Node to set the parameters for pNodeParams : :py:obj:`~.cudaHostNodeParams` Parameters to copy Returns ------- cudaError_t :py:obj:`~.cudaSuccess`, :py:obj:`~.cudaErrorInvalidValue` See Also -------- :py:obj:`~.cudaGraphNodeSetParams`, :py:obj:`~.cudaLaunchHostFunc`, :py:obj:`~.cudaGraphAddHostNode`, :py:obj:`~.cudaGraphHostNodeGetParams` """ cdef cyruntime.cudaGraphNode_t cynode if node is None: pnode = 0 elif isinstance(node, (cudaGraphNode_t,driver.CUgraphNode)): pnode = int(node) else: pnode = int(cudaGraphNode_t(node)) cynode = pnode cdef cyruntime.cudaHostNodeParams* cypNodeParams_ptr = pNodeParams._pvt_ptr if pNodeParams is not None else NULL with nogil: err = cyruntime.cudaGraphHostNodeSetParams(cynode, cypNodeParams_ptr) return (_dict_cudaError_t[err],) @cython.embedsignature(True) def cudaGraphAddChildGraphNode(graph, pDependencies : Optional[tuple[cudaGraphNode_t] | list[cudaGraphNode_t]], size_t numDependencies, childGraph): """ Creates a child graph node and adds it to a graph. Creates a new node which executes an embedded graph, and adds it to `graph` with `numDependencies` dependencies specified via `pDependencies`. It is possible for `numDependencies` to be 0, in which case the node will be placed at the root of the graph. `pDependencies` may not have any duplicate entries. A handle to the new node will be returned in `pGraphNode`. If `childGraph` contains allocation nodes, free nodes, or conditional nodes, this call will return an error. The node executes an embedded child graph. The child graph is cloned in this call. Parameters ---------- graph : :py:obj:`~.CUgraph` or :py:obj:`~.cudaGraph_t` Graph to which to add the node pDependencies : list[:py:obj:`~.cudaGraphNode_t`] Dependencies of the node numDependencies : size_t Number of dependencies childGraph : :py:obj:`~.CUgraph` or :py:obj:`~.cudaGraph_t` The graph to clone into this node Returns ------- cudaError_t :py:obj:`~.cudaSuccess`, :py:obj:`~.cudaErrorInvalidValue` pGraphNode : :py:obj:`~.cudaGraphNode_t` Returns newly created node See Also -------- :py:obj:`~.cudaGraphAddNode`, :py:obj:`~.cudaGraphChildGraphNodeGetGraph`, :py:obj:`~.cudaGraphCreate`, :py:obj:`~.cudaGraphDestroyNode`, :py:obj:`~.cudaGraphAddEmptyNode`, :py:obj:`~.cudaGraphAddKernelNode`, :py:obj:`~.cudaGraphAddHostNode`, :py:obj:`~.cudaGraphAddMemcpyNode`, :py:obj:`~.cudaGraphAddMemsetNode`, :py:obj:`~.cudaGraphClone` """ cdef cyruntime.cudaGraph_t cychildGraph if childGraph is None: pchildGraph = 0 elif isinstance(childGraph, (cudaGraph_t,driver.CUgraph)): pchildGraph = int(childGraph) else: pchildGraph = int(cudaGraph_t(childGraph)) cychildGraph = pchildGraph pDependencies = [] if pDependencies is None else pDependencies if not all(isinstance(_x, (cudaGraphNode_t,driver.CUgraphNode)) for _x in pDependencies): raise TypeError("Argument 'pDependencies' is not instance of type (expected tuple[cyruntime.cudaGraphNode_t,driver.CUgraphNode] or list[cyruntime.cudaGraphNode_t,driver.CUgraphNode]") cdef cyruntime.cudaGraph_t cygraph if graph is None: pgraph = 0 elif isinstance(graph, (cudaGraph_t,driver.CUgraph)): pgraph = int(graph) else: pgraph = int(cudaGraph_t(graph)) cygraph = pgraph cdef cudaGraphNode_t pGraphNode = cudaGraphNode_t() cdef cyruntime.cudaGraphNode_t* cypDependencies = NULL if len(pDependencies) > 1: cypDependencies = calloc(len(pDependencies), sizeof(cyruntime.cudaGraphNode_t)) if cypDependencies is NULL: raise MemoryError('Failed to allocate length x size memory: ' + str(len(pDependencies)) + 'x' + str(sizeof(cyruntime.cudaGraphNode_t))) else: for idx in range(len(pDependencies)): cypDependencies[idx] = (pDependencies[idx])._pvt_ptr[0] elif len(pDependencies) == 1: cypDependencies = (pDependencies[0])._pvt_ptr if numDependencies > len(pDependencies): raise RuntimeError("List is too small: " + str(len(pDependencies)) + " < " + str(numDependencies)) with nogil: err = cyruntime.cudaGraphAddChildGraphNode(pGraphNode._pvt_ptr, cygraph, cypDependencies, numDependencies, cychildGraph) if len(pDependencies) > 1 and cypDependencies is not NULL: free(cypDependencies) if err != cyruntime.cudaSuccess: return (_dict_cudaError_t[err], None) return (_dict_cudaError_t[err], pGraphNode) @cython.embedsignature(True) def cudaGraphChildGraphNodeGetGraph(node): """ Gets a handle to the embedded graph of a child graph node. Gets a handle to the embedded graph in a child graph node. This call does not clone the graph. Changes to the graph will be reflected in the node, and the node retains ownership of the graph. Allocation and free nodes cannot be added to the returned graph. Attempting to do so will return an error. Parameters ---------- node : :py:obj:`~.CUgraphNode` or :py:obj:`~.cudaGraphNode_t` Node to get the embedded graph for Returns ------- cudaError_t :py:obj:`~.cudaSuccess`, :py:obj:`~.cudaErrorInvalidValue` pGraph : :py:obj:`~.cudaGraph_t` Location to store a handle to the graph See Also -------- :py:obj:`~.cudaGraphAddChildGraphNode`, :py:obj:`~.cudaGraphNodeFindInClone` """ cdef cyruntime.cudaGraphNode_t cynode if node is None: pnode = 0 elif isinstance(node, (cudaGraphNode_t,driver.CUgraphNode)): pnode = int(node) else: pnode = int(cudaGraphNode_t(node)) cynode = pnode cdef cudaGraph_t pGraph = cudaGraph_t() with nogil: err = cyruntime.cudaGraphChildGraphNodeGetGraph(cynode, pGraph._pvt_ptr) if err != cyruntime.cudaSuccess: return (_dict_cudaError_t[err], None) return (_dict_cudaError_t[err], pGraph) @cython.embedsignature(True) def cudaGraphAddEmptyNode(graph, pDependencies : Optional[tuple[cudaGraphNode_t] | list[cudaGraphNode_t]], size_t numDependencies): """ Creates an empty node and adds it to a graph. Creates a new node which performs no operation, and adds it to `graph` with `numDependencies` dependencies specified via `pDependencies`. It is possible for `numDependencies` to be 0, in which case the node will be placed at the root of the graph. `pDependencies` may not have any duplicate entries. A handle to the new node will be returned in `pGraphNode`. An empty node performs no operation during execution, but can be used for transitive ordering. For example, a phased execution graph with 2 groups of n nodes with a barrier between them can be represented using an empty node and 2*n dependency edges, rather than no empty node and n^2 dependency edges. Parameters ---------- graph : :py:obj:`~.CUgraph` or :py:obj:`~.cudaGraph_t` Graph to which to add the node pDependencies : list[:py:obj:`~.cudaGraphNode_t`] Dependencies of the node numDependencies : size_t Number of dependencies Returns ------- cudaError_t :py:obj:`~.cudaSuccess`, :py:obj:`~.cudaErrorInvalidValue` pGraphNode : :py:obj:`~.cudaGraphNode_t` Returns newly created node See Also -------- :py:obj:`~.cudaGraphAddNode`, :py:obj:`~.cudaGraphCreate`, :py:obj:`~.cudaGraphDestroyNode`, :py:obj:`~.cudaGraphAddChildGraphNode`, :py:obj:`~.cudaGraphAddKernelNode`, :py:obj:`~.cudaGraphAddHostNode`, :py:obj:`~.cudaGraphAddMemcpyNode`, :py:obj:`~.cudaGraphAddMemsetNode` """ pDependencies = [] if pDependencies is None else pDependencies if not all(isinstance(_x, (cudaGraphNode_t,driver.CUgraphNode)) for _x in pDependencies): raise TypeError("Argument 'pDependencies' is not instance of type (expected tuple[cyruntime.cudaGraphNode_t,driver.CUgraphNode] or list[cyruntime.cudaGraphNode_t,driver.CUgraphNode]") cdef cyruntime.cudaGraph_t cygraph if graph is None: pgraph = 0 elif isinstance(graph, (cudaGraph_t,driver.CUgraph)): pgraph = int(graph) else: pgraph = int(cudaGraph_t(graph)) cygraph = pgraph cdef cudaGraphNode_t pGraphNode = cudaGraphNode_t() cdef cyruntime.cudaGraphNode_t* cypDependencies = NULL if len(pDependencies) > 1: cypDependencies = calloc(len(pDependencies), sizeof(cyruntime.cudaGraphNode_t)) if cypDependencies is NULL: raise MemoryError('Failed to allocate length x size memory: ' + str(len(pDependencies)) + 'x' + str(sizeof(cyruntime.cudaGraphNode_t))) else: for idx in range(len(pDependencies)): cypDependencies[idx] = (pDependencies[idx])._pvt_ptr[0] elif len(pDependencies) == 1: cypDependencies = (pDependencies[0])._pvt_ptr if numDependencies > len(pDependencies): raise RuntimeError("List is too small: " + str(len(pDependencies)) + " < " + str(numDependencies)) with nogil: err = cyruntime.cudaGraphAddEmptyNode(pGraphNode._pvt_ptr, cygraph, cypDependencies, numDependencies) if len(pDependencies) > 1 and cypDependencies is not NULL: free(cypDependencies) if err != cyruntime.cudaSuccess: return (_dict_cudaError_t[err], None) return (_dict_cudaError_t[err], pGraphNode) @cython.embedsignature(True) def cudaGraphAddEventRecordNode(graph, pDependencies : Optional[tuple[cudaGraphNode_t] | list[cudaGraphNode_t]], size_t numDependencies, event): """ Creates an event record node and adds it to a graph. Creates a new event record node and adds it to `hGraph` with `numDependencies` dependencies specified via `dependencies` and event specified in `event`. It is possible for `numDependencies` to be 0, in which case the node will be placed at the root of the graph. `dependencies` may not have any duplicate entries. A handle to the new node will be returned in `phGraphNode`. Each launch of the graph will record `event` to capture execution of the node's dependencies. These nodes may not be used in loops or conditionals. Parameters ---------- hGraph : :py:obj:`~.CUgraph` or :py:obj:`~.cudaGraph_t` Graph to which to add the node dependencies : list[:py:obj:`~.cudaGraphNode_t`] Dependencies of the node numDependencies : size_t Number of dependencies event : :py:obj:`~.CUevent` or :py:obj:`~.cudaEvent_t` Event for the node Returns ------- cudaError_t :py:obj:`~.cudaSuccess`, :py:obj:`~.cudaErrorInvalidValue` phGraphNode : :py:obj:`~.cudaGraphNode_t` Returns newly created node See Also -------- :py:obj:`~.cudaGraphAddNode`, :py:obj:`~.cudaGraphAddEventWaitNode`, :py:obj:`~.cudaEventRecordWithFlags`, :py:obj:`~.cudaStreamWaitEvent`, :py:obj:`~.cudaGraphCreate`, :py:obj:`~.cudaGraphDestroyNode`, :py:obj:`~.cudaGraphAddChildGraphNode`, :py:obj:`~.cudaGraphAddEmptyNode`, :py:obj:`~.cudaGraphAddKernelNode`, :py:obj:`~.cudaGraphAddMemcpyNode`, :py:obj:`~.cudaGraphAddMemsetNode` """ cdef cyruntime.cudaEvent_t cyevent if event is None: pevent = 0 elif isinstance(event, (cudaEvent_t,driver.CUevent)): pevent = int(event) else: pevent = int(cudaEvent_t(event)) cyevent = pevent pDependencies = [] if pDependencies is None else pDependencies if not all(isinstance(_x, (cudaGraphNode_t,driver.CUgraphNode)) for _x in pDependencies): raise TypeError("Argument 'pDependencies' is not instance of type (expected tuple[cyruntime.cudaGraphNode_t,driver.CUgraphNode] or list[cyruntime.cudaGraphNode_t,driver.CUgraphNode]") cdef cyruntime.cudaGraph_t cygraph if graph is None: pgraph = 0 elif isinstance(graph, (cudaGraph_t,driver.CUgraph)): pgraph = int(graph) else: pgraph = int(cudaGraph_t(graph)) cygraph = pgraph cdef cudaGraphNode_t pGraphNode = cudaGraphNode_t() cdef cyruntime.cudaGraphNode_t* cypDependencies = NULL if len(pDependencies) > 1: cypDependencies = calloc(len(pDependencies), sizeof(cyruntime.cudaGraphNode_t)) if cypDependencies is NULL: raise MemoryError('Failed to allocate length x size memory: ' + str(len(pDependencies)) + 'x' + str(sizeof(cyruntime.cudaGraphNode_t))) else: for idx in range(len(pDependencies)): cypDependencies[idx] = (pDependencies[idx])._pvt_ptr[0] elif len(pDependencies) == 1: cypDependencies = (pDependencies[0])._pvt_ptr if numDependencies > len(pDependencies): raise RuntimeError("List is too small: " + str(len(pDependencies)) + " < " + str(numDependencies)) with nogil: err = cyruntime.cudaGraphAddEventRecordNode(pGraphNode._pvt_ptr, cygraph, cypDependencies, numDependencies, cyevent) if len(pDependencies) > 1 and cypDependencies is not NULL: free(cypDependencies) if err != cyruntime.cudaSuccess: return (_dict_cudaError_t[err], None) return (_dict_cudaError_t[err], pGraphNode) @cython.embedsignature(True) def cudaGraphEventRecordNodeGetEvent(node): """ Returns the event associated with an event record node. Returns the event of event record node `hNode` in `event_out`. Parameters ---------- hNode : :py:obj:`~.CUgraphNode` or :py:obj:`~.cudaGraphNode_t` Node to get the event for Returns ------- cudaError_t :py:obj:`~.cudaSuccess`, :py:obj:`~.cudaErrorInvalidValue` event_out : :py:obj:`~.cudaEvent_t` Pointer to return the event See Also -------- :py:obj:`~.cudaGraphAddEventRecordNode`, :py:obj:`~.cudaGraphEventRecordNodeSetEvent`, :py:obj:`~.cudaGraphEventWaitNodeGetEvent`, :py:obj:`~.cudaEventRecordWithFlags`, :py:obj:`~.cudaStreamWaitEvent` """ cdef cyruntime.cudaGraphNode_t cynode if node is None: pnode = 0 elif isinstance(node, (cudaGraphNode_t,driver.CUgraphNode)): pnode = int(node) else: pnode = int(cudaGraphNode_t(node)) cynode = pnode cdef cudaEvent_t event_out = cudaEvent_t() with nogil: err = cyruntime.cudaGraphEventRecordNodeGetEvent(cynode, event_out._pvt_ptr) if err != cyruntime.cudaSuccess: return (_dict_cudaError_t[err], None) return (_dict_cudaError_t[err], event_out) @cython.embedsignature(True) def cudaGraphEventRecordNodeSetEvent(node, event): """ Sets an event record node's event. Sets the event of event record node `hNode` to `event`. Parameters ---------- hNode : :py:obj:`~.CUgraphNode` or :py:obj:`~.cudaGraphNode_t` Node to set the event for event : :py:obj:`~.CUevent` or :py:obj:`~.cudaEvent_t` Event to use Returns ------- cudaError_t :py:obj:`~.cudaSuccess`, :py:obj:`~.cudaErrorInvalidValue` See Also -------- :py:obj:`~.cudaGraphNodeSetParams`, :py:obj:`~.cudaGraphAddEventRecordNode`, :py:obj:`~.cudaGraphEventRecordNodeGetEvent`, :py:obj:`~.cudaGraphEventWaitNodeSetEvent`, :py:obj:`~.cudaEventRecordWithFlags`, :py:obj:`~.cudaStreamWaitEvent` """ cdef cyruntime.cudaEvent_t cyevent if event is None: pevent = 0 elif isinstance(event, (cudaEvent_t,driver.CUevent)): pevent = int(event) else: pevent = int(cudaEvent_t(event)) cyevent = pevent cdef cyruntime.cudaGraphNode_t cynode if node is None: pnode = 0 elif isinstance(node, (cudaGraphNode_t,driver.CUgraphNode)): pnode = int(node) else: pnode = int(cudaGraphNode_t(node)) cynode = pnode with nogil: err = cyruntime.cudaGraphEventRecordNodeSetEvent(cynode, cyevent) return (_dict_cudaError_t[err],) @cython.embedsignature(True) def cudaGraphAddEventWaitNode(graph, pDependencies : Optional[tuple[cudaGraphNode_t] | list[cudaGraphNode_t]], size_t numDependencies, event): """ Creates an event wait node and adds it to a graph. Creates a new event wait node and adds it to `hGraph` with `numDependencies` dependencies specified via `dependencies` and event specified in `event`. It is possible for `numDependencies` to be 0, in which case the node will be placed at the root of the graph. `dependencies` may not have any duplicate entries. A handle to the new node will be returned in `phGraphNode`. The graph node will wait for all work captured in `event`. See :py:obj:`~.cuEventRecord()` for details on what is captured by an event. The synchronization will be performed efficiently on the device when applicable. `event` may be from a different context or device than the launch stream. These nodes may not be used in loops or conditionals. Parameters ---------- hGraph : :py:obj:`~.CUgraph` or :py:obj:`~.cudaGraph_t` Graph to which to add the node dependencies : list[:py:obj:`~.cudaGraphNode_t`] Dependencies of the node numDependencies : size_t Number of dependencies event : :py:obj:`~.CUevent` or :py:obj:`~.cudaEvent_t` Event for the node Returns ------- cudaError_t :py:obj:`~.cudaSuccess`, :py:obj:`~.cudaErrorInvalidValue` phGraphNode : :py:obj:`~.cudaGraphNode_t` Returns newly created node See Also -------- :py:obj:`~.cudaGraphAddNode`, :py:obj:`~.cudaGraphAddEventRecordNode`, :py:obj:`~.cudaEventRecordWithFlags`, :py:obj:`~.cudaStreamWaitEvent`, :py:obj:`~.cudaGraphCreate`, :py:obj:`~.cudaGraphDestroyNode`, :py:obj:`~.cudaGraphAddChildGraphNode`, :py:obj:`~.cudaGraphAddEmptyNode`, :py:obj:`~.cudaGraphAddKernelNode`, :py:obj:`~.cudaGraphAddMemcpyNode`, :py:obj:`~.cudaGraphAddMemsetNode` """ cdef cyruntime.cudaEvent_t cyevent if event is None: pevent = 0 elif isinstance(event, (cudaEvent_t,driver.CUevent)): pevent = int(event) else: pevent = int(cudaEvent_t(event)) cyevent = pevent pDependencies = [] if pDependencies is None else pDependencies if not all(isinstance(_x, (cudaGraphNode_t,driver.CUgraphNode)) for _x in pDependencies): raise TypeError("Argument 'pDependencies' is not instance of type (expected tuple[cyruntime.cudaGraphNode_t,driver.CUgraphNode] or list[cyruntime.cudaGraphNode_t,driver.CUgraphNode]") cdef cyruntime.cudaGraph_t cygraph if graph is None: pgraph = 0 elif isinstance(graph, (cudaGraph_t,driver.CUgraph)): pgraph = int(graph) else: pgraph = int(cudaGraph_t(graph)) cygraph = pgraph cdef cudaGraphNode_t pGraphNode = cudaGraphNode_t() cdef cyruntime.cudaGraphNode_t* cypDependencies = NULL if len(pDependencies) > 1: cypDependencies = calloc(len(pDependencies), sizeof(cyruntime.cudaGraphNode_t)) if cypDependencies is NULL: raise MemoryError('Failed to allocate length x size memory: ' + str(len(pDependencies)) + 'x' + str(sizeof(cyruntime.cudaGraphNode_t))) else: for idx in range(len(pDependencies)): cypDependencies[idx] = (pDependencies[idx])._pvt_ptr[0] elif len(pDependencies) == 1: cypDependencies = (pDependencies[0])._pvt_ptr if numDependencies > len(pDependencies): raise RuntimeError("List is too small: " + str(len(pDependencies)) + " < " + str(numDependencies)) with nogil: err = cyruntime.cudaGraphAddEventWaitNode(pGraphNode._pvt_ptr, cygraph, cypDependencies, numDependencies, cyevent) if len(pDependencies) > 1 and cypDependencies is not NULL: free(cypDependencies) if err != cyruntime.cudaSuccess: return (_dict_cudaError_t[err], None) return (_dict_cudaError_t[err], pGraphNode) @cython.embedsignature(True) def cudaGraphEventWaitNodeGetEvent(node): """ Returns the event associated with an event wait node. Returns the event of event wait node `hNode` in `event_out`. Parameters ---------- hNode : :py:obj:`~.CUgraphNode` or :py:obj:`~.cudaGraphNode_t` Node to get the event for Returns ------- cudaError_t :py:obj:`~.cudaSuccess`, :py:obj:`~.cudaErrorInvalidValue` event_out : :py:obj:`~.cudaEvent_t` Pointer to return the event See Also -------- :py:obj:`~.cudaGraphAddEventWaitNode`, :py:obj:`~.cudaGraphEventWaitNodeSetEvent`, :py:obj:`~.cudaGraphEventRecordNodeGetEvent`, :py:obj:`~.cudaEventRecordWithFlags`, :py:obj:`~.cudaStreamWaitEvent` """ cdef cyruntime.cudaGraphNode_t cynode if node is None: pnode = 0 elif isinstance(node, (cudaGraphNode_t,driver.CUgraphNode)): pnode = int(node) else: pnode = int(cudaGraphNode_t(node)) cynode = pnode cdef cudaEvent_t event_out = cudaEvent_t() with nogil: err = cyruntime.cudaGraphEventWaitNodeGetEvent(cynode, event_out._pvt_ptr) if err != cyruntime.cudaSuccess: return (_dict_cudaError_t[err], None) return (_dict_cudaError_t[err], event_out) @cython.embedsignature(True) def cudaGraphEventWaitNodeSetEvent(node, event): """ Sets an event wait node's event. Sets the event of event wait node `hNode` to `event`. Parameters ---------- hNode : :py:obj:`~.CUgraphNode` or :py:obj:`~.cudaGraphNode_t` Node to set the event for event : :py:obj:`~.CUevent` or :py:obj:`~.cudaEvent_t` Event to use Returns ------- cudaError_t :py:obj:`~.cudaSuccess`, :py:obj:`~.cudaErrorInvalidValue` See Also -------- :py:obj:`~.cudaGraphNodeSetParams`, :py:obj:`~.cudaGraphAddEventWaitNode`, :py:obj:`~.cudaGraphEventWaitNodeGetEvent`, :py:obj:`~.cudaGraphEventRecordNodeSetEvent`, :py:obj:`~.cudaEventRecordWithFlags`, :py:obj:`~.cudaStreamWaitEvent` """ cdef cyruntime.cudaEvent_t cyevent if event is None: pevent = 0 elif isinstance(event, (cudaEvent_t,driver.CUevent)): pevent = int(event) else: pevent = int(cudaEvent_t(event)) cyevent = pevent cdef cyruntime.cudaGraphNode_t cynode if node is None: pnode = 0 elif isinstance(node, (cudaGraphNode_t,driver.CUgraphNode)): pnode = int(node) else: pnode = int(cudaGraphNode_t(node)) cynode = pnode with nogil: err = cyruntime.cudaGraphEventWaitNodeSetEvent(cynode, cyevent) return (_dict_cudaError_t[err],) @cython.embedsignature(True) def cudaGraphAddExternalSemaphoresSignalNode(graph, pDependencies : Optional[tuple[cudaGraphNode_t] | list[cudaGraphNode_t]], size_t numDependencies, nodeParams : Optional[cudaExternalSemaphoreSignalNodeParams]): """ Creates an external semaphore signal node and adds it to a graph. Creates a new external semaphore signal node and adds it to `graph` with `numDependencies` dependencies specified via `dependencies` and arguments specified in `nodeParams`. It is possible for `numDependencies` to be 0, in which case the node will be placed at the root of the graph. `dependencies` may not have any duplicate entries. A handle to the new node will be returned in `pGraphNode`. Performs a signal operation on a set of externally allocated semaphore objects when the node is launched. The operation(s) will occur after all of the node's dependencies have completed. Parameters ---------- graph : :py:obj:`~.CUgraph` or :py:obj:`~.cudaGraph_t` Graph to which to add the node pDependencies : list[:py:obj:`~.cudaGraphNode_t`] Dependencies of the node numDependencies : size_t Number of dependencies nodeParams : :py:obj:`~.cudaExternalSemaphoreSignalNodeParams` Parameters for the node Returns ------- cudaError_t :py:obj:`~.cudaSuccess`, :py:obj:`~.cudaErrorInvalidValue` pGraphNode : :py:obj:`~.cudaGraphNode_t` Returns newly created node See Also -------- :py:obj:`~.cudaGraphAddNode`, :py:obj:`~.cudaGraphExternalSemaphoresSignalNodeGetParams`, :py:obj:`~.cudaGraphExternalSemaphoresSignalNodeSetParams`, :py:obj:`~.cudaGraphExecExternalSemaphoresSignalNodeSetParams`, :py:obj:`~.cudaGraphAddExternalSemaphoresWaitNode`, :py:obj:`~.cudaImportExternalSemaphore`, :py:obj:`~.cudaSignalExternalSemaphoresAsync`, :py:obj:`~.cudaWaitExternalSemaphoresAsync`, :py:obj:`~.cudaGraphCreate`, :py:obj:`~.cudaGraphDestroyNode`, :py:obj:`~.cudaGraphAddEventRecordNode`, :py:obj:`~.cudaGraphAddEventWaitNode`, :py:obj:`~.cudaGraphAddChildGraphNode`, :py:obj:`~.cudaGraphAddEmptyNode`, :py:obj:`~.cudaGraphAddKernelNode`, :py:obj:`~.cudaGraphAddMemcpyNode`, :py:obj:`~.cudaGraphAddMemsetNode` """ pDependencies = [] if pDependencies is None else pDependencies if not all(isinstance(_x, (cudaGraphNode_t,driver.CUgraphNode)) for _x in pDependencies): raise TypeError("Argument 'pDependencies' is not instance of type (expected tuple[cyruntime.cudaGraphNode_t,driver.CUgraphNode] or list[cyruntime.cudaGraphNode_t,driver.CUgraphNode]") cdef cyruntime.cudaGraph_t cygraph if graph is None: pgraph = 0 elif isinstance(graph, (cudaGraph_t,driver.CUgraph)): pgraph = int(graph) else: pgraph = int(cudaGraph_t(graph)) cygraph = pgraph cdef cudaGraphNode_t pGraphNode = cudaGraphNode_t() cdef cyruntime.cudaGraphNode_t* cypDependencies = NULL if len(pDependencies) > 1: cypDependencies = calloc(len(pDependencies), sizeof(cyruntime.cudaGraphNode_t)) if cypDependencies is NULL: raise MemoryError('Failed to allocate length x size memory: ' + str(len(pDependencies)) + 'x' + str(sizeof(cyruntime.cudaGraphNode_t))) else: for idx in range(len(pDependencies)): cypDependencies[idx] = (pDependencies[idx])._pvt_ptr[0] elif len(pDependencies) == 1: cypDependencies = (pDependencies[0])._pvt_ptr if numDependencies > len(pDependencies): raise RuntimeError("List is too small: " + str(len(pDependencies)) + " < " + str(numDependencies)) cdef cyruntime.cudaExternalSemaphoreSignalNodeParams* cynodeParams_ptr = nodeParams._pvt_ptr if nodeParams is not None else NULL with nogil: err = cyruntime.cudaGraphAddExternalSemaphoresSignalNode(pGraphNode._pvt_ptr, cygraph, cypDependencies, numDependencies, cynodeParams_ptr) if len(pDependencies) > 1 and cypDependencies is not NULL: free(cypDependencies) if err != cyruntime.cudaSuccess: return (_dict_cudaError_t[err], None) return (_dict_cudaError_t[err], pGraphNode) @cython.embedsignature(True) def cudaGraphExternalSemaphoresSignalNodeGetParams(hNode): """ Returns an external semaphore signal node's parameters. Returns the parameters of an external semaphore signal node `hNode` in `params_out`. The `extSemArray` and `paramsArray` returned in `params_out`, are owned by the node. This memory remains valid until the node is destroyed or its parameters are modified, and should not be modified directly. Use :py:obj:`~.cudaGraphExternalSemaphoresSignalNodeSetParams` to update the parameters of this node. Parameters ---------- hNode : :py:obj:`~.CUgraphNode` or :py:obj:`~.cudaGraphNode_t` Node to get the parameters for Returns ------- cudaError_t :py:obj:`~.cudaSuccess`, :py:obj:`~.cudaErrorInvalidValue` params_out : :py:obj:`~.cudaExternalSemaphoreSignalNodeParams` Pointer to return the parameters See Also -------- :py:obj:`~.cudaLaunchKernel`, :py:obj:`~.cudaGraphAddExternalSemaphoresSignalNode`, :py:obj:`~.cudaGraphExternalSemaphoresSignalNodeSetParams`, :py:obj:`~.cudaGraphAddExternalSemaphoresWaitNode`, :py:obj:`~.cudaSignalExternalSemaphoresAsync`, :py:obj:`~.cudaWaitExternalSemaphoresAsync` """ cdef cyruntime.cudaGraphNode_t cyhNode if hNode is None: phNode = 0 elif isinstance(hNode, (cudaGraphNode_t,driver.CUgraphNode)): phNode = int(hNode) else: phNode = int(cudaGraphNode_t(hNode)) cyhNode = phNode cdef cudaExternalSemaphoreSignalNodeParams params_out = cudaExternalSemaphoreSignalNodeParams() with nogil: err = cyruntime.cudaGraphExternalSemaphoresSignalNodeGetParams(cyhNode, params_out._pvt_ptr) if err != cyruntime.cudaSuccess: return (_dict_cudaError_t[err], None) return (_dict_cudaError_t[err], params_out) @cython.embedsignature(True) def cudaGraphExternalSemaphoresSignalNodeSetParams(hNode, nodeParams : Optional[cudaExternalSemaphoreSignalNodeParams]): """ Sets an external semaphore signal node's parameters. Sets the parameters of an external semaphore signal node `hNode` to `nodeParams`. Parameters ---------- hNode : :py:obj:`~.CUgraphNode` or :py:obj:`~.cudaGraphNode_t` Node to set the parameters for nodeParams : :py:obj:`~.cudaExternalSemaphoreSignalNodeParams` Parameters to copy Returns ------- cudaError_t :py:obj:`~.cudaSuccess`, :py:obj:`~.cudaErrorInvalidValue` See Also -------- :py:obj:`~.cudaGraphNodeSetParams`, :py:obj:`~.cudaGraphAddExternalSemaphoresSignalNode`, :py:obj:`~.cudaGraphExternalSemaphoresSignalNodeSetParams`, :py:obj:`~.cudaGraphAddExternalSemaphoresWaitNode`, :py:obj:`~.cudaSignalExternalSemaphoresAsync`, :py:obj:`~.cudaWaitExternalSemaphoresAsync` """ cdef cyruntime.cudaGraphNode_t cyhNode if hNode is None: phNode = 0 elif isinstance(hNode, (cudaGraphNode_t,driver.CUgraphNode)): phNode = int(hNode) else: phNode = int(cudaGraphNode_t(hNode)) cyhNode = phNode cdef cyruntime.cudaExternalSemaphoreSignalNodeParams* cynodeParams_ptr = nodeParams._pvt_ptr if nodeParams is not None else NULL with nogil: err = cyruntime.cudaGraphExternalSemaphoresSignalNodeSetParams(cyhNode, cynodeParams_ptr) return (_dict_cudaError_t[err],) @cython.embedsignature(True) def cudaGraphAddExternalSemaphoresWaitNode(graph, pDependencies : Optional[tuple[cudaGraphNode_t] | list[cudaGraphNode_t]], size_t numDependencies, nodeParams : Optional[cudaExternalSemaphoreWaitNodeParams]): """ Creates an external semaphore wait node and adds it to a graph. Creates a new external semaphore wait node and adds it to `graph` with `numDependencies` dependencies specified via `dependencies` and arguments specified in `nodeParams`. It is possible for `numDependencies` to be 0, in which case the node will be placed at the root of the graph. `dependencies` may not have any duplicate entries. A handle to the new node will be returned in `pGraphNode`. Performs a wait operation on a set of externally allocated semaphore objects when the node is launched. The node's dependencies will not be launched until the wait operation has completed. Parameters ---------- graph : :py:obj:`~.CUgraph` or :py:obj:`~.cudaGraph_t` Graph to which to add the node pDependencies : list[:py:obj:`~.cudaGraphNode_t`] Dependencies of the node numDependencies : size_t Number of dependencies nodeParams : :py:obj:`~.cudaExternalSemaphoreWaitNodeParams` Parameters for the node Returns ------- cudaError_t :py:obj:`~.cudaSuccess`, :py:obj:`~.cudaErrorInvalidValue` pGraphNode : :py:obj:`~.cudaGraphNode_t` Returns newly created node See Also -------- :py:obj:`~.cudaGraphAddNode`, :py:obj:`~.cudaGraphExternalSemaphoresWaitNodeGetParams`, :py:obj:`~.cudaGraphExternalSemaphoresWaitNodeSetParams`, :py:obj:`~.cudaGraphExecExternalSemaphoresWaitNodeSetParams`, :py:obj:`~.cudaGraphAddExternalSemaphoresSignalNode`, :py:obj:`~.cudaImportExternalSemaphore`, :py:obj:`~.cudaSignalExternalSemaphoresAsync`, :py:obj:`~.cudaWaitExternalSemaphoresAsync`, :py:obj:`~.cudaGraphCreate`, :py:obj:`~.cudaGraphDestroyNode`, :py:obj:`~.cudaGraphAddEventRecordNode`, :py:obj:`~.cudaGraphAddEventWaitNode`, :py:obj:`~.cudaGraphAddChildGraphNode`, :py:obj:`~.cudaGraphAddEmptyNode`, :py:obj:`~.cudaGraphAddKernelNode`, :py:obj:`~.cudaGraphAddMemcpyNode`, :py:obj:`~.cudaGraphAddMemsetNode` """ pDependencies = [] if pDependencies is None else pDependencies if not all(isinstance(_x, (cudaGraphNode_t,driver.CUgraphNode)) for _x in pDependencies): raise TypeError("Argument 'pDependencies' is not instance of type (expected tuple[cyruntime.cudaGraphNode_t,driver.CUgraphNode] or list[cyruntime.cudaGraphNode_t,driver.CUgraphNode]") cdef cyruntime.cudaGraph_t cygraph if graph is None: pgraph = 0 elif isinstance(graph, (cudaGraph_t,driver.CUgraph)): pgraph = int(graph) else: pgraph = int(cudaGraph_t(graph)) cygraph = pgraph cdef cudaGraphNode_t pGraphNode = cudaGraphNode_t() cdef cyruntime.cudaGraphNode_t* cypDependencies = NULL if len(pDependencies) > 1: cypDependencies = calloc(len(pDependencies), sizeof(cyruntime.cudaGraphNode_t)) if cypDependencies is NULL: raise MemoryError('Failed to allocate length x size memory: ' + str(len(pDependencies)) + 'x' + str(sizeof(cyruntime.cudaGraphNode_t))) else: for idx in range(len(pDependencies)): cypDependencies[idx] = (pDependencies[idx])._pvt_ptr[0] elif len(pDependencies) == 1: cypDependencies = (pDependencies[0])._pvt_ptr if numDependencies > len(pDependencies): raise RuntimeError("List is too small: " + str(len(pDependencies)) + " < " + str(numDependencies)) cdef cyruntime.cudaExternalSemaphoreWaitNodeParams* cynodeParams_ptr = nodeParams._pvt_ptr if nodeParams is not None else NULL with nogil: err = cyruntime.cudaGraphAddExternalSemaphoresWaitNode(pGraphNode._pvt_ptr, cygraph, cypDependencies, numDependencies, cynodeParams_ptr) if len(pDependencies) > 1 and cypDependencies is not NULL: free(cypDependencies) if err != cyruntime.cudaSuccess: return (_dict_cudaError_t[err], None) return (_dict_cudaError_t[err], pGraphNode) @cython.embedsignature(True) def cudaGraphExternalSemaphoresWaitNodeGetParams(hNode): """ Returns an external semaphore wait node's parameters. Returns the parameters of an external semaphore wait node `hNode` in `params_out`. The `extSemArray` and `paramsArray` returned in `params_out`, are owned by the node. This memory remains valid until the node is destroyed or its parameters are modified, and should not be modified directly. Use :py:obj:`~.cudaGraphExternalSemaphoresSignalNodeSetParams` to update the parameters of this node. Parameters ---------- hNode : :py:obj:`~.CUgraphNode` or :py:obj:`~.cudaGraphNode_t` Node to get the parameters for Returns ------- cudaError_t :py:obj:`~.cudaSuccess`, :py:obj:`~.cudaErrorInvalidValue` params_out : :py:obj:`~.cudaExternalSemaphoreWaitNodeParams` Pointer to return the parameters See Also -------- :py:obj:`~.cudaLaunchKernel`, :py:obj:`~.cudaGraphAddExternalSemaphoresWaitNode`, :py:obj:`~.cudaGraphExternalSemaphoresWaitNodeSetParams`, :py:obj:`~.cudaGraphAddExternalSemaphoresWaitNode`, :py:obj:`~.cudaSignalExternalSemaphoresAsync`, :py:obj:`~.cudaWaitExternalSemaphoresAsync` """ cdef cyruntime.cudaGraphNode_t cyhNode if hNode is None: phNode = 0 elif isinstance(hNode, (cudaGraphNode_t,driver.CUgraphNode)): phNode = int(hNode) else: phNode = int(cudaGraphNode_t(hNode)) cyhNode = phNode cdef cudaExternalSemaphoreWaitNodeParams params_out = cudaExternalSemaphoreWaitNodeParams() with nogil: err = cyruntime.cudaGraphExternalSemaphoresWaitNodeGetParams(cyhNode, params_out._pvt_ptr) if err != cyruntime.cudaSuccess: return (_dict_cudaError_t[err], None) return (_dict_cudaError_t[err], params_out) @cython.embedsignature(True) def cudaGraphExternalSemaphoresWaitNodeSetParams(hNode, nodeParams : Optional[cudaExternalSemaphoreWaitNodeParams]): """ Sets an external semaphore wait node's parameters. Sets the parameters of an external semaphore wait node `hNode` to `nodeParams`. Parameters ---------- hNode : :py:obj:`~.CUgraphNode` or :py:obj:`~.cudaGraphNode_t` Node to set the parameters for nodeParams : :py:obj:`~.cudaExternalSemaphoreWaitNodeParams` Parameters to copy Returns ------- cudaError_t :py:obj:`~.cudaSuccess`, :py:obj:`~.cudaErrorInvalidValue` See Also -------- :py:obj:`~.cudaGraphNodeSetParams`, :py:obj:`~.cudaGraphAddExternalSemaphoresWaitNode`, :py:obj:`~.cudaGraphExternalSemaphoresWaitNodeSetParams`, :py:obj:`~.cudaGraphAddExternalSemaphoresWaitNode`, :py:obj:`~.cudaSignalExternalSemaphoresAsync`, :py:obj:`~.cudaWaitExternalSemaphoresAsync` """ cdef cyruntime.cudaGraphNode_t cyhNode if hNode is None: phNode = 0 elif isinstance(hNode, (cudaGraphNode_t,driver.CUgraphNode)): phNode = int(hNode) else: phNode = int(cudaGraphNode_t(hNode)) cyhNode = phNode cdef cyruntime.cudaExternalSemaphoreWaitNodeParams* cynodeParams_ptr = nodeParams._pvt_ptr if nodeParams is not None else NULL with nogil: err = cyruntime.cudaGraphExternalSemaphoresWaitNodeSetParams(cyhNode, cynodeParams_ptr) return (_dict_cudaError_t[err],) @cython.embedsignature(True) def cudaGraphAddMemAllocNode(graph, pDependencies : Optional[tuple[cudaGraphNode_t] | list[cudaGraphNode_t]], size_t numDependencies, nodeParams : Optional[cudaMemAllocNodeParams]): """ Creates an allocation node and adds it to a graph. Creates a new allocation node and adds it to `graph` with `numDependencies` dependencies specified via `pDependencies` and arguments specified in `nodeParams`. It is possible for `numDependencies` to be 0, in which case the node will be placed at the root of the graph. `pDependencies` may not have any duplicate entries. A handle to the new node will be returned in `pGraphNode`. When :py:obj:`~.cudaGraphAddMemAllocNode` creates an allocation node, it returns the address of the allocation in `nodeParams.dptr`. The allocation's address remains fixed across instantiations and launches. If the allocation is freed in the same graph, by creating a free node using :py:obj:`~.cudaGraphAddMemFreeNode`, the allocation can be accessed by nodes ordered after the allocation node but before the free node. These allocations cannot be freed outside the owning graph, and they can only be freed once in the owning graph. If the allocation is not freed in the same graph, then it can be accessed not only by nodes in the graph which are ordered after the allocation node, but also by stream operations ordered after the graph's execution but before the allocation is freed. Allocations which are not freed in the same graph can be freed by: - passing the allocation to :py:obj:`~.cudaMemFreeAsync` or :py:obj:`~.cudaMemFree`; - launching a graph with a free node for that allocation; or - specifying :py:obj:`~.cudaGraphInstantiateFlagAutoFreeOnLaunch` during instantiation, which makes each launch behave as though it called :py:obj:`~.cudaMemFreeAsync` for every unfreed allocation. It is not possible to free an allocation in both the owning graph and another graph. If the allocation is freed in the same graph, a free node cannot be added to another graph. If the allocation is freed in another graph, a free node can no longer be added to the owning graph. The following restrictions apply to graphs which contain allocation and/or memory free nodes: - Nodes and edges of the graph cannot be deleted. - The graph can only be used in a child node if the ownership is moved to the parent. - Only one instantiation of the graph may exist at any point in time. - The graph cannot be cloned. Parameters ---------- graph : :py:obj:`~.CUgraph` or :py:obj:`~.cudaGraph_t` Graph to which to add the node pDependencies : list[:py:obj:`~.cudaGraphNode_t`] Dependencies of the node numDependencies : size_t Number of dependencies nodeParams : :py:obj:`~.cudaMemAllocNodeParams` Parameters for the node Returns ------- cudaError_t :py:obj:`~.cudaSuccess`, :py:obj:`~.cudaErrorCudartUnloading`, :py:obj:`~.cudaErrorInitializationError`, :py:obj:`~.cudaErrorNotSupported`, :py:obj:`~.cudaErrorInvalidValue`, :py:obj:`~.cudaErrorOutOfMemory` pGraphNode : :py:obj:`~.cudaGraphNode_t` Returns newly created node See Also -------- :py:obj:`~.cudaGraphAddNode`, :py:obj:`~.cudaGraphAddMemFreeNode`, :py:obj:`~.cudaGraphMemAllocNodeGetParams`, :py:obj:`~.cudaDeviceGraphMemTrim`, :py:obj:`~.cudaDeviceGetGraphMemAttribute`, :py:obj:`~.cudaDeviceSetGraphMemAttribute`, :py:obj:`~.cudaMallocAsync`, :py:obj:`~.cudaFreeAsync`, :py:obj:`~.cudaGraphCreate`, :py:obj:`~.cudaGraphDestroyNode`, :py:obj:`~.cudaGraphAddChildGraphNode`, :py:obj:`~.cudaGraphAddEmptyNode`, :py:obj:`~.cudaGraphAddEventRecordNode`, :py:obj:`~.cudaGraphAddEventWaitNode`, :py:obj:`~.cudaGraphAddExternalSemaphoresSignalNode`, :py:obj:`~.cudaGraphAddExternalSemaphoresWaitNode`, :py:obj:`~.cudaGraphAddKernelNode`, :py:obj:`~.cudaGraphAddMemcpyNode`, :py:obj:`~.cudaGraphAddMemsetNode` """ pDependencies = [] if pDependencies is None else pDependencies if not all(isinstance(_x, (cudaGraphNode_t,driver.CUgraphNode)) for _x in pDependencies): raise TypeError("Argument 'pDependencies' is not instance of type (expected tuple[cyruntime.cudaGraphNode_t,driver.CUgraphNode] or list[cyruntime.cudaGraphNode_t,driver.CUgraphNode]") cdef cyruntime.cudaGraph_t cygraph if graph is None: pgraph = 0 elif isinstance(graph, (cudaGraph_t,driver.CUgraph)): pgraph = int(graph) else: pgraph = int(cudaGraph_t(graph)) cygraph = pgraph cdef cudaGraphNode_t pGraphNode = cudaGraphNode_t() cdef cyruntime.cudaGraphNode_t* cypDependencies = NULL if len(pDependencies) > 1: cypDependencies = calloc(len(pDependencies), sizeof(cyruntime.cudaGraphNode_t)) if cypDependencies is NULL: raise MemoryError('Failed to allocate length x size memory: ' + str(len(pDependencies)) + 'x' + str(sizeof(cyruntime.cudaGraphNode_t))) else: for idx in range(len(pDependencies)): cypDependencies[idx] = (pDependencies[idx])._pvt_ptr[0] elif len(pDependencies) == 1: cypDependencies = (pDependencies[0])._pvt_ptr if numDependencies > len(pDependencies): raise RuntimeError("List is too small: " + str(len(pDependencies)) + " < " + str(numDependencies)) cdef cyruntime.cudaMemAllocNodeParams* cynodeParams_ptr = nodeParams._pvt_ptr if nodeParams is not None else NULL with nogil: err = cyruntime.cudaGraphAddMemAllocNode(pGraphNode._pvt_ptr, cygraph, cypDependencies, numDependencies, cynodeParams_ptr) if len(pDependencies) > 1 and cypDependencies is not NULL: free(cypDependencies) if err != cyruntime.cudaSuccess: return (_dict_cudaError_t[err], None) return (_dict_cudaError_t[err], pGraphNode) @cython.embedsignature(True) def cudaGraphMemAllocNodeGetParams(node): """ Returns a memory alloc node's parameters. Returns the parameters of a memory alloc node `hNode` in `params_out`. The `poolProps` and `accessDescs` returned in `params_out`, are owned by the node. This memory remains valid until the node is destroyed. The returned parameters must not be modified. Parameters ---------- node : :py:obj:`~.CUgraphNode` or :py:obj:`~.cudaGraphNode_t` Node to get the parameters for Returns ------- cudaError_t :py:obj:`~.cudaSuccess`, :py:obj:`~.cudaErrorInvalidValue` params_out : :py:obj:`~.cudaMemAllocNodeParams` Pointer to return the parameters See Also -------- :py:obj:`~.cudaGraphAddMemAllocNode`, :py:obj:`~.cudaGraphMemFreeNodeGetParams` """ cdef cyruntime.cudaGraphNode_t cynode if node is None: pnode = 0 elif isinstance(node, (cudaGraphNode_t,driver.CUgraphNode)): pnode = int(node) else: pnode = int(cudaGraphNode_t(node)) cynode = pnode cdef cudaMemAllocNodeParams params_out = cudaMemAllocNodeParams() with nogil: err = cyruntime.cudaGraphMemAllocNodeGetParams(cynode, params_out._pvt_ptr) if err != cyruntime.cudaSuccess: return (_dict_cudaError_t[err], None) return (_dict_cudaError_t[err], params_out) @cython.embedsignature(True) def cudaGraphAddMemFreeNode(graph, pDependencies : Optional[tuple[cudaGraphNode_t] | list[cudaGraphNode_t]], size_t numDependencies, dptr): """ Creates a memory free node and adds it to a graph. Creates a new memory free node and adds it to `graph` with `numDependencies` dependencies specified via `pDependencies` and address specified in `dptr`. It is possible for `numDependencies` to be 0, in which case the node will be placed at the root of the graph. `pDependencies` may not have any duplicate entries. A handle to the new node will be returned in `pGraphNode`. :py:obj:`~.cudaGraphAddMemFreeNode` will return :py:obj:`~.cudaErrorInvalidValue` if the user attempts to free: - an allocation twice in the same graph. - an address that was not returned by an allocation node. - an invalid address. The following restrictions apply to graphs which contain allocation and/or memory free nodes: - Nodes and edges of the graph cannot be deleted. - The graph can only be used in a child node if the ownership is moved to the parent. - Only one instantiation of the graph may exist at any point in time. - The graph cannot be cloned. Parameters ---------- graph : :py:obj:`~.CUgraph` or :py:obj:`~.cudaGraph_t` Graph to which to add the node pDependencies : list[:py:obj:`~.cudaGraphNode_t`] Dependencies of the node numDependencies : size_t Number of dependencies dptr : Any Address of memory to free Returns ------- cudaError_t :py:obj:`~.cudaSuccess`, :py:obj:`~.cudaErrorCudartUnloading`, :py:obj:`~.cudaErrorInitializationError`, :py:obj:`~.cudaErrorNotSupported`, :py:obj:`~.cudaErrorInvalidValue`, :py:obj:`~.cudaErrorOutOfMemory` pGraphNode : :py:obj:`~.cudaGraphNode_t` Returns newly created node See Also -------- :py:obj:`~.cudaGraphAddNode`, :py:obj:`~.cudaGraphAddMemAllocNode`, :py:obj:`~.cudaGraphMemFreeNodeGetParams`, :py:obj:`~.cudaDeviceGraphMemTrim`, :py:obj:`~.cudaDeviceGetGraphMemAttribute`, :py:obj:`~.cudaDeviceSetGraphMemAttribute`, :py:obj:`~.cudaMallocAsync`, :py:obj:`~.cudaFreeAsync`, :py:obj:`~.cudaGraphCreate`, :py:obj:`~.cudaGraphDestroyNode`, :py:obj:`~.cudaGraphAddChildGraphNode`, :py:obj:`~.cudaGraphAddEmptyNode`, :py:obj:`~.cudaGraphAddEventRecordNode`, :py:obj:`~.cudaGraphAddEventWaitNode`, :py:obj:`~.cudaGraphAddExternalSemaphoresSignalNode`, :py:obj:`~.cudaGraphAddExternalSemaphoresWaitNode`, :py:obj:`~.cudaGraphAddKernelNode`, :py:obj:`~.cudaGraphAddMemcpyNode`, :py:obj:`~.cudaGraphAddMemsetNode` """ pDependencies = [] if pDependencies is None else pDependencies if not all(isinstance(_x, (cudaGraphNode_t,driver.CUgraphNode)) for _x in pDependencies): raise TypeError("Argument 'pDependencies' is not instance of type (expected tuple[cyruntime.cudaGraphNode_t,driver.CUgraphNode] or list[cyruntime.cudaGraphNode_t,driver.CUgraphNode]") cdef cyruntime.cudaGraph_t cygraph if graph is None: pgraph = 0 elif isinstance(graph, (cudaGraph_t,driver.CUgraph)): pgraph = int(graph) else: pgraph = int(cudaGraph_t(graph)) cygraph = pgraph cdef cudaGraphNode_t pGraphNode = cudaGraphNode_t() cdef cyruntime.cudaGraphNode_t* cypDependencies = NULL if len(pDependencies) > 1: cypDependencies = calloc(len(pDependencies), sizeof(cyruntime.cudaGraphNode_t)) if cypDependencies is NULL: raise MemoryError('Failed to allocate length x size memory: ' + str(len(pDependencies)) + 'x' + str(sizeof(cyruntime.cudaGraphNode_t))) else: for idx in range(len(pDependencies)): cypDependencies[idx] = (pDependencies[idx])._pvt_ptr[0] elif len(pDependencies) == 1: cypDependencies = (pDependencies[0])._pvt_ptr if numDependencies > len(pDependencies): raise RuntimeError("List is too small: " + str(len(pDependencies)) + " < " + str(numDependencies)) cydptr = _HelperInputVoidPtr(dptr) cdef void* cydptr_ptr = cydptr.cptr with nogil: err = cyruntime.cudaGraphAddMemFreeNode(pGraphNode._pvt_ptr, cygraph, cypDependencies, numDependencies, cydptr_ptr) if len(pDependencies) > 1 and cypDependencies is not NULL: free(cypDependencies) if err != cyruntime.cudaSuccess: return (_dict_cudaError_t[err], None) return (_dict_cudaError_t[err], pGraphNode) @cython.embedsignature(True) def cudaGraphMemFreeNodeGetParams(node): """ Returns a memory free node's parameters. Returns the address of a memory free node `hNode` in `dptr_out`. Parameters ---------- node : :py:obj:`~.CUgraphNode` or :py:obj:`~.cudaGraphNode_t` Node to get the parameters for Returns ------- cudaError_t :py:obj:`~.cudaSuccess`, :py:obj:`~.cudaErrorInvalidValue` dptr_out : Any Pointer to return the device address See Also -------- :py:obj:`~.cudaGraphAddMemFreeNode`, :py:obj:`~.cudaGraphMemFreeNodeGetParams` """ cdef cyruntime.cudaGraphNode_t cynode if node is None: pnode = 0 elif isinstance(node, (cudaGraphNode_t,driver.CUgraphNode)): pnode = int(node) else: pnode = int(cudaGraphNode_t(node)) cynode = pnode cdef void_ptr dptr_out = 0 cdef void* cydptr_out_ptr = &dptr_out with nogil: err = cyruntime.cudaGraphMemFreeNodeGetParams(cynode, cydptr_out_ptr) if err != cyruntime.cudaSuccess: return (_dict_cudaError_t[err], None) return (_dict_cudaError_t[err], dptr_out) @cython.embedsignature(True) def cudaDeviceGraphMemTrim(int device): """ Free unused memory that was cached on the specified device for use with graphs back to the OS. Blocks which are not in use by a graph that is either currently executing or scheduled to execute are freed back to the operating system. Parameters ---------- device : int The device for which cached memory should be freed. Returns ------- cudaError_t :py:obj:`~.cudaSuccess`, :py:obj:`~.cudaErrorInvalidValue` See Also -------- :py:obj:`~.cudaGraphAddMemAllocNode`, :py:obj:`~.cudaGraphAddMemFreeNode`, :py:obj:`~.cudaDeviceGetGraphMemAttribute`, :py:obj:`~.cudaDeviceSetGraphMemAttribute`, :py:obj:`~.cudaMallocAsync`, :py:obj:`~.cudaFreeAsync` """ with nogil: err = cyruntime.cudaDeviceGraphMemTrim(device) return (_dict_cudaError_t[err],) @cython.embedsignature(True) def cudaDeviceGetGraphMemAttribute(int device, attr not None : cudaGraphMemAttributeType): """ Query asynchronous allocation attributes related to graphs. Valid attributes are: - :py:obj:`~.cudaGraphMemAttrUsedMemCurrent`: Amount of memory, in bytes, currently associated with graphs - :py:obj:`~.cudaGraphMemAttrUsedMemHigh`: High watermark of memory, in bytes, associated with graphs since the last time it was reset. High watermark can only be reset to zero. - :py:obj:`~.cudaGraphMemAttrReservedMemCurrent`: Amount of memory, in bytes, currently allocated for use by the CUDA graphs asynchronous allocator. - :py:obj:`~.cudaGraphMemAttrReservedMemHigh`: High watermark of memory, in bytes, currently allocated for use by the CUDA graphs asynchronous allocator. Parameters ---------- device : int Specifies the scope of the query attr : :py:obj:`~.cudaGraphMemAttributeType` attribute to get Returns ------- cudaError_t :py:obj:`~.cudaSuccess`, :py:obj:`~.cudaErrorInvalidDevice` value : Any retrieved value See Also -------- :py:obj:`~.cudaDeviceSetGraphMemAttribute`, :py:obj:`~.cudaGraphAddMemAllocNode`, :py:obj:`~.cudaGraphAddMemFreeNode`, :py:obj:`~.cudaDeviceGraphMemTrim`, :py:obj:`~.cudaMallocAsync`, :py:obj:`~.cudaFreeAsync` """ cdef cyruntime.cudaGraphMemAttributeType cyattr = attr.value cdef _HelperCUgraphMem_attribute cyvalue = _HelperCUgraphMem_attribute(attr, 0, is_getter=True) cdef void* cyvalue_ptr = cyvalue.cptr with nogil: err = cyruntime.cudaDeviceGetGraphMemAttribute(device, cyattr, cyvalue_ptr) if err != cyruntime.cudaSuccess: return (_dict_cudaError_t[err], None) return (_dict_cudaError_t[err], cyvalue.pyObj()) @cython.embedsignature(True) def cudaDeviceSetGraphMemAttribute(int device, attr not None : cudaGraphMemAttributeType, value): """ Set asynchronous allocation attributes related to graphs. Valid attributes are: - :py:obj:`~.cudaGraphMemAttrUsedMemHigh`: High watermark of memory, in bytes, associated with graphs since the last time it was reset. High watermark can only be reset to zero. - :py:obj:`~.cudaGraphMemAttrReservedMemHigh`: High watermark of memory, in bytes, currently allocated for use by the CUDA graphs asynchronous allocator. Parameters ---------- device : int Specifies the scope of the query attr : :py:obj:`~.cudaGraphMemAttributeType` attribute to get value : Any pointer to value to set Returns ------- cudaError_t :py:obj:`~.cudaSuccess`, :py:obj:`~.cudaErrorInvalidDevice` See Also -------- :py:obj:`~.cudaDeviceGetGraphMemAttribute`, :py:obj:`~.cudaGraphAddMemAllocNode`, :py:obj:`~.cudaGraphAddMemFreeNode`, :py:obj:`~.cudaDeviceGraphMemTrim`, :py:obj:`~.cudaMallocAsync`, :py:obj:`~.cudaFreeAsync` """ cdef cyruntime.cudaGraphMemAttributeType cyattr = attr.value cdef _HelperCUgraphMem_attribute cyvalue = _HelperCUgraphMem_attribute(attr, value, is_getter=False) cdef void* cyvalue_ptr = cyvalue.cptr with nogil: err = cyruntime.cudaDeviceSetGraphMemAttribute(device, cyattr, cyvalue_ptr) return (_dict_cudaError_t[err],) @cython.embedsignature(True) def cudaGraphClone(originalGraph): """ Clones a graph. This function creates a copy of `originalGraph` and returns it in `pGraphClone`. All parameters are copied into the cloned graph. The original graph may be modified after this call without affecting the clone. Child graph nodes in the original graph are recursively copied into the clone. Parameters ---------- originalGraph : :py:obj:`~.CUgraph` or :py:obj:`~.cudaGraph_t` Graph to clone Returns ------- cudaError_t :py:obj:`~.cudaSuccess`, :py:obj:`~.cudaErrorInvalidValue`, :py:obj:`~.cudaErrorMemoryAllocation` pGraphClone : :py:obj:`~.cudaGraph_t` Returns newly created cloned graph See Also -------- :py:obj:`~.cudaGraphCreate`, :py:obj:`~.cudaGraphNodeFindInClone` Notes ----- : Cloning is not supported for graphs which contain memory allocation nodes, memory free nodes, or conditional nodes. """ cdef cyruntime.cudaGraph_t cyoriginalGraph if originalGraph is None: poriginalGraph = 0 elif isinstance(originalGraph, (cudaGraph_t,driver.CUgraph)): poriginalGraph = int(originalGraph) else: poriginalGraph = int(cudaGraph_t(originalGraph)) cyoriginalGraph = poriginalGraph cdef cudaGraph_t pGraphClone = cudaGraph_t() with nogil: err = cyruntime.cudaGraphClone(pGraphClone._pvt_ptr, cyoriginalGraph) if err != cyruntime.cudaSuccess: return (_dict_cudaError_t[err], None) return (_dict_cudaError_t[err], pGraphClone) @cython.embedsignature(True) def cudaGraphNodeFindInClone(originalNode, clonedGraph): """ Finds a cloned version of a node. This function returns the node in `clonedGraph` corresponding to `originalNode` in the original graph. `clonedGraph` must have been cloned from `originalGraph` via :py:obj:`~.cudaGraphClone`. `originalNode` must have been in `originalGraph` at the time of the call to :py:obj:`~.cudaGraphClone`, and the corresponding cloned node in `clonedGraph` must not have been removed. The cloned node is then returned via `pClonedNode`. Parameters ---------- originalNode : :py:obj:`~.CUgraphNode` or :py:obj:`~.cudaGraphNode_t` Handle to the original node clonedGraph : :py:obj:`~.CUgraph` or :py:obj:`~.cudaGraph_t` Cloned graph to query Returns ------- cudaError_t :py:obj:`~.cudaSuccess`, :py:obj:`~.cudaErrorInvalidValue` pNode : :py:obj:`~.cudaGraphNode_t` Returns handle to the cloned node See Also -------- :py:obj:`~.cudaGraphClone` """ cdef cyruntime.cudaGraph_t cyclonedGraph if clonedGraph is None: pclonedGraph = 0 elif isinstance(clonedGraph, (cudaGraph_t,driver.CUgraph)): pclonedGraph = int(clonedGraph) else: pclonedGraph = int(cudaGraph_t(clonedGraph)) cyclonedGraph = pclonedGraph cdef cyruntime.cudaGraphNode_t cyoriginalNode if originalNode is None: poriginalNode = 0 elif isinstance(originalNode, (cudaGraphNode_t,driver.CUgraphNode)): poriginalNode = int(originalNode) else: poriginalNode = int(cudaGraphNode_t(originalNode)) cyoriginalNode = poriginalNode cdef cudaGraphNode_t pNode = cudaGraphNode_t() with nogil: err = cyruntime.cudaGraphNodeFindInClone(pNode._pvt_ptr, cyoriginalNode, cyclonedGraph) if err != cyruntime.cudaSuccess: return (_dict_cudaError_t[err], None) return (_dict_cudaError_t[err], pNode) @cython.embedsignature(True) def cudaGraphNodeGetType(node): """ Returns a node's type. Returns the node type of `node` in `pType`. Parameters ---------- node : :py:obj:`~.CUgraphNode` or :py:obj:`~.cudaGraphNode_t` Node to query Returns ------- cudaError_t :py:obj:`~.cudaSuccess`, :py:obj:`~.cudaErrorInvalidValue` pType : :py:obj:`~.cudaGraphNodeType` Pointer to return the node type See Also -------- :py:obj:`~.cudaGraphGetNodes`, :py:obj:`~.cudaGraphGetRootNodes`, :py:obj:`~.cudaGraphChildGraphNodeGetGraph`, :py:obj:`~.cudaGraphKernelNodeGetParams`, :py:obj:`~.cudaGraphKernelNodeSetParams`, :py:obj:`~.cudaGraphHostNodeGetParams`, :py:obj:`~.cudaGraphHostNodeSetParams`, :py:obj:`~.cudaGraphMemcpyNodeGetParams`, :py:obj:`~.cudaGraphMemcpyNodeSetParams`, :py:obj:`~.cudaGraphMemsetNodeGetParams`, :py:obj:`~.cudaGraphMemsetNodeSetParams` """ cdef cyruntime.cudaGraphNode_t cynode if node is None: pnode = 0 elif isinstance(node, (cudaGraphNode_t,driver.CUgraphNode)): pnode = int(node) else: pnode = int(cudaGraphNode_t(node)) cynode = pnode cdef cyruntime.cudaGraphNodeType pType with nogil: err = cyruntime.cudaGraphNodeGetType(cynode, &pType) if err != cyruntime.cudaSuccess: return (_dict_cudaError_t[err], None) return (_dict_cudaError_t[err], cudaGraphNodeType(pType)) @cython.embedsignature(True) def cudaGraphGetNodes(graph, size_t numNodes = 0): """ Returns a graph's nodes. Returns a list of `graph's` nodes. `nodes` may be NULL, in which case this function will return the number of nodes in `numNodes`. Otherwise, `numNodes` entries will be filled in. If `numNodes` is higher than the actual number of nodes, the remaining entries in `nodes` will be set to NULL, and the number of nodes actually obtained will be returned in `numNodes`. Parameters ---------- graph : :py:obj:`~.CUgraph` or :py:obj:`~.cudaGraph_t` Graph to query numNodes : int See description Returns ------- cudaError_t :py:obj:`~.cudaSuccess`, :py:obj:`~.cudaErrorInvalidValue` nodes : list[:py:obj:`~.cudaGraphNode_t`] Pointer to return the nodes numNodes : int See description See Also -------- :py:obj:`~.cudaGraphCreate`, :py:obj:`~.cudaGraphGetRootNodes`, :py:obj:`~.cudaGraphGetEdges`, :py:obj:`~.cudaGraphNodeGetType`, :py:obj:`~.cudaGraphNodeGetDependencies`, :py:obj:`~.cudaGraphNodeGetDependentNodes` """ cdef size_t _graph_length = numNodes cdef cyruntime.cudaGraph_t cygraph if graph is None: pgraph = 0 elif isinstance(graph, (cudaGraph_t,driver.CUgraph)): pgraph = int(graph) else: pgraph = int(cudaGraph_t(graph)) cygraph = pgraph cdef cyruntime.cudaGraphNode_t* cynodes = NULL pynodes = [] if _graph_length != 0: cynodes = calloc(_graph_length, sizeof(cyruntime.cudaGraphNode_t)) if cynodes is NULL: raise MemoryError('Failed to allocate length x size memory: ' + str(_graph_length) + 'x' + str(sizeof(cyruntime.cudaGraphNode_t))) with nogil: err = cyruntime.cudaGraphGetNodes(cygraph, cynodes, &numNodes) if cudaError_t(err) == cudaError_t(0): pynodes = [cudaGraphNode_t(init_value=cynodes[idx]) for idx in range(_graph_length)] if cynodes is not NULL: free(cynodes) if err != cyruntime.cudaSuccess: return (_dict_cudaError_t[err], None, None) return (_dict_cudaError_t[err], pynodes, numNodes) @cython.embedsignature(True) def cudaGraphGetRootNodes(graph, size_t pNumRootNodes = 0): """ Returns a graph's root nodes. Returns a list of `graph's` root nodes. `pRootNodes` may be NULL, in which case this function will return the number of root nodes in `pNumRootNodes`. Otherwise, `pNumRootNodes` entries will be filled in. If `pNumRootNodes` is higher than the actual number of root nodes, the remaining entries in `pRootNodes` will be set to NULL, and the number of nodes actually obtained will be returned in `pNumRootNodes`. Parameters ---------- graph : :py:obj:`~.CUgraph` or :py:obj:`~.cudaGraph_t` Graph to query pNumRootNodes : int See description Returns ------- cudaError_t :py:obj:`~.cudaSuccess`, :py:obj:`~.cudaErrorInvalidValue` pRootNodes : list[:py:obj:`~.cudaGraphNode_t`] Pointer to return the root nodes pNumRootNodes : int See description See Also -------- :py:obj:`~.cudaGraphCreate`, :py:obj:`~.cudaGraphGetNodes`, :py:obj:`~.cudaGraphGetEdges`, :py:obj:`~.cudaGraphNodeGetType`, :py:obj:`~.cudaGraphNodeGetDependencies`, :py:obj:`~.cudaGraphNodeGetDependentNodes` """ cdef size_t _graph_length = pNumRootNodes cdef cyruntime.cudaGraph_t cygraph if graph is None: pgraph = 0 elif isinstance(graph, (cudaGraph_t,driver.CUgraph)): pgraph = int(graph) else: pgraph = int(cudaGraph_t(graph)) cygraph = pgraph cdef cyruntime.cudaGraphNode_t* cypRootNodes = NULL pypRootNodes = [] if _graph_length != 0: cypRootNodes = calloc(_graph_length, sizeof(cyruntime.cudaGraphNode_t)) if cypRootNodes is NULL: raise MemoryError('Failed to allocate length x size memory: ' + str(_graph_length) + 'x' + str(sizeof(cyruntime.cudaGraphNode_t))) with nogil: err = cyruntime.cudaGraphGetRootNodes(cygraph, cypRootNodes, &pNumRootNodes) if cudaError_t(err) == cudaError_t(0): pypRootNodes = [cudaGraphNode_t(init_value=cypRootNodes[idx]) for idx in range(_graph_length)] if cypRootNodes is not NULL: free(cypRootNodes) if err != cyruntime.cudaSuccess: return (_dict_cudaError_t[err], None, None) return (_dict_cudaError_t[err], pypRootNodes, pNumRootNodes) @cython.embedsignature(True) def cudaGraphGetEdges(graph, size_t numEdges = 0): """ Returns a graph's dependency edges. Returns a list of `graph's` dependency edges. Edges are returned via corresponding indices in `from` and `to`; that is, the node in `to`[i] has a dependency on the node in `from`[i]. `from` and `to` may both be NULL, in which case this function only returns the number of edges in `numEdges`. Otherwise, `numEdges` entries will be filled in. If `numEdges` is higher than the actual number of edges, the remaining entries in `from` and `to` will be set to NULL, and the number of edges actually returned will be written to `numEdges`. Parameters ---------- graph : :py:obj:`~.CUgraph` or :py:obj:`~.cudaGraph_t` Graph to get the edges from numEdges : int See description Returns ------- cudaError_t :py:obj:`~.cudaSuccess`, :py:obj:`~.cudaErrorInvalidValue` from : list[:py:obj:`~.cudaGraphNode_t`] Location to return edge endpoints to : list[:py:obj:`~.cudaGraphNode_t`] Location to return edge endpoints numEdges : int See description See Also -------- :py:obj:`~.cudaGraphGetNodes`, :py:obj:`~.cudaGraphGetRootNodes`, :py:obj:`~.cudaGraphAddDependencies`, :py:obj:`~.cudaGraphRemoveDependencies`, :py:obj:`~.cudaGraphNodeGetDependencies`, :py:obj:`~.cudaGraphNodeGetDependentNodes` """ cdef size_t _graph_length = numEdges cdef cyruntime.cudaGraph_t cygraph if graph is None: pgraph = 0 elif isinstance(graph, (cudaGraph_t,driver.CUgraph)): pgraph = int(graph) else: pgraph = int(cudaGraph_t(graph)) cygraph = pgraph cdef cyruntime.cudaGraphNode_t* cyfrom_ = NULL pyfrom_ = [] if _graph_length != 0: cyfrom_ = calloc(_graph_length, sizeof(cyruntime.cudaGraphNode_t)) if cyfrom_ is NULL: raise MemoryError('Failed to allocate length x size memory: ' + str(_graph_length) + 'x' + str(sizeof(cyruntime.cudaGraphNode_t))) cdef cyruntime.cudaGraphNode_t* cyto = NULL pyto = [] if _graph_length != 0: cyto = calloc(_graph_length, sizeof(cyruntime.cudaGraphNode_t)) if cyto is NULL: raise MemoryError('Failed to allocate length x size memory: ' + str(_graph_length) + 'x' + str(sizeof(cyruntime.cudaGraphNode_t))) with nogil: err = cyruntime.cudaGraphGetEdges(cygraph, cyfrom_, cyto, &numEdges) if cudaError_t(err) == cudaError_t(0): pyfrom_ = [cudaGraphNode_t(init_value=cyfrom_[idx]) for idx in range(_graph_length)] if cyfrom_ is not NULL: free(cyfrom_) if cudaError_t(err) == cudaError_t(0): pyto = [cudaGraphNode_t(init_value=cyto[idx]) for idx in range(_graph_length)] if cyto is not NULL: free(cyto) if err != cyruntime.cudaSuccess: return (_dict_cudaError_t[err], None, None, None) return (_dict_cudaError_t[err], pyfrom_, pyto, numEdges) @cython.embedsignature(True) def cudaGraphGetEdges_v2(graph, size_t numEdges = 0): """ Returns a graph's dependency edges (12.3+) Returns a list of `graph's` dependency edges. Edges are returned via corresponding indices in `from`, `to` and `edgeData`; that is, the node in `to`[i] has a dependency on the node in `from`[i] with data `edgeData`[i]. `from` and `to` may both be NULL, in which case this function only returns the number of edges in `numEdges`. Otherwise, `numEdges` entries will be filled in. If `numEdges` is higher than the actual number of edges, the remaining entries in `from` and `to` will be set to NULL, and the number of edges actually returned will be written to `numEdges`. `edgeData` may alone be NULL, in which case the edges must all have default (zeroed) edge data. Attempting a losst query via NULL `edgeData` will result in :py:obj:`~.cudaErrorLossyQuery`. If `edgeData` is non-NULL then `from` and `to` must be as well. Parameters ---------- graph : :py:obj:`~.CUgraph` or :py:obj:`~.cudaGraph_t` Graph to get the edges from numEdges : int See description Returns ------- cudaError_t :py:obj:`~.cudaSuccess`, :py:obj:`~.cudaErrorLossyQuery`, :py:obj:`~.cudaErrorInvalidValue` from : list[:py:obj:`~.cudaGraphNode_t`] Location to return edge endpoints to : list[:py:obj:`~.cudaGraphNode_t`] Location to return edge endpoints edgeData : list[:py:obj:`~.cudaGraphEdgeData`] Optional location to return edge data numEdges : int See description See Also -------- :py:obj:`~.cudaGraphGetNodes`, :py:obj:`~.cudaGraphGetRootNodes`, :py:obj:`~.cudaGraphAddDependencies`, :py:obj:`~.cudaGraphRemoveDependencies`, :py:obj:`~.cudaGraphNodeGetDependencies`, :py:obj:`~.cudaGraphNodeGetDependentNodes` """ cdef size_t _graph_length = numEdges cdef cyruntime.cudaGraph_t cygraph if graph is None: pgraph = 0 elif isinstance(graph, (cudaGraph_t,driver.CUgraph)): pgraph = int(graph) else: pgraph = int(cudaGraph_t(graph)) cygraph = pgraph cdef cyruntime.cudaGraphNode_t* cyfrom_ = NULL pyfrom_ = [] if _graph_length != 0: cyfrom_ = calloc(_graph_length, sizeof(cyruntime.cudaGraphNode_t)) if cyfrom_ is NULL: raise MemoryError('Failed to allocate length x size memory: ' + str(_graph_length) + 'x' + str(sizeof(cyruntime.cudaGraphNode_t))) cdef cyruntime.cudaGraphNode_t* cyto = NULL pyto = [] if _graph_length != 0: cyto = calloc(_graph_length, sizeof(cyruntime.cudaGraphNode_t)) if cyto is NULL: raise MemoryError('Failed to allocate length x size memory: ' + str(_graph_length) + 'x' + str(sizeof(cyruntime.cudaGraphNode_t))) cdef cyruntime.cudaGraphEdgeData* cyedgeData = NULL pyedgeData = [] if _graph_length != 0: cyedgeData = calloc(_graph_length, sizeof(cyruntime.cudaGraphEdgeData)) if cyedgeData is NULL: raise MemoryError('Failed to allocate length x size memory: ' + str(_graph_length) + 'x' + str(sizeof(cyruntime.cudaGraphEdgeData))) with nogil: err = cyruntime.cudaGraphGetEdges_v2(cygraph, cyfrom_, cyto, cyedgeData, &numEdges) if cudaError_t(err) == cudaError_t(0): pyfrom_ = [cudaGraphNode_t(init_value=cyfrom_[idx]) for idx in range(_graph_length)] if cyfrom_ is not NULL: free(cyfrom_) if cudaError_t(err) == cudaError_t(0): pyto = [cudaGraphNode_t(init_value=cyto[idx]) for idx in range(_graph_length)] if cyto is not NULL: free(cyto) if cudaError_t(err) == cudaError_t(0): pyedgeData = [cudaGraphEdgeData(_ptr=&cyedgeData[idx]) for idx in range(_graph_length)] if cyedgeData is not NULL: free(cyedgeData) if err != cyruntime.cudaSuccess: return (_dict_cudaError_t[err], None, None, None, None) return (_dict_cudaError_t[err], pyfrom_, pyto, pyedgeData, numEdges) @cython.embedsignature(True) def cudaGraphNodeGetDependencies(node, size_t pNumDependencies = 0): """ Returns a node's dependencies. Returns a list of `node's` dependencies. `pDependencies` may be NULL, in which case this function will return the number of dependencies in `pNumDependencies`. Otherwise, `pNumDependencies` entries will be filled in. If `pNumDependencies` is higher than the actual number of dependencies, the remaining entries in `pDependencies` will be set to NULL, and the number of nodes actually obtained will be returned in `pNumDependencies`. Parameters ---------- node : :py:obj:`~.CUgraphNode` or :py:obj:`~.cudaGraphNode_t` Node to query pNumDependencies : int See description Returns ------- cudaError_t :py:obj:`~.cudaSuccess`, :py:obj:`~.cudaErrorInvalidValue` pDependencies : list[:py:obj:`~.cudaGraphNode_t`] Pointer to return the dependencies pNumDependencies : int See description See Also -------- :py:obj:`~.cudaGraphNodeGetDependentNodes`, :py:obj:`~.cudaGraphGetNodes`, :py:obj:`~.cudaGraphGetRootNodes`, :py:obj:`~.cudaGraphGetEdges`, :py:obj:`~.cudaGraphAddDependencies`, :py:obj:`~.cudaGraphRemoveDependencies` """ cdef size_t _graph_length = pNumDependencies cdef cyruntime.cudaGraphNode_t cynode if node is None: pnode = 0 elif isinstance(node, (cudaGraphNode_t,driver.CUgraphNode)): pnode = int(node) else: pnode = int(cudaGraphNode_t(node)) cynode = pnode cdef cyruntime.cudaGraphNode_t* cypDependencies = NULL pypDependencies = [] if _graph_length != 0: cypDependencies = calloc(_graph_length, sizeof(cyruntime.cudaGraphNode_t)) if cypDependencies is NULL: raise MemoryError('Failed to allocate length x size memory: ' + str(_graph_length) + 'x' + str(sizeof(cyruntime.cudaGraphNode_t))) with nogil: err = cyruntime.cudaGraphNodeGetDependencies(cynode, cypDependencies, &pNumDependencies) if cudaError_t(err) == cudaError_t(0): pypDependencies = [cudaGraphNode_t(init_value=cypDependencies[idx]) for idx in range(_graph_length)] if cypDependencies is not NULL: free(cypDependencies) if err != cyruntime.cudaSuccess: return (_dict_cudaError_t[err], None, None) return (_dict_cudaError_t[err], pypDependencies, pNumDependencies) @cython.embedsignature(True) def cudaGraphNodeGetDependencies_v2(node, size_t pNumDependencies = 0): """ Returns a node's dependencies (12.3+) Returns a list of `node's` dependencies. `pDependencies` may be NULL, in which case this function will return the number of dependencies in `pNumDependencies`. Otherwise, `pNumDependencies` entries will be filled in. If `pNumDependencies` is higher than the actual number of dependencies, the remaining entries in `pDependencies` will be set to NULL, and the number of nodes actually obtained will be returned in `pNumDependencies`. Note that if an edge has non-zero (non-default) edge data and `edgeData` is NULL, this API will return :py:obj:`~.cudaErrorLossyQuery`. If `edgeData` is non-NULL, then `pDependencies` must be as well. Parameters ---------- node : :py:obj:`~.CUgraphNode` or :py:obj:`~.cudaGraphNode_t` Node to query pNumDependencies : int See description Returns ------- cudaError_t :py:obj:`~.cudaSuccess`, :py:obj:`~.cudaErrorLossyQuery`, :py:obj:`~.cudaErrorInvalidValue` pDependencies : list[:py:obj:`~.cudaGraphNode_t`] Pointer to return the dependencies edgeData : list[:py:obj:`~.cudaGraphEdgeData`] Optional array to return edge data for each dependency pNumDependencies : int See description See Also -------- :py:obj:`~.cudaGraphNodeGetDependentNodes`, :py:obj:`~.cudaGraphGetNodes`, :py:obj:`~.cudaGraphGetRootNodes`, :py:obj:`~.cudaGraphGetEdges`, :py:obj:`~.cudaGraphAddDependencies`, :py:obj:`~.cudaGraphRemoveDependencies` """ cdef size_t _graph_length = pNumDependencies cdef cyruntime.cudaGraphNode_t cynode if node is None: pnode = 0 elif isinstance(node, (cudaGraphNode_t,driver.CUgraphNode)): pnode = int(node) else: pnode = int(cudaGraphNode_t(node)) cynode = pnode cdef cyruntime.cudaGraphNode_t* cypDependencies = NULL pypDependencies = [] if _graph_length != 0: cypDependencies = calloc(_graph_length, sizeof(cyruntime.cudaGraphNode_t)) if cypDependencies is NULL: raise MemoryError('Failed to allocate length x size memory: ' + str(_graph_length) + 'x' + str(sizeof(cyruntime.cudaGraphNode_t))) cdef cyruntime.cudaGraphEdgeData* cyedgeData = NULL pyedgeData = [] if _graph_length != 0: cyedgeData = calloc(_graph_length, sizeof(cyruntime.cudaGraphEdgeData)) if cyedgeData is NULL: raise MemoryError('Failed to allocate length x size memory: ' + str(_graph_length) + 'x' + str(sizeof(cyruntime.cudaGraphEdgeData))) with nogil: err = cyruntime.cudaGraphNodeGetDependencies_v2(cynode, cypDependencies, cyedgeData, &pNumDependencies) if cudaError_t(err) == cudaError_t(0): pypDependencies = [cudaGraphNode_t(init_value=cypDependencies[idx]) for idx in range(_graph_length)] if cypDependencies is not NULL: free(cypDependencies) if cudaError_t(err) == cudaError_t(0): pyedgeData = [cudaGraphEdgeData(_ptr=&cyedgeData[idx]) for idx in range(_graph_length)] if cyedgeData is not NULL: free(cyedgeData) if err != cyruntime.cudaSuccess: return (_dict_cudaError_t[err], None, None, None) return (_dict_cudaError_t[err], pypDependencies, pyedgeData, pNumDependencies) @cython.embedsignature(True) def cudaGraphNodeGetDependentNodes(node, size_t pNumDependentNodes = 0): """ Returns a node's dependent nodes. Returns a list of `node's` dependent nodes. `pDependentNodes` may be NULL, in which case this function will return the number of dependent nodes in `pNumDependentNodes`. Otherwise, `pNumDependentNodes` entries will be filled in. If `pNumDependentNodes` is higher than the actual number of dependent nodes, the remaining entries in `pDependentNodes` will be set to NULL, and the number of nodes actually obtained will be returned in `pNumDependentNodes`. Parameters ---------- node : :py:obj:`~.CUgraphNode` or :py:obj:`~.cudaGraphNode_t` Node to query pNumDependentNodes : int See description Returns ------- cudaError_t :py:obj:`~.cudaSuccess`, :py:obj:`~.cudaErrorInvalidValue` pDependentNodes : list[:py:obj:`~.cudaGraphNode_t`] Pointer to return the dependent nodes pNumDependentNodes : int See description See Also -------- :py:obj:`~.cudaGraphNodeGetDependencies`, :py:obj:`~.cudaGraphGetNodes`, :py:obj:`~.cudaGraphGetRootNodes`, :py:obj:`~.cudaGraphGetEdges`, :py:obj:`~.cudaGraphAddDependencies`, :py:obj:`~.cudaGraphRemoveDependencies` """ cdef size_t _graph_length = pNumDependentNodes cdef cyruntime.cudaGraphNode_t cynode if node is None: pnode = 0 elif isinstance(node, (cudaGraphNode_t,driver.CUgraphNode)): pnode = int(node) else: pnode = int(cudaGraphNode_t(node)) cynode = pnode cdef cyruntime.cudaGraphNode_t* cypDependentNodes = NULL pypDependentNodes = [] if _graph_length != 0: cypDependentNodes = calloc(_graph_length, sizeof(cyruntime.cudaGraphNode_t)) if cypDependentNodes is NULL: raise MemoryError('Failed to allocate length x size memory: ' + str(_graph_length) + 'x' + str(sizeof(cyruntime.cudaGraphNode_t))) with nogil: err = cyruntime.cudaGraphNodeGetDependentNodes(cynode, cypDependentNodes, &pNumDependentNodes) if cudaError_t(err) == cudaError_t(0): pypDependentNodes = [cudaGraphNode_t(init_value=cypDependentNodes[idx]) for idx in range(_graph_length)] if cypDependentNodes is not NULL: free(cypDependentNodes) if err != cyruntime.cudaSuccess: return (_dict_cudaError_t[err], None, None) return (_dict_cudaError_t[err], pypDependentNodes, pNumDependentNodes) @cython.embedsignature(True) def cudaGraphNodeGetDependentNodes_v2(node, size_t pNumDependentNodes = 0): """ Returns a node's dependent nodes (12.3+) Returns a list of `node's` dependent nodes. `pDependentNodes` may be NULL, in which case this function will return the number of dependent nodes in `pNumDependentNodes`. Otherwise, `pNumDependentNodes` entries will be filled in. If `pNumDependentNodes` is higher than the actual number of dependent nodes, the remaining entries in `pDependentNodes` will be set to NULL, and the number of nodes actually obtained will be returned in `pNumDependentNodes`. Note that if an edge has non-zero (non-default) edge data and `edgeData` is NULL, this API will return :py:obj:`~.cudaErrorLossyQuery`. If `edgeData` is non-NULL, then `pDependentNodes` must be as well. Parameters ---------- node : :py:obj:`~.CUgraphNode` or :py:obj:`~.cudaGraphNode_t` Node to query pNumDependentNodes : int See description Returns ------- cudaError_t :py:obj:`~.cudaSuccess`, :py:obj:`~.cudaErrorLossyQuery`, :py:obj:`~.cudaErrorInvalidValue` pDependentNodes : list[:py:obj:`~.cudaGraphNode_t`] Pointer to return the dependent nodes edgeData : list[:py:obj:`~.cudaGraphEdgeData`] Optional pointer to return edge data for dependent nodes pNumDependentNodes : int See description See Also -------- :py:obj:`~.cudaGraphNodeGetDependencies`, :py:obj:`~.cudaGraphGetNodes`, :py:obj:`~.cudaGraphGetRootNodes`, :py:obj:`~.cudaGraphGetEdges`, :py:obj:`~.cudaGraphAddDependencies`, :py:obj:`~.cudaGraphRemoveDependencies` """ cdef size_t _graph_length = pNumDependentNodes cdef cyruntime.cudaGraphNode_t cynode if node is None: pnode = 0 elif isinstance(node, (cudaGraphNode_t,driver.CUgraphNode)): pnode = int(node) else: pnode = int(cudaGraphNode_t(node)) cynode = pnode cdef cyruntime.cudaGraphNode_t* cypDependentNodes = NULL pypDependentNodes = [] if _graph_length != 0: cypDependentNodes = calloc(_graph_length, sizeof(cyruntime.cudaGraphNode_t)) if cypDependentNodes is NULL: raise MemoryError('Failed to allocate length x size memory: ' + str(_graph_length) + 'x' + str(sizeof(cyruntime.cudaGraphNode_t))) cdef cyruntime.cudaGraphEdgeData* cyedgeData = NULL pyedgeData = [] if _graph_length != 0: cyedgeData = calloc(_graph_length, sizeof(cyruntime.cudaGraphEdgeData)) if cyedgeData is NULL: raise MemoryError('Failed to allocate length x size memory: ' + str(_graph_length) + 'x' + str(sizeof(cyruntime.cudaGraphEdgeData))) with nogil: err = cyruntime.cudaGraphNodeGetDependentNodes_v2(cynode, cypDependentNodes, cyedgeData, &pNumDependentNodes) if cudaError_t(err) == cudaError_t(0): pypDependentNodes = [cudaGraphNode_t(init_value=cypDependentNodes[idx]) for idx in range(_graph_length)] if cypDependentNodes is not NULL: free(cypDependentNodes) if cudaError_t(err) == cudaError_t(0): pyedgeData = [cudaGraphEdgeData(_ptr=&cyedgeData[idx]) for idx in range(_graph_length)] if cyedgeData is not NULL: free(cyedgeData) if err != cyruntime.cudaSuccess: return (_dict_cudaError_t[err], None, None, None) return (_dict_cudaError_t[err], pypDependentNodes, pyedgeData, pNumDependentNodes) @cython.embedsignature(True) def cudaGraphAddDependencies(graph, from_ : Optional[tuple[cudaGraphNode_t] | list[cudaGraphNode_t]], to : Optional[tuple[cudaGraphNode_t] | list[cudaGraphNode_t]], size_t numDependencies): """ Adds dependency edges to a graph. The number of dependencies to be added is defined by `numDependencies` Elements in `pFrom` and `pTo` at corresponding indices define a dependency. Each node in `pFrom` and `pTo` must belong to `graph`. If `numDependencies` is 0, elements in `pFrom` and `pTo` will be ignored. Specifying an existing dependency will return an error. Parameters ---------- graph : :py:obj:`~.CUgraph` or :py:obj:`~.cudaGraph_t` Graph to which dependencies are added from : list[:py:obj:`~.cudaGraphNode_t`] Array of nodes that provide the dependencies to : list[:py:obj:`~.cudaGraphNode_t`] Array of dependent nodes numDependencies : size_t Number of dependencies to be added Returns ------- cudaError_t :py:obj:`~.cudaSuccess`, :py:obj:`~.cudaErrorInvalidValue` See Also -------- :py:obj:`~.cudaGraphRemoveDependencies`, :py:obj:`~.cudaGraphGetEdges`, :py:obj:`~.cudaGraphNodeGetDependencies`, :py:obj:`~.cudaGraphNodeGetDependentNodes` """ to = [] if to is None else to if not all(isinstance(_x, (cudaGraphNode_t,driver.CUgraphNode)) for _x in to): raise TypeError("Argument 'to' is not instance of type (expected tuple[cyruntime.cudaGraphNode_t,driver.CUgraphNode] or list[cyruntime.cudaGraphNode_t,driver.CUgraphNode]") from_ = [] if from_ is None else from_ if not all(isinstance(_x, (cudaGraphNode_t,driver.CUgraphNode)) for _x in from_): raise TypeError("Argument 'from_' is not instance of type (expected tuple[cyruntime.cudaGraphNode_t,driver.CUgraphNode] or list[cyruntime.cudaGraphNode_t,driver.CUgraphNode]") cdef cyruntime.cudaGraph_t cygraph if graph is None: pgraph = 0 elif isinstance(graph, (cudaGraph_t,driver.CUgraph)): pgraph = int(graph) else: pgraph = int(cudaGraph_t(graph)) cygraph = pgraph cdef cyruntime.cudaGraphNode_t* cyfrom_ = NULL if len(from_) > 1: cyfrom_ = calloc(len(from_), sizeof(cyruntime.cudaGraphNode_t)) if cyfrom_ is NULL: raise MemoryError('Failed to allocate length x size memory: ' + str(len(from_)) + 'x' + str(sizeof(cyruntime.cudaGraphNode_t))) else: for idx in range(len(from_)): cyfrom_[idx] = (from_[idx])._pvt_ptr[0] elif len(from_) == 1: cyfrom_ = (from_[0])._pvt_ptr cdef cyruntime.cudaGraphNode_t* cyto = NULL if len(to) > 1: cyto = calloc(len(to), sizeof(cyruntime.cudaGraphNode_t)) if cyto is NULL: raise MemoryError('Failed to allocate length x size memory: ' + str(len(to)) + 'x' + str(sizeof(cyruntime.cudaGraphNode_t))) else: for idx in range(len(to)): cyto[idx] = (to[idx])._pvt_ptr[0] elif len(to) == 1: cyto = (to[0])._pvt_ptr if numDependencies > len(from_): raise RuntimeError("List is too small: " + str(len(from_)) + " < " + str(numDependencies)) if numDependencies > len(to): raise RuntimeError("List is too small: " + str(len(to)) + " < " + str(numDependencies)) with nogil: err = cyruntime.cudaGraphAddDependencies(cygraph, cyfrom_, cyto, numDependencies) if len(from_) > 1 and cyfrom_ is not NULL: free(cyfrom_) if len(to) > 1 and cyto is not NULL: free(cyto) return (_dict_cudaError_t[err],) @cython.embedsignature(True) def cudaGraphAddDependencies_v2(graph, from_ : Optional[tuple[cudaGraphNode_t] | list[cudaGraphNode_t]], to : Optional[tuple[cudaGraphNode_t] | list[cudaGraphNode_t]], edgeData : Optional[tuple[cudaGraphEdgeData] | list[cudaGraphEdgeData]], size_t numDependencies): """ Adds dependency edges to a graph. (12.3+) The number of dependencies to be added is defined by `numDependencies` Elements in `pFrom` and `pTo` at corresponding indices define a dependency. Each node in `pFrom` and `pTo` must belong to `graph`. If `numDependencies` is 0, elements in `pFrom` and `pTo` will be ignored. Specifying an existing dependency will return an error. Parameters ---------- graph : :py:obj:`~.CUgraph` or :py:obj:`~.cudaGraph_t` Graph to which dependencies are added from : list[:py:obj:`~.cudaGraphNode_t`] Array of nodes that provide the dependencies to : list[:py:obj:`~.cudaGraphNode_t`] Array of dependent nodes edgeData : list[:py:obj:`~.cudaGraphEdgeData`] Optional array of edge data. If NULL, default (zeroed) edge data is assumed. numDependencies : size_t Number of dependencies to be added Returns ------- cudaError_t :py:obj:`~.cudaSuccess`, :py:obj:`~.cudaErrorInvalidValue` See Also -------- :py:obj:`~.cudaGraphRemoveDependencies`, :py:obj:`~.cudaGraphGetEdges`, :py:obj:`~.cudaGraphNodeGetDependencies`, :py:obj:`~.cudaGraphNodeGetDependentNodes` """ edgeData = [] if edgeData is None else edgeData if not all(isinstance(_x, (cudaGraphEdgeData,)) for _x in edgeData): raise TypeError("Argument 'edgeData' is not instance of type (expected tuple[cyruntime.cudaGraphEdgeData,] or list[cyruntime.cudaGraphEdgeData,]") to = [] if to is None else to if not all(isinstance(_x, (cudaGraphNode_t,driver.CUgraphNode)) for _x in to): raise TypeError("Argument 'to' is not instance of type (expected tuple[cyruntime.cudaGraphNode_t,driver.CUgraphNode] or list[cyruntime.cudaGraphNode_t,driver.CUgraphNode]") from_ = [] if from_ is None else from_ if not all(isinstance(_x, (cudaGraphNode_t,driver.CUgraphNode)) for _x in from_): raise TypeError("Argument 'from_' is not instance of type (expected tuple[cyruntime.cudaGraphNode_t,driver.CUgraphNode] or list[cyruntime.cudaGraphNode_t,driver.CUgraphNode]") cdef cyruntime.cudaGraph_t cygraph if graph is None: pgraph = 0 elif isinstance(graph, (cudaGraph_t,driver.CUgraph)): pgraph = int(graph) else: pgraph = int(cudaGraph_t(graph)) cygraph = pgraph cdef cyruntime.cudaGraphNode_t* cyfrom_ = NULL if len(from_) > 1: cyfrom_ = calloc(len(from_), sizeof(cyruntime.cudaGraphNode_t)) if cyfrom_ is NULL: raise MemoryError('Failed to allocate length x size memory: ' + str(len(from_)) + 'x' + str(sizeof(cyruntime.cudaGraphNode_t))) else: for idx in range(len(from_)): cyfrom_[idx] = (from_[idx])._pvt_ptr[0] elif len(from_) == 1: cyfrom_ = (from_[0])._pvt_ptr cdef cyruntime.cudaGraphNode_t* cyto = NULL if len(to) > 1: cyto = calloc(len(to), sizeof(cyruntime.cudaGraphNode_t)) if cyto is NULL: raise MemoryError('Failed to allocate length x size memory: ' + str(len(to)) + 'x' + str(sizeof(cyruntime.cudaGraphNode_t))) else: for idx in range(len(to)): cyto[idx] = (to[idx])._pvt_ptr[0] elif len(to) == 1: cyto = (to[0])._pvt_ptr cdef cyruntime.cudaGraphEdgeData* cyedgeData = NULL if len(edgeData) > 1: cyedgeData = calloc(len(edgeData), sizeof(cyruntime.cudaGraphEdgeData)) if cyedgeData is NULL: raise MemoryError('Failed to allocate length x size memory: ' + str(len(edgeData)) + 'x' + str(sizeof(cyruntime.cudaGraphEdgeData))) for idx in range(len(edgeData)): string.memcpy(&cyedgeData[idx], (edgeData[idx])._pvt_ptr, sizeof(cyruntime.cudaGraphEdgeData)) elif len(edgeData) == 1: cyedgeData = (edgeData[0])._pvt_ptr with nogil: err = cyruntime.cudaGraphAddDependencies_v2(cygraph, cyfrom_, cyto, cyedgeData, numDependencies) if len(from_) > 1 and cyfrom_ is not NULL: free(cyfrom_) if len(to) > 1 and cyto is not NULL: free(cyto) if len(edgeData) > 1 and cyedgeData is not NULL: free(cyedgeData) return (_dict_cudaError_t[err],) @cython.embedsignature(True) def cudaGraphRemoveDependencies(graph, from_ : Optional[tuple[cudaGraphNode_t] | list[cudaGraphNode_t]], to : Optional[tuple[cudaGraphNode_t] | list[cudaGraphNode_t]], size_t numDependencies): """ Removes dependency edges from a graph. The number of `pDependencies` to be removed is defined by `numDependencies`. Elements in `pFrom` and `pTo` at corresponding indices define a dependency. Each node in `pFrom` and `pTo` must belong to `graph`. If `numDependencies` is 0, elements in `pFrom` and `pTo` will be ignored. Specifying a non-existing dependency will return an error. Parameters ---------- graph : :py:obj:`~.CUgraph` or :py:obj:`~.cudaGraph_t` Graph from which to remove dependencies from : list[:py:obj:`~.cudaGraphNode_t`] Array of nodes that provide the dependencies to : list[:py:obj:`~.cudaGraphNode_t`] Array of dependent nodes numDependencies : size_t Number of dependencies to be removed Returns ------- cudaError_t :py:obj:`~.cudaSuccess`, :py:obj:`~.cudaErrorInvalidValue` See Also -------- :py:obj:`~.cudaGraphAddDependencies`, :py:obj:`~.cudaGraphGetEdges`, :py:obj:`~.cudaGraphNodeGetDependencies`, :py:obj:`~.cudaGraphNodeGetDependentNodes` """ to = [] if to is None else to if not all(isinstance(_x, (cudaGraphNode_t,driver.CUgraphNode)) for _x in to): raise TypeError("Argument 'to' is not instance of type (expected tuple[cyruntime.cudaGraphNode_t,driver.CUgraphNode] or list[cyruntime.cudaGraphNode_t,driver.CUgraphNode]") from_ = [] if from_ is None else from_ if not all(isinstance(_x, (cudaGraphNode_t,driver.CUgraphNode)) for _x in from_): raise TypeError("Argument 'from_' is not instance of type (expected tuple[cyruntime.cudaGraphNode_t,driver.CUgraphNode] or list[cyruntime.cudaGraphNode_t,driver.CUgraphNode]") cdef cyruntime.cudaGraph_t cygraph if graph is None: pgraph = 0 elif isinstance(graph, (cudaGraph_t,driver.CUgraph)): pgraph = int(graph) else: pgraph = int(cudaGraph_t(graph)) cygraph = pgraph cdef cyruntime.cudaGraphNode_t* cyfrom_ = NULL if len(from_) > 1: cyfrom_ = calloc(len(from_), sizeof(cyruntime.cudaGraphNode_t)) if cyfrom_ is NULL: raise MemoryError('Failed to allocate length x size memory: ' + str(len(from_)) + 'x' + str(sizeof(cyruntime.cudaGraphNode_t))) else: for idx in range(len(from_)): cyfrom_[idx] = (from_[idx])._pvt_ptr[0] elif len(from_) == 1: cyfrom_ = (from_[0])._pvt_ptr cdef cyruntime.cudaGraphNode_t* cyto = NULL if len(to) > 1: cyto = calloc(len(to), sizeof(cyruntime.cudaGraphNode_t)) if cyto is NULL: raise MemoryError('Failed to allocate length x size memory: ' + str(len(to)) + 'x' + str(sizeof(cyruntime.cudaGraphNode_t))) else: for idx in range(len(to)): cyto[idx] = (to[idx])._pvt_ptr[0] elif len(to) == 1: cyto = (to[0])._pvt_ptr if numDependencies > len(from_): raise RuntimeError("List is too small: " + str(len(from_)) + " < " + str(numDependencies)) if numDependencies > len(to): raise RuntimeError("List is too small: " + str(len(to)) + " < " + str(numDependencies)) with nogil: err = cyruntime.cudaGraphRemoveDependencies(cygraph, cyfrom_, cyto, numDependencies) if len(from_) > 1 and cyfrom_ is not NULL: free(cyfrom_) if len(to) > 1 and cyto is not NULL: free(cyto) return (_dict_cudaError_t[err],) @cython.embedsignature(True) def cudaGraphRemoveDependencies_v2(graph, from_ : Optional[tuple[cudaGraphNode_t] | list[cudaGraphNode_t]], to : Optional[tuple[cudaGraphNode_t] | list[cudaGraphNode_t]], edgeData : Optional[tuple[cudaGraphEdgeData] | list[cudaGraphEdgeData]], size_t numDependencies): """ Removes dependency edges from a graph. (12.3+) The number of `pDependencies` to be removed is defined by `numDependencies`. Elements in `pFrom` and `pTo` at corresponding indices define a dependency. Each node in `pFrom` and `pTo` must belong to `graph`. If `numDependencies` is 0, elements in `pFrom` and `pTo` will be ignored. Specifying an edge that does not exist in the graph, with data matching `edgeData`, results in an error. `edgeData` is nullable, which is equivalent to passing default (zeroed) data for each edge. Parameters ---------- graph : :py:obj:`~.CUgraph` or :py:obj:`~.cudaGraph_t` Graph from which to remove dependencies from : list[:py:obj:`~.cudaGraphNode_t`] Array of nodes that provide the dependencies to : list[:py:obj:`~.cudaGraphNode_t`] Array of dependent nodes edgeData : list[:py:obj:`~.cudaGraphEdgeData`] Optional array of edge data. If NULL, edge data is assumed to be default (zeroed). numDependencies : size_t Number of dependencies to be removed Returns ------- cudaError_t :py:obj:`~.cudaSuccess`, :py:obj:`~.cudaErrorInvalidValue` See Also -------- :py:obj:`~.cudaGraphAddDependencies`, :py:obj:`~.cudaGraphGetEdges`, :py:obj:`~.cudaGraphNodeGetDependencies`, :py:obj:`~.cudaGraphNodeGetDependentNodes` """ edgeData = [] if edgeData is None else edgeData if not all(isinstance(_x, (cudaGraphEdgeData,)) for _x in edgeData): raise TypeError("Argument 'edgeData' is not instance of type (expected tuple[cyruntime.cudaGraphEdgeData,] or list[cyruntime.cudaGraphEdgeData,]") to = [] if to is None else to if not all(isinstance(_x, (cudaGraphNode_t,driver.CUgraphNode)) for _x in to): raise TypeError("Argument 'to' is not instance of type (expected tuple[cyruntime.cudaGraphNode_t,driver.CUgraphNode] or list[cyruntime.cudaGraphNode_t,driver.CUgraphNode]") from_ = [] if from_ is None else from_ if not all(isinstance(_x, (cudaGraphNode_t,driver.CUgraphNode)) for _x in from_): raise TypeError("Argument 'from_' is not instance of type (expected tuple[cyruntime.cudaGraphNode_t,driver.CUgraphNode] or list[cyruntime.cudaGraphNode_t,driver.CUgraphNode]") cdef cyruntime.cudaGraph_t cygraph if graph is None: pgraph = 0 elif isinstance(graph, (cudaGraph_t,driver.CUgraph)): pgraph = int(graph) else: pgraph = int(cudaGraph_t(graph)) cygraph = pgraph cdef cyruntime.cudaGraphNode_t* cyfrom_ = NULL if len(from_) > 1: cyfrom_ = calloc(len(from_), sizeof(cyruntime.cudaGraphNode_t)) if cyfrom_ is NULL: raise MemoryError('Failed to allocate length x size memory: ' + str(len(from_)) + 'x' + str(sizeof(cyruntime.cudaGraphNode_t))) else: for idx in range(len(from_)): cyfrom_[idx] = (from_[idx])._pvt_ptr[0] elif len(from_) == 1: cyfrom_ = (from_[0])._pvt_ptr cdef cyruntime.cudaGraphNode_t* cyto = NULL if len(to) > 1: cyto = calloc(len(to), sizeof(cyruntime.cudaGraphNode_t)) if cyto is NULL: raise MemoryError('Failed to allocate length x size memory: ' + str(len(to)) + 'x' + str(sizeof(cyruntime.cudaGraphNode_t))) else: for idx in range(len(to)): cyto[idx] = (to[idx])._pvt_ptr[0] elif len(to) == 1: cyto = (to[0])._pvt_ptr cdef cyruntime.cudaGraphEdgeData* cyedgeData = NULL if len(edgeData) > 1: cyedgeData = calloc(len(edgeData), sizeof(cyruntime.cudaGraphEdgeData)) if cyedgeData is NULL: raise MemoryError('Failed to allocate length x size memory: ' + str(len(edgeData)) + 'x' + str(sizeof(cyruntime.cudaGraphEdgeData))) for idx in range(len(edgeData)): string.memcpy(&cyedgeData[idx], (edgeData[idx])._pvt_ptr, sizeof(cyruntime.cudaGraphEdgeData)) elif len(edgeData) == 1: cyedgeData = (edgeData[0])._pvt_ptr with nogil: err = cyruntime.cudaGraphRemoveDependencies_v2(cygraph, cyfrom_, cyto, cyedgeData, numDependencies) if len(from_) > 1 and cyfrom_ is not NULL: free(cyfrom_) if len(to) > 1 and cyto is not NULL: free(cyto) if len(edgeData) > 1 and cyedgeData is not NULL: free(cyedgeData) return (_dict_cudaError_t[err],) @cython.embedsignature(True) def cudaGraphDestroyNode(node): """ Remove a node from the graph. Removes `node` from its graph. This operation also severs any dependencies of other nodes on `node` and vice versa. Dependencies cannot be removed from graphs which contain allocation or free nodes. Any attempt to do so will return an error. Parameters ---------- node : :py:obj:`~.CUgraphNode` or :py:obj:`~.cudaGraphNode_t` Node to remove Returns ------- cudaError_t :py:obj:`~.cudaSuccess`, :py:obj:`~.cudaErrorInvalidValue` See Also -------- :py:obj:`~.cudaGraphAddChildGraphNode`, :py:obj:`~.cudaGraphAddEmptyNode`, :py:obj:`~.cudaGraphAddKernelNode`, :py:obj:`~.cudaGraphAddHostNode`, :py:obj:`~.cudaGraphAddMemcpyNode`, :py:obj:`~.cudaGraphAddMemsetNode` """ cdef cyruntime.cudaGraphNode_t cynode if node is None: pnode = 0 elif isinstance(node, (cudaGraphNode_t,driver.CUgraphNode)): pnode = int(node) else: pnode = int(cudaGraphNode_t(node)) cynode = pnode with nogil: err = cyruntime.cudaGraphDestroyNode(cynode) return (_dict_cudaError_t[err],) @cython.embedsignature(True) def cudaGraphInstantiate(graph, unsigned long long flags): """ Creates an executable graph from a graph. Instantiates `graph` as an executable graph. The graph is validated for any structural constraints or intra-node constraints which were not previously validated. If instantiation is successful, a handle to the instantiated graph is returned in `pGraphExec`. The `flags` parameter controls the behavior of instantiation and subsequent graph launches. Valid flags are: - :py:obj:`~.cudaGraphInstantiateFlagAutoFreeOnLaunch`, which configures a graph containing memory allocation nodes to automatically free any unfreed memory allocations before the graph is relaunched. - :py:obj:`~.cudaGraphInstantiateFlagDeviceLaunch`, which configures the graph for launch from the device. If this flag is passed, the executable graph handle returned can be used to launch the graph from both the host and device. This flag cannot be used in conjunction with :py:obj:`~.cudaGraphInstantiateFlagAutoFreeOnLaunch`. - :py:obj:`~.cudaGraphInstantiateFlagUseNodePriority`, which causes the graph to use the priorities from the per-node attributes rather than the priority of the launch stream during execution. Note that priorities are only available on kernel nodes, and are copied from stream priority during stream capture. If `graph` contains any allocation or free nodes, there can be at most one executable graph in existence for that graph at a time. An attempt to instantiate a second executable graph before destroying the first with :py:obj:`~.cudaGraphExecDestroy` will result in an error. The same also applies if `graph` contains any device-updatable kernel nodes. Graphs instantiated for launch on the device have additional restrictions which do not apply to host graphs: - The graph's nodes must reside on a single device. - The graph can only contain kernel nodes, memcpy nodes, memset nodes, and child graph nodes. - The graph cannot be empty and must contain at least one kernel, memcpy, or memset node. Operation-specific restrictions are outlined below. - Kernel nodes: - Use of CUDA Dynamic Parallelism is not permitted. - Cooperative launches are permitted as long as MPS is not in use. - Memcpy nodes: - Only copies involving device memory and/or pinned device-mapped host memory are permitted. - Copies involving CUDA arrays are not permitted. - Both operands must be accessible from the current device, and the current device must match the device of other nodes in the graph. If `graph` is not instantiated for launch on the device but contains kernels which call device-side :py:obj:`~.cudaGraphLaunch()` from multiple devices, this will result in an error. Parameters ---------- graph : :py:obj:`~.CUgraph` or :py:obj:`~.cudaGraph_t` Graph to instantiate flags : unsigned long long Flags to control instantiation. See :py:obj:`~.CUgraphInstantiate_flags`. Returns ------- cudaError_t :py:obj:`~.cudaSuccess`, :py:obj:`~.cudaErrorInvalidValue` pGraphExec : :py:obj:`~.cudaGraphExec_t` Returns instantiated graph See Also -------- :py:obj:`~.cudaGraphInstantiateWithFlags`, :py:obj:`~.cudaGraphCreate`, :py:obj:`~.cudaGraphUpload`, :py:obj:`~.cudaGraphLaunch`, :py:obj:`~.cudaGraphExecDestroy` """ cdef cyruntime.cudaGraph_t cygraph if graph is None: pgraph = 0 elif isinstance(graph, (cudaGraph_t,driver.CUgraph)): pgraph = int(graph) else: pgraph = int(cudaGraph_t(graph)) cygraph = pgraph cdef cudaGraphExec_t pGraphExec = cudaGraphExec_t() with nogil: err = cyruntime.cudaGraphInstantiate(pGraphExec._pvt_ptr, cygraph, flags) if err != cyruntime.cudaSuccess: return (_dict_cudaError_t[err], None) return (_dict_cudaError_t[err], pGraphExec) @cython.embedsignature(True) def cudaGraphInstantiateWithFlags(graph, unsigned long long flags): """ Creates an executable graph from a graph. Instantiates `graph` as an executable graph. The graph is validated for any structural constraints or intra-node constraints which were not previously validated. If instantiation is successful, a handle to the instantiated graph is returned in `pGraphExec`. The `flags` parameter controls the behavior of instantiation and subsequent graph launches. Valid flags are: - :py:obj:`~.cudaGraphInstantiateFlagAutoFreeOnLaunch`, which configures a graph containing memory allocation nodes to automatically free any unfreed memory allocations before the graph is relaunched. - :py:obj:`~.cudaGraphInstantiateFlagDeviceLaunch`, which configures the graph for launch from the device. If this flag is passed, the executable graph handle returned can be used to launch the graph from both the host and device. This flag can only be used on platforms which support unified addressing. This flag cannot be used in conjunction with :py:obj:`~.cudaGraphInstantiateFlagAutoFreeOnLaunch`. - :py:obj:`~.cudaGraphInstantiateFlagUseNodePriority`, which causes the graph to use the priorities from the per-node attributes rather than the priority of the launch stream during execution. Note that priorities are only available on kernel nodes, and are copied from stream priority during stream capture. If `graph` contains any allocation or free nodes, there can be at most one executable graph in existence for that graph at a time. An attempt to instantiate a second executable graph before destroying the first with :py:obj:`~.cudaGraphExecDestroy` will result in an error. The same also applies if `graph` contains any device-updatable kernel nodes. If `graph` contains kernels which call device-side :py:obj:`~.cudaGraphLaunch()` from multiple devices, this will result in an error. Graphs instantiated for launch on the device have additional restrictions which do not apply to host graphs: - The graph's nodes must reside on a single device. - The graph can only contain kernel nodes, memcpy nodes, memset nodes, and child graph nodes. - The graph cannot be empty and must contain at least one kernel, memcpy, or memset node. Operation-specific restrictions are outlined below. - Kernel nodes: - Use of CUDA Dynamic Parallelism is not permitted. - Cooperative launches are permitted as long as MPS is not in use. - Memcpy nodes: - Only copies involving device memory and/or pinned device-mapped host memory are permitted. - Copies involving CUDA arrays are not permitted. - Both operands must be accessible from the current device, and the current device must match the device of other nodes in the graph. Parameters ---------- graph : :py:obj:`~.CUgraph` or :py:obj:`~.cudaGraph_t` Graph to instantiate flags : unsigned long long Flags to control instantiation. See :py:obj:`~.CUgraphInstantiate_flags`. Returns ------- cudaError_t :py:obj:`~.cudaSuccess`, :py:obj:`~.cudaErrorInvalidValue` pGraphExec : :py:obj:`~.cudaGraphExec_t` Returns instantiated graph See Also -------- :py:obj:`~.cudaGraphInstantiate`, :py:obj:`~.cudaGraphCreate`, :py:obj:`~.cudaGraphUpload`, :py:obj:`~.cudaGraphLaunch`, :py:obj:`~.cudaGraphExecDestroy` """ cdef cyruntime.cudaGraph_t cygraph if graph is None: pgraph = 0 elif isinstance(graph, (cudaGraph_t,driver.CUgraph)): pgraph = int(graph) else: pgraph = int(cudaGraph_t(graph)) cygraph = pgraph cdef cudaGraphExec_t pGraphExec = cudaGraphExec_t() with nogil: err = cyruntime.cudaGraphInstantiateWithFlags(pGraphExec._pvt_ptr, cygraph, flags) if err != cyruntime.cudaSuccess: return (_dict_cudaError_t[err], None) return (_dict_cudaError_t[err], pGraphExec) @cython.embedsignature(True) def cudaGraphInstantiateWithParams(graph, instantiateParams : Optional[cudaGraphInstantiateParams]): """ Creates an executable graph from a graph. Instantiates `graph` as an executable graph according to the `instantiateParams` structure. The graph is validated for any structural constraints or intra-node constraints which were not previously validated. If instantiation is successful, a handle to the instantiated graph is returned in `pGraphExec`. `instantiateParams` controls the behavior of instantiation and subsequent graph launches, as well as returning more detailed information in the event of an error. :py:obj:`~.cudaGraphInstantiateParams` is defined as: **View CUDA Toolkit Documentation for a C++ code example** The `flags` field controls the behavior of instantiation and subsequent graph launches. Valid flags are: - :py:obj:`~.cudaGraphInstantiateFlagAutoFreeOnLaunch`, which configures a graph containing memory allocation nodes to automatically free any unfreed memory allocations before the graph is relaunched. - :py:obj:`~.cudaGraphInstantiateFlagUpload`, which will perform an upload of the graph into `uploadStream` once the graph has been instantiated. - :py:obj:`~.cudaGraphInstantiateFlagDeviceLaunch`, which configures the graph for launch from the device. If this flag is passed, the executable graph handle returned can be used to launch the graph from both the host and device. This flag can only be used on platforms which support unified addressing. This flag cannot be used in conjunction with :py:obj:`~.cudaGraphInstantiateFlagAutoFreeOnLaunch`. - :py:obj:`~.cudaGraphInstantiateFlagUseNodePriority`, which causes the graph to use the priorities from the per-node attributes rather than the priority of the launch stream during execution. Note that priorities are only available on kernel nodes, and are copied from stream priority during stream capture. If `graph` contains any allocation or free nodes, there can be at most one executable graph in existence for that graph at a time. An attempt to instantiate a second executable graph before destroying the first with :py:obj:`~.cudaGraphExecDestroy` will result in an error. The same also applies if `graph` contains any device-updatable kernel nodes. If `graph` contains kernels which call device-side :py:obj:`~.cudaGraphLaunch()` from multiple devices, this will result in an error. Graphs instantiated for launch on the device have additional restrictions which do not apply to host graphs: - The graph's nodes must reside on a single device. - The graph can only contain kernel nodes, memcpy nodes, memset nodes, and child graph nodes. - The graph cannot be empty and must contain at least one kernel, memcpy, or memset node. Operation-specific restrictions are outlined below. - Kernel nodes: - Use of CUDA Dynamic Parallelism is not permitted. - Cooperative launches are permitted as long as MPS is not in use. - Memcpy nodes: - Only copies involving device memory and/or pinned device-mapped host memory are permitted. - Copies involving CUDA arrays are not permitted. - Both operands must be accessible from the current device, and the current device must match the device of other nodes in the graph. In the event of an error, the `result_out` and `errNode_out` fields will contain more information about the nature of the error. Possible error reporting includes: - :py:obj:`~.cudaGraphInstantiateError`, if passed an invalid value or if an unexpected error occurred which is described by the return value of the function. `errNode_out` will be set to NULL. - :py:obj:`~.cudaGraphInstantiateInvalidStructure`, if the graph structure is invalid. `errNode_out` will be set to one of the offending nodes. - :py:obj:`~.cudaGraphInstantiateNodeOperationNotSupported`, if the graph is instantiated for device launch but contains a node of an unsupported node type, or a node which performs unsupported operations, such as use of CUDA dynamic parallelism within a kernel node. `errNode_out` will be set to this node. - :py:obj:`~.cudaGraphInstantiateMultipleDevicesNotSupported`, if the graph is instantiated for device launch but a node’s device differs from that of another node. This error can also be returned if a graph is not instantiated for device launch and it contains kernels which call device-side :py:obj:`~.cudaGraphLaunch()` from multiple devices. `errNode_out` will be set to this node. If instantiation is successful, `result_out` will be set to :py:obj:`~.cudaGraphInstantiateSuccess`, and `hErrNode_out` will be set to NULL. Parameters ---------- graph : :py:obj:`~.CUgraph` or :py:obj:`~.cudaGraph_t` Graph to instantiate instantiateParams : :py:obj:`~.cudaGraphInstantiateParams` Instantiation parameters Returns ------- cudaError_t :py:obj:`~.cudaSuccess`, :py:obj:`~.cudaErrorInvalidValue` pGraphExec : :py:obj:`~.cudaGraphExec_t` Returns instantiated graph See Also -------- :py:obj:`~.cudaGraphCreate`, :py:obj:`~.cudaGraphInstantiate`, :py:obj:`~.cudaGraphInstantiateWithFlags`, :py:obj:`~.cudaGraphExecDestroy` """ cdef cyruntime.cudaGraph_t cygraph if graph is None: pgraph = 0 elif isinstance(graph, (cudaGraph_t,driver.CUgraph)): pgraph = int(graph) else: pgraph = int(cudaGraph_t(graph)) cygraph = pgraph cdef cudaGraphExec_t pGraphExec = cudaGraphExec_t() cdef cyruntime.cudaGraphInstantiateParams* cyinstantiateParams_ptr = instantiateParams._pvt_ptr if instantiateParams is not None else NULL with nogil: err = cyruntime.cudaGraphInstantiateWithParams(pGraphExec._pvt_ptr, cygraph, cyinstantiateParams_ptr) if err != cyruntime.cudaSuccess: return (_dict_cudaError_t[err], None) return (_dict_cudaError_t[err], pGraphExec) @cython.embedsignature(True) def cudaGraphExecGetFlags(graphExec): """ Query the instantiation flags of an executable graph. Returns the flags that were passed to instantiation for the given executable graph. :py:obj:`~.cudaGraphInstantiateFlagUpload` will not be returned by this API as it does not affect the resulting executable graph. Parameters ---------- graphExec : :py:obj:`~.CUgraphExec` or :py:obj:`~.cudaGraphExec_t` The executable graph to query Returns ------- cudaError_t :py:obj:`~.cudaSuccess`, :py:obj:`~.cudaErrorInvalidValue` flags : unsigned long long Returns the instantiation flags See Also -------- :py:obj:`~.cudaGraphInstantiate`, :py:obj:`~.cudaGraphInstantiateWithFlags`, :py:obj:`~.cudaGraphInstantiateWithParams` """ cdef cyruntime.cudaGraphExec_t cygraphExec if graphExec is None: pgraphExec = 0 elif isinstance(graphExec, (cudaGraphExec_t,driver.CUgraphExec)): pgraphExec = int(graphExec) else: pgraphExec = int(cudaGraphExec_t(graphExec)) cygraphExec = pgraphExec cdef unsigned long long flags = 0 with nogil: err = cyruntime.cudaGraphExecGetFlags(cygraphExec, &flags) if err != cyruntime.cudaSuccess: return (_dict_cudaError_t[err], None) return (_dict_cudaError_t[err], flags) @cython.embedsignature(True) def cudaGraphExecKernelNodeSetParams(hGraphExec, node, pNodeParams : Optional[cudaKernelNodeParams]): """ Sets the parameters for a kernel node in the given graphExec. Sets the parameters of a kernel node in an executable graph `hGraphExec`. The node is identified by the corresponding node `node` in the non-executable graph, from which the executable graph was instantiated. `node` must not have been removed from the original graph. All `nodeParams` fields may change, but the following restrictions apply to `func` updates: - The owning device of the function cannot change. - A node whose function originally did not use CUDA dynamic parallelism cannot be updated to a function which uses CDP - A node whose function originally did not make device-side update calls cannot be updated to a function which makes device-side update calls. - If `hGraphExec` was not instantiated for device launch, a node whose function originally did not use device-side :py:obj:`~.cudaGraphLaunch()` cannot be updated to a function which uses device-side :py:obj:`~.cudaGraphLaunch()` unless the node resides on the same device as nodes which contained such calls at instantiate-time. If no such calls were present at instantiation, these updates cannot be performed at all. The modifications only affect future launches of `hGraphExec`. Already enqueued or running launches of `hGraphExec` are not affected by this call. `node` is also not modified by this call. If `node` is a device-updatable kernel node, the next upload/launch of `hGraphExec` will overwrite any previous device-side updates. Additionally, applying host updates to a device-updatable kernel node while it is being updated from the device will result in undefined behavior. Parameters ---------- hGraphExec : :py:obj:`~.CUgraphExec` or :py:obj:`~.cudaGraphExec_t` The executable graph in which to set the specified node node : :py:obj:`~.CUgraphNode` or :py:obj:`~.cudaGraphNode_t` kernel node from the graph from which graphExec was instantiated pNodeParams : :py:obj:`~.cudaKernelNodeParams` Updated Parameters to set Returns ------- cudaError_t :py:obj:`~.cudaSuccess`, :py:obj:`~.cudaErrorInvalidValue`, See Also -------- :py:obj:`~.cudaGraphExecNodeSetParams`, :py:obj:`~.cudaGraphAddKernelNode`, :py:obj:`~.cudaGraphKernelNodeSetParams`, :py:obj:`~.cudaGraphExecMemcpyNodeSetParams`, :py:obj:`~.cudaGraphExecMemsetNodeSetParams`, :py:obj:`~.cudaGraphExecHostNodeSetParams`, :py:obj:`~.cudaGraphExecChildGraphNodeSetParams`, :py:obj:`~.cudaGraphExecEventRecordNodeSetEvent`, :py:obj:`~.cudaGraphExecEventWaitNodeSetEvent`, :py:obj:`~.cudaGraphExecExternalSemaphoresSignalNodeSetParams`, :py:obj:`~.cudaGraphExecExternalSemaphoresWaitNodeSetParams`, :py:obj:`~.cudaGraphExecUpdate`, :py:obj:`~.cudaGraphInstantiate` """ cdef cyruntime.cudaGraphNode_t cynode if node is None: pnode = 0 elif isinstance(node, (cudaGraphNode_t,driver.CUgraphNode)): pnode = int(node) else: pnode = int(cudaGraphNode_t(node)) cynode = pnode cdef cyruntime.cudaGraphExec_t cyhGraphExec if hGraphExec is None: phGraphExec = 0 elif isinstance(hGraphExec, (cudaGraphExec_t,driver.CUgraphExec)): phGraphExec = int(hGraphExec) else: phGraphExec = int(cudaGraphExec_t(hGraphExec)) cyhGraphExec = phGraphExec cdef cyruntime.cudaKernelNodeParams* cypNodeParams_ptr = pNodeParams._pvt_ptr if pNodeParams is not None else NULL with nogil: err = cyruntime.cudaGraphExecKernelNodeSetParams(cyhGraphExec, cynode, cypNodeParams_ptr) return (_dict_cudaError_t[err],) @cython.embedsignature(True) def cudaGraphExecMemcpyNodeSetParams(hGraphExec, node, pNodeParams : Optional[cudaMemcpy3DParms]): """ Sets the parameters for a memcpy node in the given graphExec. Updates the work represented by `node` in `hGraphExec` as though `node` had contained `pNodeParams` at instantiation. `node` must remain in the graph which was used to instantiate `hGraphExec`. Changed edges to and from `node` are ignored. The source and destination memory in `pNodeParams` must be allocated from the same contexts as the original source and destination memory. Both the instantiation-time memory operands and the memory operands in `pNodeParams` must be 1-dimensional. Zero-length operations are not supported. The modifications only affect future launches of `hGraphExec`. Already enqueued or running launches of `hGraphExec` are not affected by this call. `node` is also not modified by this call. Returns :py:obj:`~.cudaErrorInvalidValue` if the memory operands' mappings changed or either the original or new memory operands are multidimensional. Parameters ---------- hGraphExec : :py:obj:`~.CUgraphExec` or :py:obj:`~.cudaGraphExec_t` The executable graph in which to set the specified node node : :py:obj:`~.CUgraphNode` or :py:obj:`~.cudaGraphNode_t` Memcpy node from the graph which was used to instantiate graphExec pNodeParams : :py:obj:`~.cudaMemcpy3DParms` Updated Parameters to set Returns ------- cudaError_t :py:obj:`~.cudaSuccess`, :py:obj:`~.cudaErrorInvalidValue`, See Also -------- :py:obj:`~.cudaGraphExecNodeSetParams`, :py:obj:`~.cudaGraphAddMemcpyNode`, :py:obj:`~.cudaGraphMemcpyNodeSetParams`, :py:obj:`~.cudaGraphExecMemcpyNodeSetParamsToSymbol`, :py:obj:`~.cudaGraphExecMemcpyNodeSetParamsFromSymbol`, :py:obj:`~.cudaGraphExecMemcpyNodeSetParams1D`, :py:obj:`~.cudaGraphExecKernelNodeSetParams`, :py:obj:`~.cudaGraphExecMemsetNodeSetParams`, :py:obj:`~.cudaGraphExecHostNodeSetParams`, :py:obj:`~.cudaGraphExecChildGraphNodeSetParams`, :py:obj:`~.cudaGraphExecEventRecordNodeSetEvent`, :py:obj:`~.cudaGraphExecEventWaitNodeSetEvent`, :py:obj:`~.cudaGraphExecExternalSemaphoresSignalNodeSetParams`, :py:obj:`~.cudaGraphExecExternalSemaphoresWaitNodeSetParams`, :py:obj:`~.cudaGraphExecUpdate`, :py:obj:`~.cudaGraphInstantiate` """ cdef cyruntime.cudaGraphNode_t cynode if node is None: pnode = 0 elif isinstance(node, (cudaGraphNode_t,driver.CUgraphNode)): pnode = int(node) else: pnode = int(cudaGraphNode_t(node)) cynode = pnode cdef cyruntime.cudaGraphExec_t cyhGraphExec if hGraphExec is None: phGraphExec = 0 elif isinstance(hGraphExec, (cudaGraphExec_t,driver.CUgraphExec)): phGraphExec = int(hGraphExec) else: phGraphExec = int(cudaGraphExec_t(hGraphExec)) cyhGraphExec = phGraphExec cdef cyruntime.cudaMemcpy3DParms* cypNodeParams_ptr = pNodeParams._pvt_ptr if pNodeParams is not None else NULL with nogil: err = cyruntime.cudaGraphExecMemcpyNodeSetParams(cyhGraphExec, cynode, cypNodeParams_ptr) return (_dict_cudaError_t[err],) @cython.embedsignature(True) def cudaGraphExecMemcpyNodeSetParams1D(hGraphExec, node, dst, src, size_t count, kind not None : cudaMemcpyKind): """ Sets the parameters for a memcpy node in the given graphExec to perform a 1-dimensional copy. Updates the work represented by `node` in `hGraphExec` as though `node` had contained the given params at instantiation. `node` must remain in the graph which was used to instantiate `hGraphExec`. Changed edges to and from `node` are ignored. `src` and `dst` must be allocated from the same contexts as the original source and destination memory. The instantiation-time memory operands must be 1-dimensional. Zero-length operations are not supported. The modifications only affect future launches of `hGraphExec`. Already enqueued or running launches of `hGraphExec` are not affected by this call. `node` is also not modified by this call. Returns :py:obj:`~.cudaErrorInvalidValue` if the memory operands' mappings changed or the original memory operands are multidimensional. Parameters ---------- hGraphExec : :py:obj:`~.CUgraphExec` or :py:obj:`~.cudaGraphExec_t` The executable graph in which to set the specified node node : :py:obj:`~.CUgraphNode` or :py:obj:`~.cudaGraphNode_t` Memcpy node from the graph which was used to instantiate graphExec dst : Any Destination memory address src : Any Source memory address count : size_t Size in bytes to copy kind : :py:obj:`~.cudaMemcpyKind` Type of transfer Returns ------- cudaError_t :py:obj:`~.cudaSuccess`, :py:obj:`~.cudaErrorInvalidValue` See Also -------- :py:obj:`~.cudaGraphAddMemcpyNode`, :py:obj:`~.cudaGraphAddMemcpyNode1D`, :py:obj:`~.cudaGraphMemcpyNodeSetParams`, :py:obj:`~.cudaGraphMemcpyNodeSetParams1D`, :py:obj:`~.cudaGraphExecMemcpyNodeSetParams`, :py:obj:`~.cudaGraphExecKernelNodeSetParams`, :py:obj:`~.cudaGraphExecMemsetNodeSetParams`, :py:obj:`~.cudaGraphExecHostNodeSetParams`, :py:obj:`~.cudaGraphExecChildGraphNodeSetParams`, :py:obj:`~.cudaGraphExecEventRecordNodeSetEvent`, :py:obj:`~.cudaGraphExecEventWaitNodeSetEvent`, :py:obj:`~.cudaGraphExecExternalSemaphoresSignalNodeSetParams`, :py:obj:`~.cudaGraphExecExternalSemaphoresWaitNodeSetParams`, :py:obj:`~.cudaGraphExecUpdate`, :py:obj:`~.cudaGraphInstantiate` """ cdef cyruntime.cudaGraphNode_t cynode if node is None: pnode = 0 elif isinstance(node, (cudaGraphNode_t,driver.CUgraphNode)): pnode = int(node) else: pnode = int(cudaGraphNode_t(node)) cynode = pnode cdef cyruntime.cudaGraphExec_t cyhGraphExec if hGraphExec is None: phGraphExec = 0 elif isinstance(hGraphExec, (cudaGraphExec_t,driver.CUgraphExec)): phGraphExec = int(hGraphExec) else: phGraphExec = int(cudaGraphExec_t(hGraphExec)) cyhGraphExec = phGraphExec cydst = _HelperInputVoidPtr(dst) cdef void* cydst_ptr = cydst.cptr cysrc = _HelperInputVoidPtr(src) cdef void* cysrc_ptr = cysrc.cptr cdef cyruntime.cudaMemcpyKind cykind = kind.value with nogil: err = cyruntime.cudaGraphExecMemcpyNodeSetParams1D(cyhGraphExec, cynode, cydst_ptr, cysrc_ptr, count, cykind) return (_dict_cudaError_t[err],) @cython.embedsignature(True) def cudaGraphExecMemsetNodeSetParams(hGraphExec, node, pNodeParams : Optional[cudaMemsetParams]): """ Sets the parameters for a memset node in the given graphExec. Updates the work represented by `node` in `hGraphExec` as though `node` had contained `pNodeParams` at instantiation. `node` must remain in the graph which was used to instantiate `hGraphExec`. Changed edges to and from `node` are ignored. Zero sized operations are not supported. The new destination pointer in `pNodeParams` must be to the same kind of allocation as the original destination pointer and have the same context association and device mapping as the original destination pointer. Both the value and pointer address may be updated. Changing other aspects of the memset (width, height, element size or pitch) may cause the update to be rejected. Specifically, for 2d memsets, all dimension changes are rejected. For 1d memsets, changes in height are explicitly rejected and other changes are opportunistically allowed if the resulting work maps onto the work resources already allocated for the node. The modifications only affect future launches of `hGraphExec`. Already enqueued or running launches of `hGraphExec` are not affected by this call. `node` is also not modified by this call. Parameters ---------- hGraphExec : :py:obj:`~.CUgraphExec` or :py:obj:`~.cudaGraphExec_t` The executable graph in which to set the specified node node : :py:obj:`~.CUgraphNode` or :py:obj:`~.cudaGraphNode_t` Memset node from the graph which was used to instantiate graphExec pNodeParams : :py:obj:`~.cudaMemsetParams` Updated Parameters to set Returns ------- cudaError_t :py:obj:`~.cudaSuccess`, :py:obj:`~.cudaErrorInvalidValue`, See Also -------- :py:obj:`~.cudaGraphExecNodeSetParams`, :py:obj:`~.cudaGraphAddMemsetNode`, :py:obj:`~.cudaGraphMemsetNodeSetParams`, :py:obj:`~.cudaGraphExecKernelNodeSetParams`, :py:obj:`~.cudaGraphExecMemcpyNodeSetParams`, :py:obj:`~.cudaGraphExecHostNodeSetParams`, :py:obj:`~.cudaGraphExecChildGraphNodeSetParams`, :py:obj:`~.cudaGraphExecEventRecordNodeSetEvent`, :py:obj:`~.cudaGraphExecEventWaitNodeSetEvent`, :py:obj:`~.cudaGraphExecExternalSemaphoresSignalNodeSetParams`, :py:obj:`~.cudaGraphExecExternalSemaphoresWaitNodeSetParams`, :py:obj:`~.cudaGraphExecUpdate`, :py:obj:`~.cudaGraphInstantiate` """ cdef cyruntime.cudaGraphNode_t cynode if node is None: pnode = 0 elif isinstance(node, (cudaGraphNode_t,driver.CUgraphNode)): pnode = int(node) else: pnode = int(cudaGraphNode_t(node)) cynode = pnode cdef cyruntime.cudaGraphExec_t cyhGraphExec if hGraphExec is None: phGraphExec = 0 elif isinstance(hGraphExec, (cudaGraphExec_t,driver.CUgraphExec)): phGraphExec = int(hGraphExec) else: phGraphExec = int(cudaGraphExec_t(hGraphExec)) cyhGraphExec = phGraphExec cdef cyruntime.cudaMemsetParams* cypNodeParams_ptr = pNodeParams._pvt_ptr if pNodeParams is not None else NULL with nogil: err = cyruntime.cudaGraphExecMemsetNodeSetParams(cyhGraphExec, cynode, cypNodeParams_ptr) return (_dict_cudaError_t[err],) @cython.embedsignature(True) def cudaGraphExecHostNodeSetParams(hGraphExec, node, pNodeParams : Optional[cudaHostNodeParams]): """ Sets the parameters for a host node in the given graphExec. Updates the work represented by `node` in `hGraphExec` as though `node` had contained `pNodeParams` at instantiation. `node` must remain in the graph which was used to instantiate `hGraphExec`. Changed edges to and from `node` are ignored. The modifications only affect future launches of `hGraphExec`. Already enqueued or running launches of `hGraphExec` are not affected by this call. `node` is also not modified by this call. Parameters ---------- hGraphExec : :py:obj:`~.CUgraphExec` or :py:obj:`~.cudaGraphExec_t` The executable graph in which to set the specified node node : :py:obj:`~.CUgraphNode` or :py:obj:`~.cudaGraphNode_t` Host node from the graph which was used to instantiate graphExec pNodeParams : :py:obj:`~.cudaHostNodeParams` Updated Parameters to set Returns ------- cudaError_t :py:obj:`~.cudaSuccess`, :py:obj:`~.cudaErrorInvalidValue`, See Also -------- :py:obj:`~.cudaGraphExecNodeSetParams`, :py:obj:`~.cudaGraphAddHostNode`, :py:obj:`~.cudaGraphHostNodeSetParams`, :py:obj:`~.cudaGraphExecKernelNodeSetParams`, :py:obj:`~.cudaGraphExecMemcpyNodeSetParams`, :py:obj:`~.cudaGraphExecMemsetNodeSetParams`, :py:obj:`~.cudaGraphExecChildGraphNodeSetParams`, :py:obj:`~.cudaGraphExecEventRecordNodeSetEvent`, :py:obj:`~.cudaGraphExecEventWaitNodeSetEvent`, :py:obj:`~.cudaGraphExecExternalSemaphoresSignalNodeSetParams`, :py:obj:`~.cudaGraphExecExternalSemaphoresWaitNodeSetParams`, :py:obj:`~.cudaGraphExecUpdate`, :py:obj:`~.cudaGraphInstantiate` """ cdef cyruntime.cudaGraphNode_t cynode if node is None: pnode = 0 elif isinstance(node, (cudaGraphNode_t,driver.CUgraphNode)): pnode = int(node) else: pnode = int(cudaGraphNode_t(node)) cynode = pnode cdef cyruntime.cudaGraphExec_t cyhGraphExec if hGraphExec is None: phGraphExec = 0 elif isinstance(hGraphExec, (cudaGraphExec_t,driver.CUgraphExec)): phGraphExec = int(hGraphExec) else: phGraphExec = int(cudaGraphExec_t(hGraphExec)) cyhGraphExec = phGraphExec cdef cyruntime.cudaHostNodeParams* cypNodeParams_ptr = pNodeParams._pvt_ptr if pNodeParams is not None else NULL with nogil: err = cyruntime.cudaGraphExecHostNodeSetParams(cyhGraphExec, cynode, cypNodeParams_ptr) return (_dict_cudaError_t[err],) @cython.embedsignature(True) def cudaGraphExecChildGraphNodeSetParams(hGraphExec, node, childGraph): """ Updates node parameters in the child graph node in the given graphExec. Updates the work represented by `node` in `hGraphExec` as though the nodes contained in `node's` graph had the parameters contained in `childGraph's` nodes at instantiation. `node` must remain in the graph which was used to instantiate `hGraphExec`. Changed edges to and from `node` are ignored. The modifications only affect future launches of `hGraphExec`. Already enqueued or running launches of `hGraphExec` are not affected by this call. `node` is also not modified by this call. The topology of `childGraph`, as well as the node insertion order, must match that of the graph contained in `node`. See :py:obj:`~.cudaGraphExecUpdate()` for a list of restrictions on what can be updated in an instantiated graph. The update is recursive, so child graph nodes contained within the top level child graph will also be updated. Parameters ---------- hGraphExec : :py:obj:`~.CUgraphExec` or :py:obj:`~.cudaGraphExec_t` The executable graph in which to set the specified node node : :py:obj:`~.CUgraphNode` or :py:obj:`~.cudaGraphNode_t` Host node from the graph which was used to instantiate graphExec childGraph : :py:obj:`~.CUgraph` or :py:obj:`~.cudaGraph_t` The graph supplying the updated parameters Returns ------- cudaError_t :py:obj:`~.cudaSuccess`, :py:obj:`~.cudaErrorInvalidValue`, See Also -------- :py:obj:`~.cudaGraphExecNodeSetParams`, :py:obj:`~.cudaGraphAddChildGraphNode`, :py:obj:`~.cudaGraphChildGraphNodeGetGraph`, :py:obj:`~.cudaGraphExecKernelNodeSetParams`, :py:obj:`~.cudaGraphExecMemcpyNodeSetParams`, :py:obj:`~.cudaGraphExecMemsetNodeSetParams`, :py:obj:`~.cudaGraphExecHostNodeSetParams`, :py:obj:`~.cudaGraphExecEventRecordNodeSetEvent`, :py:obj:`~.cudaGraphExecEventWaitNodeSetEvent`, :py:obj:`~.cudaGraphExecExternalSemaphoresSignalNodeSetParams`, :py:obj:`~.cudaGraphExecExternalSemaphoresWaitNodeSetParams`, :py:obj:`~.cudaGraphExecUpdate`, :py:obj:`~.cudaGraphInstantiate` """ cdef cyruntime.cudaGraph_t cychildGraph if childGraph is None: pchildGraph = 0 elif isinstance(childGraph, (cudaGraph_t,driver.CUgraph)): pchildGraph = int(childGraph) else: pchildGraph = int(cudaGraph_t(childGraph)) cychildGraph = pchildGraph cdef cyruntime.cudaGraphNode_t cynode if node is None: pnode = 0 elif isinstance(node, (cudaGraphNode_t,driver.CUgraphNode)): pnode = int(node) else: pnode = int(cudaGraphNode_t(node)) cynode = pnode cdef cyruntime.cudaGraphExec_t cyhGraphExec if hGraphExec is None: phGraphExec = 0 elif isinstance(hGraphExec, (cudaGraphExec_t,driver.CUgraphExec)): phGraphExec = int(hGraphExec) else: phGraphExec = int(cudaGraphExec_t(hGraphExec)) cyhGraphExec = phGraphExec with nogil: err = cyruntime.cudaGraphExecChildGraphNodeSetParams(cyhGraphExec, cynode, cychildGraph) return (_dict_cudaError_t[err],) @cython.embedsignature(True) def cudaGraphExecEventRecordNodeSetEvent(hGraphExec, hNode, event): """ Sets the event for an event record node in the given graphExec. Sets the event of an event record node in an executable graph `hGraphExec`. The node is identified by the corresponding node `hNode` in the non-executable graph, from which the executable graph was instantiated. The modifications only affect future launches of `hGraphExec`. Already enqueued or running launches of `hGraphExec` are not affected by this call. `hNode` is also not modified by this call. Parameters ---------- hGraphExec : :py:obj:`~.CUgraphExec` or :py:obj:`~.cudaGraphExec_t` The executable graph in which to set the specified node hNode : :py:obj:`~.CUgraphNode` or :py:obj:`~.cudaGraphNode_t` Event record node from the graph from which graphExec was instantiated event : :py:obj:`~.CUevent` or :py:obj:`~.cudaEvent_t` Updated event to use Returns ------- cudaError_t :py:obj:`~.cudaSuccess`, :py:obj:`~.cudaErrorInvalidValue`, See Also -------- :py:obj:`~.cudaGraphExecNodeSetParams`, :py:obj:`~.cudaGraphAddEventRecordNode`, :py:obj:`~.cudaGraphEventRecordNodeGetEvent`, :py:obj:`~.cudaGraphEventWaitNodeSetEvent`, :py:obj:`~.cudaEventRecordWithFlags`, :py:obj:`~.cudaStreamWaitEvent`, :py:obj:`~.cudaGraphExecKernelNodeSetParams`, :py:obj:`~.cudaGraphExecMemcpyNodeSetParams`, :py:obj:`~.cudaGraphExecMemsetNodeSetParams`, :py:obj:`~.cudaGraphExecHostNodeSetParams`, :py:obj:`~.cudaGraphExecChildGraphNodeSetParams`, :py:obj:`~.cudaGraphExecEventWaitNodeSetEvent`, :py:obj:`~.cudaGraphExecExternalSemaphoresSignalNodeSetParams`, :py:obj:`~.cudaGraphExecExternalSemaphoresWaitNodeSetParams`, :py:obj:`~.cudaGraphExecUpdate`, :py:obj:`~.cudaGraphInstantiate` """ cdef cyruntime.cudaEvent_t cyevent if event is None: pevent = 0 elif isinstance(event, (cudaEvent_t,driver.CUevent)): pevent = int(event) else: pevent = int(cudaEvent_t(event)) cyevent = pevent cdef cyruntime.cudaGraphNode_t cyhNode if hNode is None: phNode = 0 elif isinstance(hNode, (cudaGraphNode_t,driver.CUgraphNode)): phNode = int(hNode) else: phNode = int(cudaGraphNode_t(hNode)) cyhNode = phNode cdef cyruntime.cudaGraphExec_t cyhGraphExec if hGraphExec is None: phGraphExec = 0 elif isinstance(hGraphExec, (cudaGraphExec_t,driver.CUgraphExec)): phGraphExec = int(hGraphExec) else: phGraphExec = int(cudaGraphExec_t(hGraphExec)) cyhGraphExec = phGraphExec with nogil: err = cyruntime.cudaGraphExecEventRecordNodeSetEvent(cyhGraphExec, cyhNode, cyevent) return (_dict_cudaError_t[err],) @cython.embedsignature(True) def cudaGraphExecEventWaitNodeSetEvent(hGraphExec, hNode, event): """ Sets the event for an event wait node in the given graphExec. Sets the event of an event wait node in an executable graph `hGraphExec`. The node is identified by the corresponding node `hNode` in the non-executable graph, from which the executable graph was instantiated. The modifications only affect future launches of `hGraphExec`. Already enqueued or running launches of `hGraphExec` are not affected by this call. `hNode` is also not modified by this call. Parameters ---------- hGraphExec : :py:obj:`~.CUgraphExec` or :py:obj:`~.cudaGraphExec_t` The executable graph in which to set the specified node hNode : :py:obj:`~.CUgraphNode` or :py:obj:`~.cudaGraphNode_t` Event wait node from the graph from which graphExec was instantiated event : :py:obj:`~.CUevent` or :py:obj:`~.cudaEvent_t` Updated event to use Returns ------- cudaError_t :py:obj:`~.cudaSuccess`, :py:obj:`~.cudaErrorInvalidValue`, See Also -------- :py:obj:`~.cudaGraphExecNodeSetParams`, :py:obj:`~.cudaGraphAddEventWaitNode`, :py:obj:`~.cudaGraphEventWaitNodeGetEvent`, :py:obj:`~.cudaGraphEventRecordNodeSetEvent`, :py:obj:`~.cudaEventRecordWithFlags`, :py:obj:`~.cudaStreamWaitEvent`, :py:obj:`~.cudaGraphExecKernelNodeSetParams`, :py:obj:`~.cudaGraphExecMemcpyNodeSetParams`, :py:obj:`~.cudaGraphExecMemsetNodeSetParams`, :py:obj:`~.cudaGraphExecHostNodeSetParams`, :py:obj:`~.cudaGraphExecChildGraphNodeSetParams`, :py:obj:`~.cudaGraphExecEventRecordNodeSetEvent`, :py:obj:`~.cudaGraphExecExternalSemaphoresSignalNodeSetParams`, :py:obj:`~.cudaGraphExecExternalSemaphoresWaitNodeSetParams`, :py:obj:`~.cudaGraphExecUpdate`, :py:obj:`~.cudaGraphInstantiate` """ cdef cyruntime.cudaEvent_t cyevent if event is None: pevent = 0 elif isinstance(event, (cudaEvent_t,driver.CUevent)): pevent = int(event) else: pevent = int(cudaEvent_t(event)) cyevent = pevent cdef cyruntime.cudaGraphNode_t cyhNode if hNode is None: phNode = 0 elif isinstance(hNode, (cudaGraphNode_t,driver.CUgraphNode)): phNode = int(hNode) else: phNode = int(cudaGraphNode_t(hNode)) cyhNode = phNode cdef cyruntime.cudaGraphExec_t cyhGraphExec if hGraphExec is None: phGraphExec = 0 elif isinstance(hGraphExec, (cudaGraphExec_t,driver.CUgraphExec)): phGraphExec = int(hGraphExec) else: phGraphExec = int(cudaGraphExec_t(hGraphExec)) cyhGraphExec = phGraphExec with nogil: err = cyruntime.cudaGraphExecEventWaitNodeSetEvent(cyhGraphExec, cyhNode, cyevent) return (_dict_cudaError_t[err],) @cython.embedsignature(True) def cudaGraphExecExternalSemaphoresSignalNodeSetParams(hGraphExec, hNode, nodeParams : Optional[cudaExternalSemaphoreSignalNodeParams]): """ Sets the parameters for an external semaphore signal node in the given graphExec. Sets the parameters of an external semaphore signal node in an executable graph `hGraphExec`. The node is identified by the corresponding node `hNode` in the non-executable graph, from which the executable graph was instantiated. `hNode` must not have been removed from the original graph. The modifications only affect future launches of `hGraphExec`. Already enqueued or running launches of `hGraphExec` are not affected by this call. `hNode` is also not modified by this call. Changing `nodeParams->numExtSems` is not supported. Parameters ---------- hGraphExec : :py:obj:`~.CUgraphExec` or :py:obj:`~.cudaGraphExec_t` The executable graph in which to set the specified node hNode : :py:obj:`~.CUgraphNode` or :py:obj:`~.cudaGraphNode_t` semaphore signal node from the graph from which graphExec was instantiated nodeParams : :py:obj:`~.cudaExternalSemaphoreSignalNodeParams` Updated Parameters to set Returns ------- cudaError_t :py:obj:`~.cudaSuccess`, :py:obj:`~.cudaErrorInvalidValue`, See Also -------- :py:obj:`~.cudaGraphExecNodeSetParams`, :py:obj:`~.cudaGraphAddExternalSemaphoresSignalNode`, :py:obj:`~.cudaImportExternalSemaphore`, :py:obj:`~.cudaSignalExternalSemaphoresAsync`, :py:obj:`~.cudaWaitExternalSemaphoresAsync`, :py:obj:`~.cudaGraphExecKernelNodeSetParams`, :py:obj:`~.cudaGraphExecMemcpyNodeSetParams`, :py:obj:`~.cudaGraphExecMemsetNodeSetParams`, :py:obj:`~.cudaGraphExecHostNodeSetParams`, :py:obj:`~.cudaGraphExecChildGraphNodeSetParams`, :py:obj:`~.cudaGraphExecEventRecordNodeSetEvent`, :py:obj:`~.cudaGraphExecEventWaitNodeSetEvent`, :py:obj:`~.cudaGraphExecExternalSemaphoresWaitNodeSetParams`, :py:obj:`~.cudaGraphExecUpdate`, :py:obj:`~.cudaGraphInstantiate` """ cdef cyruntime.cudaGraphNode_t cyhNode if hNode is None: phNode = 0 elif isinstance(hNode, (cudaGraphNode_t,driver.CUgraphNode)): phNode = int(hNode) else: phNode = int(cudaGraphNode_t(hNode)) cyhNode = phNode cdef cyruntime.cudaGraphExec_t cyhGraphExec if hGraphExec is None: phGraphExec = 0 elif isinstance(hGraphExec, (cudaGraphExec_t,driver.CUgraphExec)): phGraphExec = int(hGraphExec) else: phGraphExec = int(cudaGraphExec_t(hGraphExec)) cyhGraphExec = phGraphExec cdef cyruntime.cudaExternalSemaphoreSignalNodeParams* cynodeParams_ptr = nodeParams._pvt_ptr if nodeParams is not None else NULL with nogil: err = cyruntime.cudaGraphExecExternalSemaphoresSignalNodeSetParams(cyhGraphExec, cyhNode, cynodeParams_ptr) return (_dict_cudaError_t[err],) @cython.embedsignature(True) def cudaGraphExecExternalSemaphoresWaitNodeSetParams(hGraphExec, hNode, nodeParams : Optional[cudaExternalSemaphoreWaitNodeParams]): """ Sets the parameters for an external semaphore wait node in the given graphExec. Sets the parameters of an external semaphore wait node in an executable graph `hGraphExec`. The node is identified by the corresponding node `hNode` in the non-executable graph, from which the executable graph was instantiated. `hNode` must not have been removed from the original graph. The modifications only affect future launches of `hGraphExec`. Already enqueued or running launches of `hGraphExec` are not affected by this call. `hNode` is also not modified by this call. Changing `nodeParams->numExtSems` is not supported. Parameters ---------- hGraphExec : :py:obj:`~.CUgraphExec` or :py:obj:`~.cudaGraphExec_t` The executable graph in which to set the specified node hNode : :py:obj:`~.CUgraphNode` or :py:obj:`~.cudaGraphNode_t` semaphore wait node from the graph from which graphExec was instantiated nodeParams : :py:obj:`~.cudaExternalSemaphoreWaitNodeParams` Updated Parameters to set Returns ------- cudaError_t :py:obj:`~.cudaSuccess`, :py:obj:`~.cudaErrorInvalidValue`, See Also -------- :py:obj:`~.cudaGraphExecNodeSetParams`, :py:obj:`~.cudaGraphAddExternalSemaphoresWaitNode`, :py:obj:`~.cudaImportExternalSemaphore`, :py:obj:`~.cudaSignalExternalSemaphoresAsync`, :py:obj:`~.cudaWaitExternalSemaphoresAsync`, :py:obj:`~.cudaGraphExecKernelNodeSetParams`, :py:obj:`~.cudaGraphExecMemcpyNodeSetParams`, :py:obj:`~.cudaGraphExecMemsetNodeSetParams`, :py:obj:`~.cudaGraphExecHostNodeSetParams`, :py:obj:`~.cudaGraphExecChildGraphNodeSetParams`, :py:obj:`~.cudaGraphExecEventRecordNodeSetEvent`, :py:obj:`~.cudaGraphExecEventWaitNodeSetEvent`, :py:obj:`~.cudaGraphExecExternalSemaphoresSignalNodeSetParams`, :py:obj:`~.cudaGraphExecUpdate`, :py:obj:`~.cudaGraphInstantiate` """ cdef cyruntime.cudaGraphNode_t cyhNode if hNode is None: phNode = 0 elif isinstance(hNode, (cudaGraphNode_t,driver.CUgraphNode)): phNode = int(hNode) else: phNode = int(cudaGraphNode_t(hNode)) cyhNode = phNode cdef cyruntime.cudaGraphExec_t cyhGraphExec if hGraphExec is None: phGraphExec = 0 elif isinstance(hGraphExec, (cudaGraphExec_t,driver.CUgraphExec)): phGraphExec = int(hGraphExec) else: phGraphExec = int(cudaGraphExec_t(hGraphExec)) cyhGraphExec = phGraphExec cdef cyruntime.cudaExternalSemaphoreWaitNodeParams* cynodeParams_ptr = nodeParams._pvt_ptr if nodeParams is not None else NULL with nogil: err = cyruntime.cudaGraphExecExternalSemaphoresWaitNodeSetParams(cyhGraphExec, cyhNode, cynodeParams_ptr) return (_dict_cudaError_t[err],) @cython.embedsignature(True) def cudaGraphNodeSetEnabled(hGraphExec, hNode, unsigned int isEnabled): """ Enables or disables the specified node in the given graphExec. Sets `hNode` to be either enabled or disabled. Disabled nodes are functionally equivalent to empty nodes until they are reenabled. Existing node parameters are not affected by disabling/enabling the node. The node is identified by the corresponding node `hNode` in the non- executable graph, from which the executable graph was instantiated. `hNode` must not have been removed from the original graph. The modifications only affect future launches of `hGraphExec`. Already enqueued or running launches of `hGraphExec` are not affected by this call. `hNode` is also not modified by this call. Parameters ---------- hGraphExec : :py:obj:`~.CUgraphExec` or :py:obj:`~.cudaGraphExec_t` The executable graph in which to set the specified node hNode : :py:obj:`~.CUgraphNode` or :py:obj:`~.cudaGraphNode_t` Node from the graph from which graphExec was instantiated isEnabled : unsigned int Node is enabled if != 0, otherwise the node is disabled Returns ------- cudaError_t :py:obj:`~.cudaSuccess`, :py:obj:`~.cudaErrorInvalidValue`, See Also -------- :py:obj:`~.cudaGraphNodeGetEnabled`, :py:obj:`~.cudaGraphExecUpdate`, :py:obj:`~.cudaGraphInstantiate` :py:obj:`~.cudaGraphLaunch` Notes ----- Currently only kernel, memset and memcpy nodes are supported. """ cdef cyruntime.cudaGraphNode_t cyhNode if hNode is None: phNode = 0 elif isinstance(hNode, (cudaGraphNode_t,driver.CUgraphNode)): phNode = int(hNode) else: phNode = int(cudaGraphNode_t(hNode)) cyhNode = phNode cdef cyruntime.cudaGraphExec_t cyhGraphExec if hGraphExec is None: phGraphExec = 0 elif isinstance(hGraphExec, (cudaGraphExec_t,driver.CUgraphExec)): phGraphExec = int(hGraphExec) else: phGraphExec = int(cudaGraphExec_t(hGraphExec)) cyhGraphExec = phGraphExec with nogil: err = cyruntime.cudaGraphNodeSetEnabled(cyhGraphExec, cyhNode, isEnabled) return (_dict_cudaError_t[err],) @cython.embedsignature(True) def cudaGraphNodeGetEnabled(hGraphExec, hNode): """ Query whether a node in the given graphExec is enabled. Sets isEnabled to 1 if `hNode` is enabled, or 0 if `hNode` is disabled. The node is identified by the corresponding node `hNode` in the non- executable graph, from which the executable graph was instantiated. `hNode` must not have been removed from the original graph. Parameters ---------- hGraphExec : :py:obj:`~.CUgraphExec` or :py:obj:`~.cudaGraphExec_t` The executable graph in which to set the specified node hNode : :py:obj:`~.CUgraphNode` or :py:obj:`~.cudaGraphNode_t` Node from the graph from which graphExec was instantiated Returns ------- cudaError_t :py:obj:`~.cudaSuccess`, :py:obj:`~.cudaErrorInvalidValue`, isEnabled : unsigned int Location to return the enabled status of the node See Also -------- :py:obj:`~.cudaGraphNodeSetEnabled`, :py:obj:`~.cudaGraphExecUpdate`, :py:obj:`~.cudaGraphInstantiate` :py:obj:`~.cudaGraphLaunch` Notes ----- Currently only kernel, memset and memcpy nodes are supported. """ cdef cyruntime.cudaGraphNode_t cyhNode if hNode is None: phNode = 0 elif isinstance(hNode, (cudaGraphNode_t,driver.CUgraphNode)): phNode = int(hNode) else: phNode = int(cudaGraphNode_t(hNode)) cyhNode = phNode cdef cyruntime.cudaGraphExec_t cyhGraphExec if hGraphExec is None: phGraphExec = 0 elif isinstance(hGraphExec, (cudaGraphExec_t,driver.CUgraphExec)): phGraphExec = int(hGraphExec) else: phGraphExec = int(cudaGraphExec_t(hGraphExec)) cyhGraphExec = phGraphExec cdef unsigned int isEnabled = 0 with nogil: err = cyruntime.cudaGraphNodeGetEnabled(cyhGraphExec, cyhNode, &isEnabled) if err != cyruntime.cudaSuccess: return (_dict_cudaError_t[err], None) return (_dict_cudaError_t[err], isEnabled) @cython.embedsignature(True) def cudaGraphExecUpdate(hGraphExec, hGraph): """ Check whether an executable graph can be updated with a graph and perform the update if possible. Updates the node parameters in the instantiated graph specified by `hGraphExec` with the node parameters in a topologically identical graph specified by `hGraph`. Limitations: - Kernel nodes: - The owning context of the function cannot change. - A node whose function originally did not use CUDA dynamic parallelism cannot be updated to a function which uses CDP. - A node whose function originally did not make device-side update calls cannot be updated to a function which makes device-side update calls. - A cooperative node cannot be updated to a non-cooperative node, and vice-versa. - If the graph was instantiated with cudaGraphInstantiateFlagUseNodePriority, the priority attribute cannot change. Equality is checked on the originally requested priority values, before they are clamped to the device's supported range. - If `hGraphExec` was not instantiated for device launch, a node whose function originally did not use device-side :py:obj:`~.cudaGraphLaunch()` cannot be updated to a function which uses device-side :py:obj:`~.cudaGraphLaunch()` unless the node resides on the same device as nodes which contained such calls at instantiate-time. If no such calls were present at instantiation, these updates cannot be performed at all. - Neither `hGraph` nor `hGraphExec` may contain device-updatable kernel nodes. - Memset and memcpy nodes: - The CUDA device(s) to which the operand(s) was allocated/mapped cannot change. - The source/destination memory must be allocated from the same contexts as the original source/destination memory. - For 2d memsets, only address and assigned value may be updated. - For 1d memsets, updating dimensions is also allowed, but may fail if the resulting operation doesn't map onto the work resources already allocated for the node. - Additional memcpy node restrictions: - Changing either the source or destination memory type(i.e. CU_MEMORYTYPE_DEVICE, CU_MEMORYTYPE_ARRAY, etc.) is not supported. - Conditional nodes: - Changing node parameters is not supported. - Changing parameters of nodes within the conditional body graph is subject to the rules above. - Conditional handle flags and default values are updated as part of the graph update. Note: The API may add further restrictions in future releases. The return code should always be checked. cudaGraphExecUpdate sets the result member of `resultInfo` to cudaGraphExecUpdateErrorTopologyChanged under the following conditions: - The count of nodes directly in `hGraphExec` and `hGraph` differ, in which case resultInfo->errorNode is set to NULL. - `hGraph` has more exit nodes than `hGraph`, in which case resultInfo->errorNode is set to one of the exit nodes in hGraph. - A node in `hGraph` has a different number of dependencies than the node from `hGraphExec` it is paired with, in which case resultInfo->errorNode is set to the node from `hGraph`. - A node in `hGraph` has a dependency that does not match with the corresponding dependency of the paired node from `hGraphExec`. resultInfo->errorNode will be set to the node from `hGraph`. resultInfo->errorFromNode will be set to the mismatched dependency. The dependencies are paired based on edge order and a dependency does not match when the nodes are already paired based on other edges examined in the graph. cudaGraphExecUpdate sets `the` result member of `resultInfo` to: - cudaGraphExecUpdateError if passed an invalid value. - cudaGraphExecUpdateErrorTopologyChanged if the graph topology changed - cudaGraphExecUpdateErrorNodeTypeChanged if the type of a node changed, in which case `hErrorNode_out` is set to the node from `hGraph`. - cudaGraphExecUpdateErrorFunctionChanged if the function of a kernel node changed (CUDA driver < 11.2) - cudaGraphExecUpdateErrorUnsupportedFunctionChange if the func field of a kernel changed in an unsupported way(see note above), in which case `hErrorNode_out` is set to the node from `hGraph` - cudaGraphExecUpdateErrorParametersChanged if any parameters to a node changed in a way that is not supported, in which case `hErrorNode_out` is set to the node from `hGraph` - cudaGraphExecUpdateErrorAttributesChanged if any attributes of a node changed in a way that is not supported, in which case `hErrorNode_out` is set to the node from `hGraph` - cudaGraphExecUpdateErrorNotSupported if something about a node is unsupported, like the node's type or configuration, in which case `hErrorNode_out` is set to the node from `hGraph` If the update fails for a reason not listed above, the result member of `resultInfo` will be set to cudaGraphExecUpdateError. If the update succeeds, the result member will be set to cudaGraphExecUpdateSuccess. cudaGraphExecUpdate returns cudaSuccess when the updated was performed successfully. It returns cudaErrorGraphExecUpdateFailure if the graph update was not performed because it included changes which violated constraints specific to instantiated graph update. Parameters ---------- hGraphExec : :py:obj:`~.CUgraphExec` or :py:obj:`~.cudaGraphExec_t` The instantiated graph to be updated hGraph : :py:obj:`~.CUgraph` or :py:obj:`~.cudaGraph_t` The graph containing the updated parameters Returns ------- cudaError_t :py:obj:`~.cudaSuccess`, :py:obj:`~.cudaErrorGraphExecUpdateFailure`, resultInfo : :py:obj:`~.cudaGraphExecUpdateResultInfo` the error info structure See Also -------- :py:obj:`~.cudaGraphInstantiate` """ cdef cyruntime.cudaGraph_t cyhGraph if hGraph is None: phGraph = 0 elif isinstance(hGraph, (cudaGraph_t,driver.CUgraph)): phGraph = int(hGraph) else: phGraph = int(cudaGraph_t(hGraph)) cyhGraph = phGraph cdef cyruntime.cudaGraphExec_t cyhGraphExec if hGraphExec is None: phGraphExec = 0 elif isinstance(hGraphExec, (cudaGraphExec_t,driver.CUgraphExec)): phGraphExec = int(hGraphExec) else: phGraphExec = int(cudaGraphExec_t(hGraphExec)) cyhGraphExec = phGraphExec cdef cudaGraphExecUpdateResultInfo resultInfo = cudaGraphExecUpdateResultInfo() with nogil: err = cyruntime.cudaGraphExecUpdate(cyhGraphExec, cyhGraph, resultInfo._pvt_ptr) if err != cyruntime.cudaSuccess: return (_dict_cudaError_t[err], None) return (_dict_cudaError_t[err], resultInfo) @cython.embedsignature(True) def cudaGraphUpload(graphExec, stream): """ Uploads an executable graph in a stream. Uploads `hGraphExec` to the device in `hStream` without executing it. Uploads of the same `hGraphExec` will be serialized. Each upload is ordered behind both any previous work in `hStream` and any previous launches of `hGraphExec`. Uses memory cached by `stream` to back the allocations owned by `graphExec`. Parameters ---------- hGraphExec : :py:obj:`~.CUgraphExec` or :py:obj:`~.cudaGraphExec_t` Executable graph to upload hStream : :py:obj:`~.CUstream` or :py:obj:`~.cudaStream_t` Stream in which to upload the graph Returns ------- cudaError_t :py:obj:`~.cudaSuccess`, :py:obj:`~.cudaErrorInvalidValue`, See Also -------- :py:obj:`~.cudaGraphInstantiate`, :py:obj:`~.cudaGraphLaunch`, :py:obj:`~.cudaGraphExecDestroy` """ cdef cyruntime.cudaStream_t cystream if stream is None: pstream = 0 elif isinstance(stream, (cudaStream_t,driver.CUstream)): pstream = int(stream) else: pstream = int(cudaStream_t(stream)) cystream = pstream cdef cyruntime.cudaGraphExec_t cygraphExec if graphExec is None: pgraphExec = 0 elif isinstance(graphExec, (cudaGraphExec_t,driver.CUgraphExec)): pgraphExec = int(graphExec) else: pgraphExec = int(cudaGraphExec_t(graphExec)) cygraphExec = pgraphExec with nogil: err = cyruntime.cudaGraphUpload(cygraphExec, cystream) return (_dict_cudaError_t[err],) @cython.embedsignature(True) def cudaGraphLaunch(graphExec, stream): """ Launches an executable graph in a stream. Executes `graphExec` in `stream`. Only one instance of `graphExec` may be executing at a time. Each launch is ordered behind both any previous work in `stream` and any previous launches of `graphExec`. To execute a graph concurrently, it must be instantiated multiple times into multiple executable graphs. If any allocations created by `graphExec` remain unfreed (from a previous launch) and `graphExec` was not instantiated with :py:obj:`~.cudaGraphInstantiateFlagAutoFreeOnLaunch`, the launch will fail with :py:obj:`~.cudaErrorInvalidValue`. Parameters ---------- graphExec : :py:obj:`~.CUgraphExec` or :py:obj:`~.cudaGraphExec_t` Executable graph to launch stream : :py:obj:`~.CUstream` or :py:obj:`~.cudaStream_t` Stream in which to launch the graph Returns ------- cudaError_t :py:obj:`~.cudaSuccess`, :py:obj:`~.cudaErrorInvalidValue` See Also -------- :py:obj:`~.cudaGraphInstantiate`, :py:obj:`~.cudaGraphUpload`, :py:obj:`~.cudaGraphExecDestroy` """ cdef cyruntime.cudaStream_t cystream if stream is None: pstream = 0 elif isinstance(stream, (cudaStream_t,driver.CUstream)): pstream = int(stream) else: pstream = int(cudaStream_t(stream)) cystream = pstream cdef cyruntime.cudaGraphExec_t cygraphExec if graphExec is None: pgraphExec = 0 elif isinstance(graphExec, (cudaGraphExec_t,driver.CUgraphExec)): pgraphExec = int(graphExec) else: pgraphExec = int(cudaGraphExec_t(graphExec)) cygraphExec = pgraphExec with nogil: err = cyruntime.cudaGraphLaunch(cygraphExec, cystream) return (_dict_cudaError_t[err],) @cython.embedsignature(True) def cudaGraphExecDestroy(graphExec): """ Destroys an executable graph. Destroys the executable graph specified by `graphExec`. Parameters ---------- graphExec : :py:obj:`~.CUgraphExec` or :py:obj:`~.cudaGraphExec_t` Executable graph to destroy Returns ------- cudaError_t :py:obj:`~.cudaSuccess`, :py:obj:`~.cudaErrorInvalidValue` See Also -------- :py:obj:`~.cudaGraphInstantiate`, :py:obj:`~.cudaGraphUpload`, :py:obj:`~.cudaGraphLaunch` """ cdef cyruntime.cudaGraphExec_t cygraphExec if graphExec is None: pgraphExec = 0 elif isinstance(graphExec, (cudaGraphExec_t,driver.CUgraphExec)): pgraphExec = int(graphExec) else: pgraphExec = int(cudaGraphExec_t(graphExec)) cygraphExec = pgraphExec with nogil: err = cyruntime.cudaGraphExecDestroy(cygraphExec) return (_dict_cudaError_t[err],) @cython.embedsignature(True) def cudaGraphDestroy(graph): """ Destroys a graph. Destroys the graph specified by `graph`, as well as all of its nodes. Parameters ---------- graph : :py:obj:`~.CUgraph` or :py:obj:`~.cudaGraph_t` Graph to destroy Returns ------- cudaError_t :py:obj:`~.cudaSuccess`, :py:obj:`~.cudaErrorInvalidValue` See Also -------- :py:obj:`~.cudaGraphCreate` """ cdef cyruntime.cudaGraph_t cygraph if graph is None: pgraph = 0 elif isinstance(graph, (cudaGraph_t,driver.CUgraph)): pgraph = int(graph) else: pgraph = int(cudaGraph_t(graph)) cygraph = pgraph with nogil: err = cyruntime.cudaGraphDestroy(cygraph) return (_dict_cudaError_t[err],) @cython.embedsignature(True) def cudaGraphDebugDotPrint(graph, char* path, unsigned int flags): """ Write a DOT file describing graph structure. Using the provided `graph`, write to `path` a DOT formatted description of the graph. By default this includes the graph topology, node types, node id, kernel names and memcpy direction. `flags` can be specified to write more detailed information about each node type such as parameter values, kernel attributes, node and function handles. Parameters ---------- graph : :py:obj:`~.CUgraph` or :py:obj:`~.cudaGraph_t` The graph to create a DOT file from path : bytes The path to write the DOT file to flags : unsigned int Flags from cudaGraphDebugDotFlags for specifying which additional node information to write Returns ------- cudaError_t :py:obj:`~.cudaSuccess`, :py:obj:`~.cudaErrorInvalidValue`, :py:obj:`~.cudaErrorOperatingSystem` """ cdef cyruntime.cudaGraph_t cygraph if graph is None: pgraph = 0 elif isinstance(graph, (cudaGraph_t,driver.CUgraph)): pgraph = int(graph) else: pgraph = int(cudaGraph_t(graph)) cygraph = pgraph with nogil: err = cyruntime.cudaGraphDebugDotPrint(cygraph, path, flags) return (_dict_cudaError_t[err],) @cython.embedsignature(True) def cudaUserObjectCreate(ptr, destroy, unsigned int initialRefcount, unsigned int flags): """ Create a user object. Create a user object with the specified destructor callback and initial reference count. The initial references are owned by the caller. Destructor callbacks cannot make CUDA API calls and should avoid blocking behavior, as they are executed by a shared internal thread. Another thread may be signaled to perform such actions, if it does not block forward progress of tasks scheduled through CUDA. See CUDA User Objects in the CUDA C++ Programming Guide for more information on user objects. Parameters ---------- ptr : Any The pointer to pass to the destroy function destroy : :py:obj:`~.cudaHostFn_t` Callback to free the user object when it is no longer in use initialRefcount : unsigned int The initial refcount to create the object with, typically 1. The initial references are owned by the calling thread. flags : unsigned int Currently it is required to pass :py:obj:`~.cudaUserObjectNoDestructorSync`, which is the only defined flag. This indicates that the destroy callback cannot be waited on by any CUDA API. Users requiring synchronization of the callback should signal its completion manually. Returns ------- cudaError_t :py:obj:`~.cudaSuccess`, :py:obj:`~.cudaErrorInvalidValue` object_out : :py:obj:`~.cudaUserObject_t` Location to return the user object handle See Also -------- :py:obj:`~.cudaUserObjectRetain`, :py:obj:`~.cudaUserObjectRelease`, :py:obj:`~.cudaGraphRetainUserObject`, :py:obj:`~.cudaGraphReleaseUserObject`, :py:obj:`~.cudaGraphCreate` """ cdef cyruntime.cudaHostFn_t cydestroy if destroy is None: pdestroy = 0 elif isinstance(destroy, (cudaHostFn_t,)): pdestroy = int(destroy) else: pdestroy = int(cudaHostFn_t(destroy)) cydestroy = pdestroy cdef cudaUserObject_t object_out = cudaUserObject_t() cyptr = _HelperInputVoidPtr(ptr) cdef void* cyptr_ptr = cyptr.cptr with nogil: err = cyruntime.cudaUserObjectCreate(object_out._pvt_ptr, cyptr_ptr, cydestroy, initialRefcount, flags) if err != cyruntime.cudaSuccess: return (_dict_cudaError_t[err], None) return (_dict_cudaError_t[err], object_out) @cython.embedsignature(True) def cudaUserObjectRetain(object, unsigned int count): """ Retain a reference to a user object. Retains new references to a user object. The new references are owned by the caller. See CUDA User Objects in the CUDA C++ Programming Guide for more information on user objects. Parameters ---------- object : :py:obj:`~.cudaUserObject_t` The object to retain count : unsigned int The number of references to retain, typically 1. Must be nonzero and not larger than INT_MAX. Returns ------- cudaError_t :py:obj:`~.cudaSuccess`, :py:obj:`~.cudaErrorInvalidValue` See Also -------- :py:obj:`~.cudaUserObjectCreate`, :py:obj:`~.cudaUserObjectRelease`, :py:obj:`~.cudaGraphRetainUserObject`, :py:obj:`~.cudaGraphReleaseUserObject`, :py:obj:`~.cudaGraphCreate` """ cdef cyruntime.cudaUserObject_t cyobject if object is None: pobject = 0 elif isinstance(object, (cudaUserObject_t,driver.CUuserObject)): pobject = int(object) else: pobject = int(cudaUserObject_t(object)) cyobject = pobject with nogil: err = cyruntime.cudaUserObjectRetain(cyobject, count) return (_dict_cudaError_t[err],) @cython.embedsignature(True) def cudaUserObjectRelease(object, unsigned int count): """ Release a reference to a user object. Releases user object references owned by the caller. The object's destructor is invoked if the reference count reaches zero. It is undefined behavior to release references not owned by the caller, or to use a user object handle after all references are released. See CUDA User Objects in the CUDA C++ Programming Guide for more information on user objects. Parameters ---------- object : :py:obj:`~.cudaUserObject_t` The object to release count : unsigned int The number of references to release, typically 1. Must be nonzero and not larger than INT_MAX. Returns ------- cudaError_t :py:obj:`~.cudaSuccess`, :py:obj:`~.cudaErrorInvalidValue` See Also -------- :py:obj:`~.cudaUserObjectCreate`, :py:obj:`~.cudaUserObjectRetain`, :py:obj:`~.cudaGraphRetainUserObject`, :py:obj:`~.cudaGraphReleaseUserObject`, :py:obj:`~.cudaGraphCreate` """ cdef cyruntime.cudaUserObject_t cyobject if object is None: pobject = 0 elif isinstance(object, (cudaUserObject_t,driver.CUuserObject)): pobject = int(object) else: pobject = int(cudaUserObject_t(object)) cyobject = pobject with nogil: err = cyruntime.cudaUserObjectRelease(cyobject, count) return (_dict_cudaError_t[err],) @cython.embedsignature(True) def cudaGraphRetainUserObject(graph, object, unsigned int count, unsigned int flags): """ Retain a reference to a user object from a graph. Creates or moves user object references that will be owned by a CUDA graph. See CUDA User Objects in the CUDA C++ Programming Guide for more information on user objects. Parameters ---------- graph : :py:obj:`~.CUgraph` or :py:obj:`~.cudaGraph_t` The graph to associate the reference with object : :py:obj:`~.cudaUserObject_t` The user object to retain a reference for count : unsigned int The number of references to add to the graph, typically 1. Must be nonzero and not larger than INT_MAX. flags : unsigned int The optional flag :py:obj:`~.cudaGraphUserObjectMove` transfers references from the calling thread, rather than create new references. Pass 0 to create new references. Returns ------- cudaError_t :py:obj:`~.cudaSuccess`, :py:obj:`~.cudaErrorInvalidValue` See Also -------- :py:obj:`~.cudaUserObjectCreate` :py:obj:`~.cudaUserObjectRetain`, :py:obj:`~.cudaUserObjectRelease`, :py:obj:`~.cudaGraphReleaseUserObject`, :py:obj:`~.cudaGraphCreate` """ cdef cyruntime.cudaUserObject_t cyobject if object is None: pobject = 0 elif isinstance(object, (cudaUserObject_t,driver.CUuserObject)): pobject = int(object) else: pobject = int(cudaUserObject_t(object)) cyobject = pobject cdef cyruntime.cudaGraph_t cygraph if graph is None: pgraph = 0 elif isinstance(graph, (cudaGraph_t,driver.CUgraph)): pgraph = int(graph) else: pgraph = int(cudaGraph_t(graph)) cygraph = pgraph with nogil: err = cyruntime.cudaGraphRetainUserObject(cygraph, cyobject, count, flags) return (_dict_cudaError_t[err],) @cython.embedsignature(True) def cudaGraphReleaseUserObject(graph, object, unsigned int count): """ Release a user object reference from a graph. Releases user object references owned by a graph. See CUDA User Objects in the CUDA C++ Programming Guide for more information on user objects. Parameters ---------- graph : :py:obj:`~.CUgraph` or :py:obj:`~.cudaGraph_t` The graph that will release the reference object : :py:obj:`~.cudaUserObject_t` The user object to release a reference for count : unsigned int The number of references to release, typically 1. Must be nonzero and not larger than INT_MAX. Returns ------- cudaError_t :py:obj:`~.cudaSuccess`, :py:obj:`~.cudaErrorInvalidValue` See Also -------- :py:obj:`~.cudaUserObjectCreate` :py:obj:`~.cudaUserObjectRetain`, :py:obj:`~.cudaUserObjectRelease`, :py:obj:`~.cudaGraphRetainUserObject`, :py:obj:`~.cudaGraphCreate` """ cdef cyruntime.cudaUserObject_t cyobject if object is None: pobject = 0 elif isinstance(object, (cudaUserObject_t,driver.CUuserObject)): pobject = int(object) else: pobject = int(cudaUserObject_t(object)) cyobject = pobject cdef cyruntime.cudaGraph_t cygraph if graph is None: pgraph = 0 elif isinstance(graph, (cudaGraph_t,driver.CUgraph)): pgraph = int(graph) else: pgraph = int(cudaGraph_t(graph)) cygraph = pgraph with nogil: err = cyruntime.cudaGraphReleaseUserObject(cygraph, cyobject, count) return (_dict_cudaError_t[err],) @cython.embedsignature(True) def cudaGraphAddNode(graph, pDependencies : Optional[tuple[cudaGraphNode_t] | list[cudaGraphNode_t]], size_t numDependencies, nodeParams : Optional[cudaGraphNodeParams]): """ Adds a node of arbitrary type to a graph. Creates a new node in `graph` described by `nodeParams` with `numDependencies` dependencies specified via `pDependencies`. `numDependencies` may be 0. `pDependencies` may be null if `numDependencies` is 0. `pDependencies` may not have any duplicate entries. `nodeParams` is a tagged union. The node type should be specified in the `typename` field, and type-specific parameters in the corresponding union member. All unused bytes - that is, `reserved0` and all bytes past the utilized union member - must be set to zero. It is recommended to use brace initialization or memset to ensure all bytes are initialized. Note that for some node types, `nodeParams` may contain "out parameters" which are modified during the call, such as `nodeParams->alloc.dptr`. A handle to the new node will be returned in `phGraphNode`. Parameters ---------- graph : :py:obj:`~.CUgraph` or :py:obj:`~.cudaGraph_t` Graph to which to add the node pDependencies : list[:py:obj:`~.cudaGraphNode_t`] Dependencies of the node numDependencies : size_t Number of dependencies nodeParams : :py:obj:`~.cudaGraphNodeParams` Specification of the node Returns ------- cudaError_t :py:obj:`~.cudaSuccess`, :py:obj:`~.cudaErrorInvalidValue`, :py:obj:`~.cudaErrorInvalidDeviceFunction`, :py:obj:`~.cudaErrorNotSupported` pGraphNode : :py:obj:`~.cudaGraphNode_t` Returns newly created node See Also -------- :py:obj:`~.cudaGraphCreate`, :py:obj:`~.cudaGraphNodeSetParams`, :py:obj:`~.cudaGraphExecNodeSetParams` """ pDependencies = [] if pDependencies is None else pDependencies if not all(isinstance(_x, (cudaGraphNode_t,driver.CUgraphNode)) for _x in pDependencies): raise TypeError("Argument 'pDependencies' is not instance of type (expected tuple[cyruntime.cudaGraphNode_t,driver.CUgraphNode] or list[cyruntime.cudaGraphNode_t,driver.CUgraphNode]") cdef cyruntime.cudaGraph_t cygraph if graph is None: pgraph = 0 elif isinstance(graph, (cudaGraph_t,driver.CUgraph)): pgraph = int(graph) else: pgraph = int(cudaGraph_t(graph)) cygraph = pgraph cdef cudaGraphNode_t pGraphNode = cudaGraphNode_t() cdef cyruntime.cudaGraphNode_t* cypDependencies = NULL if len(pDependencies) > 1: cypDependencies = calloc(len(pDependencies), sizeof(cyruntime.cudaGraphNode_t)) if cypDependencies is NULL: raise MemoryError('Failed to allocate length x size memory: ' + str(len(pDependencies)) + 'x' + str(sizeof(cyruntime.cudaGraphNode_t))) else: for idx in range(len(pDependencies)): cypDependencies[idx] = (pDependencies[idx])._pvt_ptr[0] elif len(pDependencies) == 1: cypDependencies = (pDependencies[0])._pvt_ptr if numDependencies > len(pDependencies): raise RuntimeError("List is too small: " + str(len(pDependencies)) + " < " + str(numDependencies)) cdef cyruntime.cudaGraphNodeParams* cynodeParams_ptr = nodeParams._pvt_ptr if nodeParams is not None else NULL with nogil: err = cyruntime.cudaGraphAddNode(pGraphNode._pvt_ptr, cygraph, cypDependencies, numDependencies, cynodeParams_ptr) if len(pDependencies) > 1 and cypDependencies is not NULL: free(cypDependencies) if err != cyruntime.cudaSuccess: return (_dict_cudaError_t[err], None) return (_dict_cudaError_t[err], pGraphNode) @cython.embedsignature(True) def cudaGraphAddNode_v2(graph, pDependencies : Optional[tuple[cudaGraphNode_t] | list[cudaGraphNode_t]], dependencyData : Optional[tuple[cudaGraphEdgeData] | list[cudaGraphEdgeData]], size_t numDependencies, nodeParams : Optional[cudaGraphNodeParams]): """ Adds a node of arbitrary type to a graph (12.3+) Creates a new node in `graph` described by `nodeParams` with `numDependencies` dependencies specified via `pDependencies`. `numDependencies` may be 0. `pDependencies` may be null if `numDependencies` is 0. `pDependencies` may not have any duplicate entries. `nodeParams` is a tagged union. The node type should be specified in the `typename` field, and type-specific parameters in the corresponding union member. All unused bytes - that is, `reserved0` and all bytes past the utilized union member - must be set to zero. It is recommended to use brace initialization or memset to ensure all bytes are initialized. Note that for some node types, `nodeParams` may contain "out parameters" which are modified during the call, such as `nodeParams->alloc.dptr`. A handle to the new node will be returned in `phGraphNode`. Parameters ---------- graph : :py:obj:`~.CUgraph` or :py:obj:`~.cudaGraph_t` Graph to which to add the node pDependencies : list[:py:obj:`~.cudaGraphNode_t`] Dependencies of the node dependencyData : list[:py:obj:`~.cudaGraphEdgeData`] Optional edge data for the dependencies. If NULL, the data is assumed to be default (zeroed) for all dependencies. numDependencies : size_t Number of dependencies nodeParams : :py:obj:`~.cudaGraphNodeParams` Specification of the node Returns ------- cudaError_t :py:obj:`~.cudaSuccess`, :py:obj:`~.cudaErrorInvalidValue`, :py:obj:`~.cudaErrorInvalidDeviceFunction`, :py:obj:`~.cudaErrorNotSupported` pGraphNode : :py:obj:`~.cudaGraphNode_t` Returns newly created node See Also -------- :py:obj:`~.cudaGraphCreate`, :py:obj:`~.cudaGraphNodeSetParams`, :py:obj:`~.cudaGraphExecNodeSetParams` """ dependencyData = [] if dependencyData is None else dependencyData if not all(isinstance(_x, (cudaGraphEdgeData,)) for _x in dependencyData): raise TypeError("Argument 'dependencyData' is not instance of type (expected tuple[cyruntime.cudaGraphEdgeData,] or list[cyruntime.cudaGraphEdgeData,]") pDependencies = [] if pDependencies is None else pDependencies if not all(isinstance(_x, (cudaGraphNode_t,driver.CUgraphNode)) for _x in pDependencies): raise TypeError("Argument 'pDependencies' is not instance of type (expected tuple[cyruntime.cudaGraphNode_t,driver.CUgraphNode] or list[cyruntime.cudaGraphNode_t,driver.CUgraphNode]") cdef cyruntime.cudaGraph_t cygraph if graph is None: pgraph = 0 elif isinstance(graph, (cudaGraph_t,driver.CUgraph)): pgraph = int(graph) else: pgraph = int(cudaGraph_t(graph)) cygraph = pgraph cdef cudaGraphNode_t pGraphNode = cudaGraphNode_t() cdef cyruntime.cudaGraphNode_t* cypDependencies = NULL if len(pDependencies) > 1: cypDependencies = calloc(len(pDependencies), sizeof(cyruntime.cudaGraphNode_t)) if cypDependencies is NULL: raise MemoryError('Failed to allocate length x size memory: ' + str(len(pDependencies)) + 'x' + str(sizeof(cyruntime.cudaGraphNode_t))) else: for idx in range(len(pDependencies)): cypDependencies[idx] = (pDependencies[idx])._pvt_ptr[0] elif len(pDependencies) == 1: cypDependencies = (pDependencies[0])._pvt_ptr cdef cyruntime.cudaGraphEdgeData* cydependencyData = NULL if len(dependencyData) > 1: cydependencyData = calloc(len(dependencyData), sizeof(cyruntime.cudaGraphEdgeData)) if cydependencyData is NULL: raise MemoryError('Failed to allocate length x size memory: ' + str(len(dependencyData)) + 'x' + str(sizeof(cyruntime.cudaGraphEdgeData))) for idx in range(len(dependencyData)): string.memcpy(&cydependencyData[idx], (dependencyData[idx])._pvt_ptr, sizeof(cyruntime.cudaGraphEdgeData)) elif len(dependencyData) == 1: cydependencyData = (dependencyData[0])._pvt_ptr if numDependencies > len(pDependencies): raise RuntimeError("List is too small: " + str(len(pDependencies)) + " < " + str(numDependencies)) if numDependencies > len(dependencyData): raise RuntimeError("List is too small: " + str(len(dependencyData)) + " < " + str(numDependencies)) cdef cyruntime.cudaGraphNodeParams* cynodeParams_ptr = nodeParams._pvt_ptr if nodeParams is not None else NULL with nogil: err = cyruntime.cudaGraphAddNode_v2(pGraphNode._pvt_ptr, cygraph, cypDependencies, cydependencyData, numDependencies, cynodeParams_ptr) if len(pDependencies) > 1 and cypDependencies is not NULL: free(cypDependencies) if len(dependencyData) > 1 and cydependencyData is not NULL: free(cydependencyData) if err != cyruntime.cudaSuccess: return (_dict_cudaError_t[err], None) return (_dict_cudaError_t[err], pGraphNode) @cython.embedsignature(True) def cudaGraphNodeSetParams(node, nodeParams : Optional[cudaGraphNodeParams]): """ Update's a graph node's parameters. Sets the parameters of graph node `node` to `nodeParams`. The node type specified by `nodeParams->type` must match the type of `node`. `nodeParams` must be fully initialized and all unused bytes (reserved, padding) zeroed. Modifying parameters is not supported for node types cudaGraphNodeTypeMemAlloc and cudaGraphNodeTypeMemFree. Parameters ---------- node : :py:obj:`~.CUgraphNode` or :py:obj:`~.cudaGraphNode_t` Node to set the parameters for nodeParams : :py:obj:`~.cudaGraphNodeParams` Parameters to copy Returns ------- cudaError_t :py:obj:`~.cudaSuccess`, :py:obj:`~.cudaErrorInvalidValue`, :py:obj:`~.cudaErrorInvalidDeviceFunction`, :py:obj:`~.cudaErrorNotSupported` See Also -------- :py:obj:`~.cudaGraphAddNode`, :py:obj:`~.cudaGraphExecNodeSetParams` """ cdef cyruntime.cudaGraphNode_t cynode if node is None: pnode = 0 elif isinstance(node, (cudaGraphNode_t,driver.CUgraphNode)): pnode = int(node) else: pnode = int(cudaGraphNode_t(node)) cynode = pnode cdef cyruntime.cudaGraphNodeParams* cynodeParams_ptr = nodeParams._pvt_ptr if nodeParams is not None else NULL with nogil: err = cyruntime.cudaGraphNodeSetParams(cynode, cynodeParams_ptr) return (_dict_cudaError_t[err],) @cython.embedsignature(True) def cudaGraphExecNodeSetParams(graphExec, node, nodeParams : Optional[cudaGraphNodeParams]): """ Update's a graph node's parameters in an instantiated graph. Sets the parameters of a node in an executable graph `graphExec`. The node is identified by the corresponding node `node` in the non- executable graph from which the executable graph was instantiated. `node` must not have been removed from the original graph. The modifications only affect future launches of `graphExec`. Already enqueued or running launches of `graphExec` are not affected by this call. `node` is also not modified by this call. Allowed changes to parameters on executable graphs are as follows: **View CUDA Toolkit Documentation for a table example** Parameters ---------- graphExec : :py:obj:`~.CUgraphExec` or :py:obj:`~.cudaGraphExec_t` The executable graph in which to update the specified node node : :py:obj:`~.CUgraphNode` or :py:obj:`~.cudaGraphNode_t` Corresponding node from the graph from which graphExec was instantiated nodeParams : :py:obj:`~.cudaGraphNodeParams` Updated Parameters to set Returns ------- cudaError_t :py:obj:`~.cudaSuccess`, :py:obj:`~.cudaErrorInvalidValue`, :py:obj:`~.cudaErrorInvalidDeviceFunction`, :py:obj:`~.cudaErrorNotSupported` See Also -------- :py:obj:`~.cudaGraphAddNode`, :py:obj:`~.cudaGraphNodeSetParams` :py:obj:`~.cudaGraphExecUpdate`, :py:obj:`~.cudaGraphInstantiate` """ cdef cyruntime.cudaGraphNode_t cynode if node is None: pnode = 0 elif isinstance(node, (cudaGraphNode_t,driver.CUgraphNode)): pnode = int(node) else: pnode = int(cudaGraphNode_t(node)) cynode = pnode cdef cyruntime.cudaGraphExec_t cygraphExec if graphExec is None: pgraphExec = 0 elif isinstance(graphExec, (cudaGraphExec_t,driver.CUgraphExec)): pgraphExec = int(graphExec) else: pgraphExec = int(cudaGraphExec_t(graphExec)) cygraphExec = pgraphExec cdef cyruntime.cudaGraphNodeParams* cynodeParams_ptr = nodeParams._pvt_ptr if nodeParams is not None else NULL with nogil: err = cyruntime.cudaGraphExecNodeSetParams(cygraphExec, cynode, cynodeParams_ptr) return (_dict_cudaError_t[err],) @cython.embedsignature(True) def cudaGraphConditionalHandleCreate(graph, unsigned int defaultLaunchValue, unsigned int flags): """ Create a conditional handle. Creates a conditional handle associated with `hGraph`. The conditional handle must be associated with a conditional node in this graph or one of its children. Handles not associated with a conditional node may cause graph instantiation to fail. Parameters ---------- hGraph : :py:obj:`~.CUgraph` or :py:obj:`~.cudaGraph_t` Graph which will contain the conditional node using this handle. defaultLaunchValue : unsigned int Optional initial value for the conditional variable. Applied at the beginning of each graph execution if cudaGraphCondAssignDefault is set in `flags`. flags : unsigned int Currently must be cudaGraphCondAssignDefault or 0. Returns ------- cudaError_t :py:obj:`~.CUDA_SUCCESS`, :py:obj:`~.CUDA_ERROR_INVALID_VALUE`, :py:obj:`~.CUDA_ERROR_NOT_SUPPORTED` pHandle_out : :py:obj:`~.cudaGraphConditionalHandle` Pointer used to return the handle to the caller. See Also -------- :py:obj:`~.cuGraphAddNode`, """ cdef cyruntime.cudaGraph_t cygraph if graph is None: pgraph = 0 elif isinstance(graph, (cudaGraph_t,driver.CUgraph)): pgraph = int(graph) else: pgraph = int(cudaGraph_t(graph)) cygraph = pgraph cdef cudaGraphConditionalHandle pHandle_out = cudaGraphConditionalHandle() with nogil: err = cyruntime.cudaGraphConditionalHandleCreate(pHandle_out._pvt_ptr, cygraph, defaultLaunchValue, flags) if err != cyruntime.cudaSuccess: return (_dict_cudaError_t[err], None) return (_dict_cudaError_t[err], pHandle_out) @cython.embedsignature(True) def cudaGetDriverEntryPoint(char* symbol, unsigned long long flags): """ Returns the requested driver API function pointer. Returns in `**funcPtr` the address of the CUDA driver function for the requested flags. For a requested driver symbol, if the CUDA version in which the driver symbol was introduced is less than or equal to the CUDA runtime version, the API will return the function pointer to the corresponding versioned driver function. The pointer returned by the API should be cast to a function pointer matching the requested driver function's definition in the API header file. The function pointer typedef can be picked up from the corresponding typedefs header file. For example, cudaTypedefs.h consists of function pointer typedefs for driver APIs defined in cuda.h. The API will return :py:obj:`~.cudaSuccess` and set the returned `funcPtr` if the requested driver function is valid and supported on the platform. The API will return :py:obj:`~.cudaSuccess` and set the returned `funcPtr` to NULL if the requested driver function is not supported on the platform, no ABI compatible driver function exists for the CUDA runtime version or if the driver symbol is invalid. It will also set the optional `driverStatus` to one of the values in :py:obj:`~.cudaDriverEntryPointQueryResult` with the following meanings: - :py:obj:`~.cudaDriverEntryPointSuccess` - The requested symbol was succesfully found based on input arguments and `pfn` is valid - :py:obj:`~.cudaDriverEntryPointSymbolNotFound` - The requested symbol was not found - :py:obj:`~.cudaDriverEntryPointVersionNotSufficent` - The requested symbol was found but is not supported by the current runtime version (CUDART_VERSION) The requested flags can be: - :py:obj:`~.cudaEnableDefault`: This is the default mode. This is equivalent to :py:obj:`~.cudaEnablePerThreadDefaultStream` if the code is compiled with --default-stream per-thread compilation flag or the macro CUDA_API_PER_THREAD_DEFAULT_STREAM is defined; :py:obj:`~.cudaEnableLegacyStream` otherwise. - :py:obj:`~.cudaEnableLegacyStream`: This will enable the search for all driver symbols that match the requested driver symbol name except the corresponding per-thread versions. - :py:obj:`~.cudaEnablePerThreadDefaultStream`: This will enable the search for all driver symbols that match the requested driver symbol name including the per-thread versions. If a per-thread version is not found, the API will return the legacy version of the driver function. Parameters ---------- symbol : bytes The base name of the driver API function to look for. As an example, for the driver API :py:obj:`~.cuMemAlloc_v2`, `symbol` would be cuMemAlloc. Note that the API will use the CUDA runtime version to return the address to the most recent ABI compatible driver symbol, :py:obj:`~.cuMemAlloc` or :py:obj:`~.cuMemAlloc_v2`. flags : unsigned long long Flags to specify search options. Returns ------- cudaError_t :py:obj:`~.cudaSuccess`, :py:obj:`~.cudaErrorInvalidValue`, :py:obj:`~.cudaErrorNotSupported` funcPtr : Any Location to return the function pointer to the requested driver function driverStatus : :py:obj:`~.cudaDriverEntryPointQueryResult` Optional location to store the status of finding the symbol from the driver. See :py:obj:`~.cudaDriverEntryPointQueryResult` for possible values. See Also -------- :py:obj:`~.cuGetProcAddress` """ cdef void_ptr funcPtr = 0 cdef cyruntime.cudaDriverEntryPointQueryResult driverStatus with nogil: err = cyruntime.cudaGetDriverEntryPoint(symbol, &funcPtr, flags, &driverStatus) if err != cyruntime.cudaSuccess: return (_dict_cudaError_t[err], None, None) return (_dict_cudaError_t[err], funcPtr, cudaDriverEntryPointQueryResult(driverStatus)) @cython.embedsignature(True) def cudaGetDriverEntryPointByVersion(char* symbol, unsigned int cudaVersion, unsigned long long flags): """ Returns the requested driver API function pointer by CUDA version. Returns in `**funcPtr` the address of the CUDA driver function for the requested flags and CUDA driver version. The CUDA version is specified as (1000 * major + 10 * minor), so CUDA 11.2 should be specified as 11020. For a requested driver symbol, if the specified CUDA version is greater than or equal to the CUDA version in which the driver symbol was introduced, this API will return the function pointer to the corresponding versioned function. The pointer returned by the API should be cast to a function pointer matching the requested driver function's definition in the API header file. The function pointer typedef can be picked up from the corresponding typedefs header file. For example, cudaTypedefs.h consists of function pointer typedefs for driver APIs defined in cuda.h. For the case where the CUDA version requested is greater than the CUDA Toolkit installed, there may not be an appropriate function pointer typedef in the corresponding header file and may need a custom typedef to match the driver function signature returned. This can be done by getting the typedefs from a later toolkit or creating appropriately matching custom function typedefs. The API will return :py:obj:`~.cudaSuccess` and set the returned `funcPtr` if the requested driver function is valid and supported on the platform. The API will return :py:obj:`~.cudaSuccess` and set the returned `funcPtr` to NULL if the requested driver function is not supported on the platform, no ABI compatible driver function exists for the requested version or if the driver symbol is invalid. It will also set the optional `driverStatus` to one of the values in :py:obj:`~.cudaDriverEntryPointQueryResult` with the following meanings: - :py:obj:`~.cudaDriverEntryPointSuccess` - The requested symbol was succesfully found based on input arguments and `pfn` is valid - :py:obj:`~.cudaDriverEntryPointSymbolNotFound` - The requested symbol was not found - :py:obj:`~.cudaDriverEntryPointVersionNotSufficent` - The requested symbol was found but is not supported by the specified version `cudaVersion` The requested flags can be: - :py:obj:`~.cudaEnableDefault`: This is the default mode. This is equivalent to :py:obj:`~.cudaEnablePerThreadDefaultStream` if the code is compiled with --default-stream per-thread compilation flag or the macro CUDA_API_PER_THREAD_DEFAULT_STREAM is defined; :py:obj:`~.cudaEnableLegacyStream` otherwise. - :py:obj:`~.cudaEnableLegacyStream`: This will enable the search for all driver symbols that match the requested driver symbol name except the corresponding per-thread versions. - :py:obj:`~.cudaEnablePerThreadDefaultStream`: This will enable the search for all driver symbols that match the requested driver symbol name including the per-thread versions. If a per-thread version is not found, the API will return the legacy version of the driver function. Parameters ---------- symbol : bytes The base name of the driver API function to look for. As an example, for the driver API :py:obj:`~.cuMemAlloc_v2`, `symbol` would be cuMemAlloc. cudaVersion : unsigned int The CUDA version to look for the requested driver symbol flags : unsigned long long Flags to specify search options. Returns ------- cudaError_t :py:obj:`~.cudaSuccess`, :py:obj:`~.cudaErrorInvalidValue`, :py:obj:`~.cudaErrorNotSupported` funcPtr : Any Location to return the function pointer to the requested driver function driverStatus : :py:obj:`~.cudaDriverEntryPointQueryResult` Optional location to store the status of finding the symbol from the driver. See :py:obj:`~.cudaDriverEntryPointQueryResult` for possible values. See Also -------- :py:obj:`~.cuGetProcAddress` """ cdef void_ptr funcPtr = 0 cdef cyruntime.cudaDriverEntryPointQueryResult driverStatus with nogil: err = cyruntime.cudaGetDriverEntryPointByVersion(symbol, &funcPtr, cudaVersion, flags, &driverStatus) if err != cyruntime.cudaSuccess: return (_dict_cudaError_t[err], None, None) return (_dict_cudaError_t[err], funcPtr, cudaDriverEntryPointQueryResult(driverStatus)) @cython.embedsignature(True) def cudaLibraryLoadData(code, jitOptions : Optional[tuple[cudaJitOption] | list[cudaJitOption]], jitOptionsValues : Optional[tuple[Any] | list[Any]], unsigned int numJitOptions, libraryOptions : Optional[tuple[cudaLibraryOption] | list[cudaLibraryOption]], libraryOptionValues : Optional[tuple[Any] | list[Any]], unsigned int numLibraryOptions): """ Load a library with specified code and options. Takes a pointer `code` and loads the corresponding library `library` based on the application defined library loading mode: - If module loading is set to EAGER, via the environment variables described in "Module loading", `library` is loaded eagerly into all contexts at the time of the call and future contexts at the time of creation until the library is unloaded with :py:obj:`~.cudaLibraryUnload()`. - If the environment variables are set to LAZY, `library` is not immediately loaded onto all existent contexts and will only be loaded when a function is needed for that context, such as a kernel launch. These environment variables are described in the CUDA programming guide under the "CUDA environment variables" section. The `code` may be a `cubin` or `fatbin` as output by nvcc, or a NULL- terminated `PTX`, either as output by nvcc or hand-written. A fatbin should also contain relocatable code when doing separate compilation. Please also see the documentation for nvrtc (https://docs.nvidia.com/cuda/nvrtc/index.html), nvjitlink (https://docs.nvidia.com/cuda/nvjitlink/index.html), and nvfatbin (https://docs.nvidia.com/cuda/nvfatbin/index.html) for more information on generating loadable code at runtime. Options are passed as an array via `jitOptions` and any corresponding parameters are passed in `jitOptionsValues`. The number of total JIT options is supplied via `numJitOptions`. Any outputs will be returned via `jitOptionsValues`. Library load options are passed as an array via `libraryOptions` and any corresponding parameters are passed in `libraryOptionValues`. The number of total library load options is supplied via `numLibraryOptions`. Parameters ---------- code : Any Code to load jitOptions : list[:py:obj:`~.cudaJitOption`] Options for JIT jitOptionsValues : list[Any] Option values for JIT numJitOptions : unsigned int Number of options libraryOptions : list[:py:obj:`~.cudaLibraryOption`] Options for loading libraryOptionValues : list[Any] Option values for loading numLibraryOptions : unsigned int Number of options for loading Returns ------- cudaError_t :py:obj:`~.cudaSuccess`, :py:obj:`~.cudaErrorInvalidValue`, :py:obj:`~.cudaErrorMemoryAllocation`, :py:obj:`~.cudaErrorInitializationError`, :py:obj:`~.cudaErrorCudartUnloading`, :py:obj:`~.cudaErrorInvalidPtx`, :py:obj:`~.cudaErrorUnsupportedPtxVersion`, :py:obj:`~.cudaErrorNoKernelImageForDevice`, :py:obj:`~.cudaErrorSharedObjectSymbolNotFound`, :py:obj:`~.cudaErrorSharedObjectInitFailed`, :py:obj:`~.cudaErrorJitCompilerNotFound` library : :py:obj:`~.cudaLibrary_t` Returned library See Also -------- :py:obj:`~.cudaLibraryLoadFromFile`, :py:obj:`~.cudaLibraryUnload`, :py:obj:`~.cuLibraryLoadData` """ libraryOptionValues = [] if libraryOptionValues is None else libraryOptionValues libraryOptions = [] if libraryOptions is None else libraryOptions if not all(isinstance(_x, (cudaLibraryOption)) for _x in libraryOptions): raise TypeError("Argument 'libraryOptions' is not instance of type (expected tuple[cyruntime.cudaLibraryOption] or list[cyruntime.cudaLibraryOption]") jitOptionsValues = [] if jitOptionsValues is None else jitOptionsValues jitOptions = [] if jitOptions is None else jitOptions if not all(isinstance(_x, (cudaJitOption)) for _x in jitOptions): raise TypeError("Argument 'jitOptions' is not instance of type (expected tuple[cyruntime.cudaJitOption] or list[cyruntime.cudaJitOption]") cdef cudaLibrary_t library = cudaLibrary_t() cycode = _HelperInputVoidPtr(code) cdef void* cycode_ptr = cycode.cptr cdef vector[cyruntime.cudaJitOption] cyjitOptions = [pyjitOptions.value for pyjitOptions in (jitOptions)] pylist = [_HelperCudaJitOption(pyoptions, pyoptionValues) for pyoptions, pyoptionValues in zip(jitOptions, jitOptionsValues)] cdef _InputVoidPtrPtrHelper voidStarHelperjitOptionsValues = _InputVoidPtrPtrHelper(pylist) cdef void** cyjitOptionsValues_ptr = voidStarHelperjitOptionsValues.cptr if numJitOptions > len(jitOptions): raise RuntimeError("List is too small: " + str(len(jitOptions)) + " < " + str(numJitOptions)) if numJitOptions > len(jitOptionsValues): raise RuntimeError("List is too small: " + str(len(jitOptionsValues)) + " < " + str(numJitOptions)) cdef vector[cyruntime.cudaLibraryOption] cylibraryOptions = [pylibraryOptions.value for pylibraryOptions in (libraryOptions)] pylist = [_HelperCudaLibraryOption(pyoptions, pyoptionValues) for pyoptions, pyoptionValues in zip(libraryOptions, libraryOptionValues)] cdef _InputVoidPtrPtrHelper voidStarHelperlibraryOptionValues = _InputVoidPtrPtrHelper(pylist) cdef void** cylibraryOptionValues_ptr = voidStarHelperlibraryOptionValues.cptr if numLibraryOptions > len(libraryOptions): raise RuntimeError("List is too small: " + str(len(libraryOptions)) + " < " + str(numLibraryOptions)) if numLibraryOptions > len(libraryOptionValues): raise RuntimeError("List is too small: " + str(len(libraryOptionValues)) + " < " + str(numLibraryOptions)) with nogil: err = cyruntime.cudaLibraryLoadData(library._pvt_ptr, cycode_ptr, cyjitOptions.data(), cyjitOptionsValues_ptr, numJitOptions, cylibraryOptions.data(), cylibraryOptionValues_ptr, numLibraryOptions) if err != cyruntime.cudaSuccess: return (_dict_cudaError_t[err], None) return (_dict_cudaError_t[err], library) @cython.embedsignature(True) def cudaLibraryLoadFromFile(char* fileName, jitOptions : Optional[tuple[cudaJitOption] | list[cudaJitOption]], jitOptionsValues : Optional[tuple[Any] | list[Any]], unsigned int numJitOptions, libraryOptions : Optional[tuple[cudaLibraryOption] | list[cudaLibraryOption]], libraryOptionValues : Optional[tuple[Any] | list[Any]], unsigned int numLibraryOptions): """ Load a library with specified file and options. Takes a pointer `code` and loads the corresponding library `library` based on the application defined library loading mode: - If module loading is set to EAGER, via the environment variables described in "Module loading", `library` is loaded eagerly into all contexts at the time of the call and future contexts at the time of creation until the library is unloaded with :py:obj:`~.cudaLibraryUnload()`. - If the environment variables are set to LAZY, `library` is not immediately loaded onto all existent contexts and will only be loaded when a function is needed for that context, such as a kernel launch. These environment variables are described in the CUDA programming guide under the "CUDA environment variables" section. The file should be a `cubin` file as output by nvcc, or a `PTX` file either as output by nvcc or handwritten, or a `fatbin` file as output by nvcc. A fatbin should also contain relocatable code when doing separate compilation. Please also see the documentation for nvrtc (https://docs.nvidia.com/cuda/nvrtc/index.html), nvjitlink (https://docs.nvidia.com/cuda/nvjitlink/index.html), and nvfatbin (https://docs.nvidia.com/cuda/nvfatbin/index.html) for more information on generating loadable code at runtime. Options are passed as an array via `jitOptions` and any corresponding parameters are passed in `jitOptionsValues`. The number of total options is supplied via `numJitOptions`. Any outputs will be returned via `jitOptionsValues`. Library load options are passed as an array via `libraryOptions` and any corresponding parameters are passed in `libraryOptionValues`. The number of total library load options is supplied via `numLibraryOptions`. Parameters ---------- fileName : bytes File to load from jitOptions : list[:py:obj:`~.cudaJitOption`] Options for JIT jitOptionsValues : list[Any] Option values for JIT numJitOptions : unsigned int Number of options libraryOptions : list[:py:obj:`~.cudaLibraryOption`] Options for loading libraryOptionValues : list[Any] Option values for loading numLibraryOptions : unsigned int Number of options for loading Returns ------- cudaError_t :py:obj:`~.cudaSuccess`, :py:obj:`~.cudaErrorInvalidValue`, :py:obj:`~.cudaErrorMemoryAllocation`, :py:obj:`~.cudaErrorInitializationError`, :py:obj:`~.cudaErrorCudartUnloading`, :py:obj:`~.cudaErrorInvalidPtx`, :py:obj:`~.cudaErrorUnsupportedPtxVersion`, :py:obj:`~.cudaErrorNoKernelImageForDevice`, :py:obj:`~.cudaErrorSharedObjectSymbolNotFound`, :py:obj:`~.cudaErrorSharedObjectInitFailed`, :py:obj:`~.cudaErrorJitCompilerNotFound` library : :py:obj:`~.cudaLibrary_t` Returned library See Also -------- :py:obj:`~.cudaLibraryLoadData`, :py:obj:`~.cudaLibraryUnload`, :py:obj:`~.cuLibraryLoadFromFile` """ libraryOptionValues = [] if libraryOptionValues is None else libraryOptionValues libraryOptions = [] if libraryOptions is None else libraryOptions if not all(isinstance(_x, (cudaLibraryOption)) for _x in libraryOptions): raise TypeError("Argument 'libraryOptions' is not instance of type (expected tuple[cyruntime.cudaLibraryOption] or list[cyruntime.cudaLibraryOption]") jitOptionsValues = [] if jitOptionsValues is None else jitOptionsValues jitOptions = [] if jitOptions is None else jitOptions if not all(isinstance(_x, (cudaJitOption)) for _x in jitOptions): raise TypeError("Argument 'jitOptions' is not instance of type (expected tuple[cyruntime.cudaJitOption] or list[cyruntime.cudaJitOption]") cdef cudaLibrary_t library = cudaLibrary_t() cdef vector[cyruntime.cudaJitOption] cyjitOptions = [pyjitOptions.value for pyjitOptions in (jitOptions)] pylist = [_HelperCudaJitOption(pyoptions, pyoptionValues) for pyoptions, pyoptionValues in zip(jitOptions, jitOptionsValues)] cdef _InputVoidPtrPtrHelper voidStarHelperjitOptionsValues = _InputVoidPtrPtrHelper(pylist) cdef void** cyjitOptionsValues_ptr = voidStarHelperjitOptionsValues.cptr if numJitOptions > len(jitOptions): raise RuntimeError("List is too small: " + str(len(jitOptions)) + " < " + str(numJitOptions)) if numJitOptions > len(jitOptionsValues): raise RuntimeError("List is too small: " + str(len(jitOptionsValues)) + " < " + str(numJitOptions)) cdef vector[cyruntime.cudaLibraryOption] cylibraryOptions = [pylibraryOptions.value for pylibraryOptions in (libraryOptions)] pylist = [_HelperCudaLibraryOption(pyoptions, pyoptionValues) for pyoptions, pyoptionValues in zip(libraryOptions, libraryOptionValues)] cdef _InputVoidPtrPtrHelper voidStarHelperlibraryOptionValues = _InputVoidPtrPtrHelper(pylist) cdef void** cylibraryOptionValues_ptr = voidStarHelperlibraryOptionValues.cptr if numLibraryOptions > len(libraryOptions): raise RuntimeError("List is too small: " + str(len(libraryOptions)) + " < " + str(numLibraryOptions)) if numLibraryOptions > len(libraryOptionValues): raise RuntimeError("List is too small: " + str(len(libraryOptionValues)) + " < " + str(numLibraryOptions)) with nogil: err = cyruntime.cudaLibraryLoadFromFile(library._pvt_ptr, fileName, cyjitOptions.data(), cyjitOptionsValues_ptr, numJitOptions, cylibraryOptions.data(), cylibraryOptionValues_ptr, numLibraryOptions) if err != cyruntime.cudaSuccess: return (_dict_cudaError_t[err], None) return (_dict_cudaError_t[err], library) @cython.embedsignature(True) def cudaLibraryUnload(library): """ Unloads a library. Unloads the library specified with `library` Parameters ---------- library : :py:obj:`~.cudaLibrary_t` Library to unload Returns ------- cudaError_t :py:obj:`~.cudaSuccess`, :py:obj:`~.cudaErrorCudartUnloading`, :py:obj:`~.cudaErrorInitializationError`, :py:obj:`~.cudaErrorInvalidValue` See Also -------- :py:obj:`~.cudaLibraryLoadData`, :py:obj:`~.cudaLibraryLoadFromFile`, :py:obj:`~.cuLibraryUnload` """ cdef cyruntime.cudaLibrary_t cylibrary if library is None: plibrary = 0 elif isinstance(library, (cudaLibrary_t,)): plibrary = int(library) else: plibrary = int(cudaLibrary_t(library)) cylibrary = plibrary with nogil: err = cyruntime.cudaLibraryUnload(cylibrary) return (_dict_cudaError_t[err],) @cython.embedsignature(True) def cudaLibraryGetKernel(library, char* name): """ Returns a kernel handle. Returns in `pKernel` the handle of the kernel with name `name` located in library `library`. If kernel handle is not found, the call returns :py:obj:`~.cudaErrorSymbolNotFound`. Parameters ---------- library : :py:obj:`~.cudaLibrary_t` Library to retrieve kernel from name : bytes Name of kernel to retrieve Returns ------- cudaError_t :py:obj:`~.cudaSuccess`, :py:obj:`~.cudaErrorCudartUnloading`, :py:obj:`~.cudaErrorInitializationError`, :py:obj:`~.cudaErrorInvalidValue`, :py:obj:`~.cudaErrorInvalidResourceHandle`, :py:obj:`~.cudaErrorSymbolNotFound` pKernel : :py:obj:`~.cudaKernel_t` Returned kernel handle See Also -------- :py:obj:`~.cudaLibraryLoadData`, :py:obj:`~.cudaLibraryLoadFromFile`, :py:obj:`~.cudaLibraryUnload`, :py:obj:`~.cuLibraryGetKernel` """ cdef cyruntime.cudaLibrary_t cylibrary if library is None: plibrary = 0 elif isinstance(library, (cudaLibrary_t,)): plibrary = int(library) else: plibrary = int(cudaLibrary_t(library)) cylibrary = plibrary cdef cudaKernel_t pKernel = cudaKernel_t() with nogil: err = cyruntime.cudaLibraryGetKernel(pKernel._pvt_ptr, cylibrary, name) if err != cyruntime.cudaSuccess: return (_dict_cudaError_t[err], None) return (_dict_cudaError_t[err], pKernel) @cython.embedsignature(True) def cudaLibraryGetGlobal(library, char* name): """ Returns a global device pointer. Returns in `*dptr` and `*bytes` the base pointer and size of the global with name `name` for the requested library `library` and the current device. If no global for the requested name `name` exists, the call returns :py:obj:`~.cudaErrorSymbolNotFound`. One of the parameters `dptr` or `numbytes` (not both) can be NULL in which case it is ignored. The returned `dptr` cannot be passed to the Symbol APIs such as :py:obj:`~.cudaMemcpyToSymbol`, :py:obj:`~.cudaMemcpyFromSymbol`, :py:obj:`~.cudaGetSymbolAddress`, or :py:obj:`~.cudaGetSymbolSize`. Parameters ---------- library : :py:obj:`~.cudaLibrary_t` Library to retrieve global from name : bytes Name of global to retrieve Returns ------- cudaError_t :py:obj:`~.cudaSuccess`, :py:obj:`~.cudaErrorCudartUnloading`, :py:obj:`~.cudaErrorInitializationError`, :py:obj:`~.cudaErrorInvalidValue`, :py:obj:`~.cudaErrorInvalidResourceHandle`, :py:obj:`~.cudaErrorSymbolNotFound` :py:obj:`~.cudaErrorDeviceUninitialized`, :py:obj:`~.cudaErrorContextIsDestroyed` dptr : Any Returned global device pointer for the requested library numbytes : int Returned global size in bytes See Also -------- :py:obj:`~.cudaLibraryLoadData`, :py:obj:`~.cudaLibraryLoadFromFile`, :py:obj:`~.cudaLibraryUnload`, :py:obj:`~.cudaLibraryGetManaged`, :py:obj:`~.cuLibraryGetGlobal` """ cdef cyruntime.cudaLibrary_t cylibrary if library is None: plibrary = 0 elif isinstance(library, (cudaLibrary_t,)): plibrary = int(library) else: plibrary = int(cudaLibrary_t(library)) cylibrary = plibrary cdef void_ptr dptr = 0 cdef size_t numbytes = 0 with nogil: err = cyruntime.cudaLibraryGetGlobal(&dptr, &numbytes, cylibrary, name) if err != cyruntime.cudaSuccess: return (_dict_cudaError_t[err], None, None) return (_dict_cudaError_t[err], dptr, numbytes) @cython.embedsignature(True) def cudaLibraryGetManaged(library, char* name): """ Returns a pointer to managed memory. Returns in `*dptr` and `*bytes` the base pointer and size of the managed memory with name `name` for the requested library `library`. If no managed memory with the requested name `name` exists, the call returns :py:obj:`~.cudaErrorSymbolNotFound`. One of the parameters `dptr` or `numbytes` (not both) can be NULL in which case it is ignored. Note that managed memory for library `library` is shared across devices and is registered when the library is loaded. The returned `dptr` cannot be passed to the Symbol APIs such as :py:obj:`~.cudaMemcpyToSymbol`, :py:obj:`~.cudaMemcpyFromSymbol`, :py:obj:`~.cudaGetSymbolAddress`, or :py:obj:`~.cudaGetSymbolSize`. Parameters ---------- library : :py:obj:`~.cudaLibrary_t` Library to retrieve managed memory from name : bytes Name of managed memory to retrieve Returns ------- cudaError_t :py:obj:`~.cudaSuccess`, :py:obj:`~.cudaErrorCudartUnloading`, :py:obj:`~.cudaErrorInitializationError`, :py:obj:`~.cudaErrorInvalidValue`, :py:obj:`~.cudaErrorInvalidResourceHandle`, :py:obj:`~.cudaErrorSymbolNotFound` dptr : Any Returned pointer to the managed memory numbytes : int Returned memory size in bytes See Also -------- :py:obj:`~.cudaLibraryLoadData`, :py:obj:`~.cudaLibraryLoadFromFile`, :py:obj:`~.cudaLibraryUnload`, :py:obj:`~.cudaLibraryGetGlobal`, :py:obj:`~.cuLibraryGetManaged` """ cdef cyruntime.cudaLibrary_t cylibrary if library is None: plibrary = 0 elif isinstance(library, (cudaLibrary_t,)): plibrary = int(library) else: plibrary = int(cudaLibrary_t(library)) cylibrary = plibrary cdef void_ptr dptr = 0 cdef size_t numbytes = 0 with nogil: err = cyruntime.cudaLibraryGetManaged(&dptr, &numbytes, cylibrary, name) if err != cyruntime.cudaSuccess: return (_dict_cudaError_t[err], None, None) return (_dict_cudaError_t[err], dptr, numbytes) @cython.embedsignature(True) def cudaLibraryGetUnifiedFunction(library, char* symbol): """ Returns a pointer to a unified function. Returns in `*fptr` the function pointer to a unified function denoted by `symbol`. If no unified function with name `symbol` exists, the call returns :py:obj:`~.cudaErrorSymbolNotFound`. If there is no device with attribute :py:obj:`~.cudaDeviceProp.unifiedFunctionPointers` present in the system, the call may return :py:obj:`~.cudaErrorSymbolNotFound`. Parameters ---------- library : :py:obj:`~.cudaLibrary_t` Library to retrieve function pointer memory from symbol : bytes Name of function pointer to retrieve Returns ------- cudaError_t :py:obj:`~.cudaSuccess`, :py:obj:`~.cudaErrorCudartUnloading`, :py:obj:`~.cudaErrorInitializationError`, :py:obj:`~.cudaErrorInvalidValue`, :py:obj:`~.cudaErrorInvalidResourceHandle`, :py:obj:`~.cudaErrorSymbolNotFound` fptr : Any Returned pointer to a unified function See Also -------- :py:obj:`~.cudaLibraryLoadData`, :py:obj:`~.cudaLibraryLoadFromFile`, :py:obj:`~.cudaLibraryUnload`, :py:obj:`~.cuLibraryGetUnifiedFunction` """ cdef cyruntime.cudaLibrary_t cylibrary if library is None: plibrary = 0 elif isinstance(library, (cudaLibrary_t,)): plibrary = int(library) else: plibrary = int(cudaLibrary_t(library)) cylibrary = plibrary cdef void_ptr fptr = 0 with nogil: err = cyruntime.cudaLibraryGetUnifiedFunction(&fptr, cylibrary, symbol) if err != cyruntime.cudaSuccess: return (_dict_cudaError_t[err], None) return (_dict_cudaError_t[err], fptr) @cython.embedsignature(True) def cudaLibraryGetKernelCount(lib): """ Returns the number of kernels within a library. Returns in `count` the number of kernels in `lib`. Parameters ---------- lib : :py:obj:`~.cudaLibrary_t` Library to query Returns ------- cudaError_t :py:obj:`~.cudaSuccess`, :py:obj:`~.cudaErrorCudartUnloading`, :py:obj:`~.cudaErrorInitializationError`, :py:obj:`~.cudaErrorInvalidValue`, :py:obj:`~.cudaErrorInvalidResourceHandle` count : unsigned int Number of kernels found within the library See Also -------- :py:obj:`~.cudaLibraryEnumerateKernels`, :py:obj:`~.cudaLibraryLoadFromFile`, :py:obj:`~.cudaLibraryLoadData`, :py:obj:`~.cuLibraryGetKernelCount` """ cdef cyruntime.cudaLibrary_t cylib if lib is None: plib = 0 elif isinstance(lib, (cudaLibrary_t,)): plib = int(lib) else: plib = int(cudaLibrary_t(lib)) cylib = plib cdef unsigned int count = 0 with nogil: err = cyruntime.cudaLibraryGetKernelCount(&count, cylib) if err != cyruntime.cudaSuccess: return (_dict_cudaError_t[err], None) return (_dict_cudaError_t[err], count) @cython.embedsignature(True) def cudaLibraryEnumerateKernels(unsigned int numKernels, lib): """ Retrieve the kernel handles within a library. Returns in `kernels` a maximum number of `numKernels` kernel handles within `lib`. The returned kernel handle becomes invalid when the library is unloaded. Parameters ---------- numKernels : unsigned int Maximum number of kernel handles may be returned to the buffer lib : :py:obj:`~.cudaLibrary_t` Library to query from Returns ------- cudaError_t :py:obj:`~.cudaSuccess`, :py:obj:`~.cudaErrorCudartUnloading`, :py:obj:`~.cudaErrorInitializationError`, :py:obj:`~.cudaErrorInvalidValue`, :py:obj:`~.cudaErrorInvalidResourceHandle` kernels : list[:py:obj:`~.cudaKernel_t`] Buffer where the kernel handles are returned to See Also -------- :py:obj:`~.cudaLibraryGetKernelCount`, :py:obj:`~.cuLibraryEnumerateKernels` """ cdef cyruntime.cudaLibrary_t cylib if lib is None: plib = 0 elif isinstance(lib, (cudaLibrary_t,)): plib = int(lib) else: plib = int(cudaLibrary_t(lib)) cylib = plib cdef cyruntime.cudaKernel_t* cykernels = NULL pykernels = [] if numKernels != 0: cykernels = calloc(numKernels, sizeof(cyruntime.cudaKernel_t)) if cykernels is NULL: raise MemoryError('Failed to allocate length x size memory: ' + str(numKernels) + 'x' + str(sizeof(cyruntime.cudaKernel_t))) with nogil: err = cyruntime.cudaLibraryEnumerateKernels(cykernels, numKernels, cylib) if cudaError_t(err) == cudaError_t(0): pykernels = [cudaKernel_t(init_value=cykernels[idx]) for idx in range(numKernels)] if cykernels is not NULL: free(cykernels) if err != cyruntime.cudaSuccess: return (_dict_cudaError_t[err], None) return (_dict_cudaError_t[err], pykernels) @cython.embedsignature(True) def cudaKernelSetAttributeForDevice(kernel, attr not None : cudaFuncAttribute, int value, int device): """ Sets information about a kernel. This call sets the value of a specified attribute `attr` on the kernel `kernel` for the requested device `device` to an integer value specified by `value`. This function returns :py:obj:`~.cudaSuccess` if the new value of the attribute could be successfully set. If the set fails, this call will return an error. Not all attributes can have values set. Attempting to set a value on a read-only attribute will result in an error (:py:obj:`~.cudaErrorInvalidValue`) Note that attributes set using :py:obj:`~.cudaFuncSetAttribute()` will override the attribute set by this API irrespective of whether the call to :py:obj:`~.cudaFuncSetAttribute()` is made before or after this API call. Because of this and the stricter locking requirements mentioned below it is suggested that this call be used during the initialization path and not on each thread accessing `kernel` such as on kernel launches or on the critical path. Valid values for `attr` are: - :py:obj:`~.cudaFuncAttributeMaxDynamicSharedMemorySize` - The requested maximum size in bytes of dynamically-allocated shared memory. The sum of this value and the function attribute :py:obj:`~.sharedSizeBytes` cannot exceed the device attribute :py:obj:`~.cudaDevAttrMaxSharedMemoryPerBlockOptin`. The maximal size of requestable dynamic shared memory may differ by GPU architecture. - :py:obj:`~.cudaFuncAttributePreferredSharedMemoryCarveout` - On devices where the L1 cache and shared memory use the same hardware resources, this sets the shared memory carveout preference, in percent of the total shared memory. See :py:obj:`~.cudaDevAttrMaxSharedMemoryPerMultiprocessor`. This is only a hint, and the driver can choose a different ratio if required to execute the function. - :py:obj:`~.cudaFuncAttributeRequiredClusterWidth`: The required cluster width in blocks. The width, height, and depth values must either all be 0 or all be positive. The validity of the cluster dimensions is checked at launch time. If the value is set during compile time, it cannot be set at runtime. Setting it at runtime will return cudaErrorNotPermitted. - :py:obj:`~.cudaFuncAttributeRequiredClusterHeight`: The required cluster height in blocks. The width, height, and depth values must either all be 0 or all be positive. The validity of the cluster dimensions is checked at launch time. If the value is set during compile time, it cannot be set at runtime. Setting it at runtime will return cudaErrorNotPermitted. - :py:obj:`~.cudaFuncAttributeRequiredClusterDepth`: The required cluster depth in blocks. The width, height, and depth values must either all be 0 or all be positive. The validity of the cluster dimensions is checked at launch time. If the value is set during compile time, it cannot be set at runtime. Setting it at runtime will return cudaErrorNotPermitted. - :py:obj:`~.cudaFuncAttributeNonPortableClusterSizeAllowed`: Indicates whether the function can be launched with non-portable cluster size. 1 is allowed, 0 is disallowed. - :py:obj:`~.cudaFuncAttributeClusterSchedulingPolicyPreference`: The block scheduling policy of a function. The value type is cudaClusterSchedulingPolicy. Parameters ---------- kernel : :py:obj:`~.cudaKernel_t` Kernel to set attribute of attr : :py:obj:`~.cudaFuncAttribute` Attribute requested value : int Value to set device : int Device to set attribute of Returns ------- cudaError_t :py:obj:`~.cudaSuccess`, :py:obj:`~.cudaErrorInvalidDeviceFunction`, :py:obj:`~.cudaErrorInvalidValue` See Also -------- :py:obj:`~.cudaLibraryLoadData`, :py:obj:`~.cudaLibraryLoadFromFile`, :py:obj:`~.cudaLibraryUnload`, :py:obj:`~.cudaLibraryGetKernel`, :py:obj:`~.cudaLaunchKernel`, :py:obj:`~.cudaFuncSetAttribute`, :py:obj:`~.cuKernelSetAttribute` Notes ----- The API has stricter locking requirements in comparison to its legacy counterpart :py:obj:`~.cudaFuncSetAttribute()` due to device-wide semantics. If multiple threads are trying to set the same attribute on the same device simultaneously, the attribute setting will depend on the interleavings chosen by the OS scheduler and memory consistency. """ cdef cyruntime.cudaKernel_t cykernel if kernel is None: pkernel = 0 elif isinstance(kernel, (cudaKernel_t,)): pkernel = int(kernel) else: pkernel = int(cudaKernel_t(kernel)) cykernel = pkernel cdef cyruntime.cudaFuncAttribute cyattr = attr.value with nogil: err = cyruntime.cudaKernelSetAttributeForDevice(cykernel, cyattr, value, device) return (_dict_cudaError_t[err],) @cython.embedsignature(True) def cudaGetExportTable(pExportTableId : Optional[cudaUUID_t]): """""" cdef void_ptr ppExportTable = 0 cdef cyruntime.cudaUUID_t* cypExportTableId_ptr = pExportTableId._pvt_ptr if pExportTableId is not None else NULL with nogil: err = cyruntime.cudaGetExportTable(&ppExportTable, cypExportTableId_ptr) if err != cyruntime.cudaSuccess: return (_dict_cudaError_t[err], None) return (_dict_cudaError_t[err], ppExportTable) @cython.embedsignature(True) def cudaGetKernel(entryFuncAddr): """ Get pointer to device kernel that matches entry function `entryFuncAddr`. Returns in `kernelPtr` the device kernel corresponding to the entry function `entryFuncAddr`. Note that it is possible that there are multiple symbols belonging to different translation units with the same `entryFuncAddr` registered with this CUDA Runtime and so the order which the translation units are loaded and registered with the CUDA Runtime can lead to differing return pointers in `kernelPtr` . Suggested methods of ensuring uniqueness are to limit visibility of global device functions by using static or hidden visibility attribute in the respective translation units. Parameters ---------- entryFuncAddr : Any Address of device entry function to search kernel for Returns ------- cudaError_t :py:obj:`~.cudaSuccess` kernelPtr : :py:obj:`~.cudaKernel_t` Returns the device kernel See Also -------- cudaGetKernel (C++ API) """ cdef cudaKernel_t kernelPtr = cudaKernel_t() cyentryFuncAddr = _HelperInputVoidPtr(entryFuncAddr) cdef void* cyentryFuncAddr_ptr = cyentryFuncAddr.cptr with nogil: err = cyruntime.cudaGetKernel(kernelPtr._pvt_ptr, cyentryFuncAddr_ptr) if err != cyruntime.cudaSuccess: return (_dict_cudaError_t[err], None) return (_dict_cudaError_t[err], kernelPtr) @cython.embedsignature(True) def make_cudaPitchedPtr(d, size_t p, size_t xsz, size_t ysz): """ Returns a :py:obj:`~.cudaPitchedPtr` based on input parameters. Returns a :py:obj:`~.cudaPitchedPtr` based on the specified input parameters `d`, `p`, `xsz`, and `ysz`. Parameters ---------- d : Any Pointer to allocated memory p : size_t Pitch of allocated memory in bytes xsz : size_t Logical width of allocation in elements ysz : size_t Logical height of allocation in elements Returns ------- cudaError_t.cudaSuccess cudaError_t.cudaSuccess :py:obj:`~.cudaPitchedPtr` :py:obj:`~.cudaPitchedPtr` specified by `d`, `p`, `xsz`, and `ysz` See Also -------- make_cudaExtent, make_cudaPos """ cyd = _HelperInputVoidPtr(d) cdef void* cyd_ptr = cyd.cptr with nogil: err = cyruntime.make_cudaPitchedPtr(cyd_ptr, p, xsz, ysz) cdef cudaPitchedPtr wrapper = cudaPitchedPtr() wrapper._pvt_ptr[0] = err return wrapper @cython.embedsignature(True) def make_cudaPos(size_t x, size_t y, size_t z): """ Returns a :py:obj:`~.cudaPos` based on input parameters. Returns a :py:obj:`~.cudaPos` based on the specified input parameters `x`, `y`, and `z`. Parameters ---------- x : size_t X position y : size_t Y position z : size_t Z position Returns ------- cudaError_t.cudaSuccess cudaError_t.cudaSuccess :py:obj:`~.cudaPos` :py:obj:`~.cudaPos` specified by `x`, `y`, and `z` See Also -------- make_cudaExtent, make_cudaPitchedPtr """ with nogil: err = cyruntime.make_cudaPos(x, y, z) cdef cudaPos wrapper = cudaPos() wrapper._pvt_ptr[0] = err return wrapper @cython.embedsignature(True) def make_cudaExtent(size_t w, size_t h, size_t d): """ Returns a :py:obj:`~.cudaExtent` based on input parameters. Returns a :py:obj:`~.cudaExtent` based on the specified input parameters `w`, `h`, and `d`. Parameters ---------- w : size_t Width in elements when referring to array memory, in bytes when referring to linear memory h : size_t Height in elements d : size_t Depth in elements Returns ------- cudaError_t.cudaSuccess cudaError_t.cudaSuccess :py:obj:`~.cudaExtent` :py:obj:`~.cudaExtent` specified by `w`, `h`, and `d` See Also -------- make_cudaPitchedPtr, make_cudaPos """ with nogil: err = cyruntime.make_cudaExtent(w, h, d) cdef cudaExtent wrapper = cudaExtent() wrapper._pvt_ptr[0] = err return wrapper @cython.embedsignature(True) def cudaGraphicsEGLRegisterImage(image, unsigned int flags): """ Registers an EGL image. Registers the EGLImageKHR specified by `image` for access by CUDA. A handle to the registered object is returned as `pCudaResource`. Additional Mapping/Unmapping is not required for the registered resource and :py:obj:`~.cudaGraphicsResourceGetMappedEglFrame` can be directly called on the `pCudaResource`. The application will be responsible for synchronizing access to shared objects. The application must ensure that any pending operation which access the objects have completed before passing control to CUDA. This may be accomplished by issuing and waiting for glFinish command on all GLcontexts (for OpenGL and likewise for other APIs). The application will be also responsible for ensuring that any pending operation on the registered CUDA resource has completed prior to executing subsequent commands in other APIs accesing the same memory objects. This can be accomplished by calling cuCtxSynchronize or cuEventSynchronize (preferably). The surface's intended usage is specified using `flags`, as follows: - :py:obj:`~.cudaGraphicsRegisterFlagsNone`: Specifies no hints about how this resource will be used. It is therefore assumed that this resource will be read from and written to by CUDA. This is the default value. - :py:obj:`~.cudaGraphicsRegisterFlagsReadOnly`: Specifies that CUDA will not write to this resource. - :py:obj:`~.cudaGraphicsRegisterFlagsWriteDiscard`: Specifies that CUDA will not read from this resource and will write over the entire contents of the resource, so none of the data previously stored in the resource will be preserved. The EGLImageKHR is an object which can be used to create EGLImage target resource. It is defined as a void pointer. typedef void* EGLImageKHR Parameters ---------- image : :py:obj:`~.EGLImageKHR` An EGLImageKHR image which can be used to create target resource. flags : unsigned int Map flags Returns ------- cudaError_t :py:obj:`~.cudaSuccess`, :py:obj:`~.cudaErrorInvalidResourceHandle`, :py:obj:`~.cudaErrorInvalidValue`, :py:obj:`~.cudaErrorUnknown` pCudaResource : :py:obj:`~.cudaGraphicsResource` Pointer to the returned object handle See Also -------- :py:obj:`~.cudaGraphicsUnregisterResource`, :py:obj:`~.cudaGraphicsResourceGetMappedEglFrame`, :py:obj:`~.cuGraphicsEGLRegisterImage` """ cdef cyruntime.EGLImageKHR cyimage if image is None: pimage = 0 elif isinstance(image, (EGLImageKHR,)): pimage = int(image) else: pimage = int(EGLImageKHR(image)) cyimage = pimage cdef cudaGraphicsResource_t pCudaResource = cudaGraphicsResource_t() with nogil: err = cyruntime.cudaGraphicsEGLRegisterImage(pCudaResource._pvt_ptr, cyimage, flags) if err != cyruntime.cudaSuccess: return (_dict_cudaError_t[err], None) return (_dict_cudaError_t[err], pCudaResource) @cython.embedsignature(True) def cudaEGLStreamConsumerConnect(eglStream): """ Connect CUDA to EGLStream as a consumer. Connect CUDA as a consumer to EGLStreamKHR specified by `eglStream`. The EGLStreamKHR is an EGL object that transfers a sequence of image frames from one API to another. Parameters ---------- eglStream : :py:obj:`~.EGLStreamKHR` EGLStreamKHR handle Returns ------- cudaError_t :py:obj:`~.cudaSuccess`, :py:obj:`~.cudaErrorInvalidValue`, :py:obj:`~.cudaErrorUnknown` conn : :py:obj:`~.cudaEglStreamConnection` Pointer to the returned connection handle See Also -------- :py:obj:`~.cudaEGLStreamConsumerDisconnect`, :py:obj:`~.cudaEGLStreamConsumerAcquireFrame`, :py:obj:`~.cudaEGLStreamConsumerReleaseFrame`, :py:obj:`~.cuEGLStreamConsumerConnect` """ cdef cyruntime.EGLStreamKHR cyeglStream if eglStream is None: peglStream = 0 elif isinstance(eglStream, (EGLStreamKHR,)): peglStream = int(eglStream) else: peglStream = int(EGLStreamKHR(eglStream)) cyeglStream = peglStream cdef cudaEglStreamConnection conn = cudaEglStreamConnection() with nogil: err = cyruntime.cudaEGLStreamConsumerConnect(conn._pvt_ptr, cyeglStream) if err != cyruntime.cudaSuccess: return (_dict_cudaError_t[err], None) return (_dict_cudaError_t[err], conn) @cython.embedsignature(True) def cudaEGLStreamConsumerConnectWithFlags(eglStream, unsigned int flags): """ Connect CUDA to EGLStream as a consumer with given flags. Connect CUDA as a consumer to EGLStreamKHR specified by `stream` with specified `flags` defined by :py:obj:`~.cudaEglResourceLocationFlags`. The flags specify whether the consumer wants to access frames from system memory or video memory. Default is :py:obj:`~.cudaEglResourceLocationVidmem`. Parameters ---------- eglStream : :py:obj:`~.EGLStreamKHR` EGLStreamKHR handle flags : unsigned int Flags denote intended location - system or video. Returns ------- cudaError_t :py:obj:`~.cudaSuccess`, :py:obj:`~.cudaErrorInvalidValue`, :py:obj:`~.cudaErrorUnknown` conn : :py:obj:`~.cudaEglStreamConnection` Pointer to the returned connection handle See Also -------- :py:obj:`~.cudaEGLStreamConsumerDisconnect`, :py:obj:`~.cudaEGLStreamConsumerAcquireFrame`, :py:obj:`~.cudaEGLStreamConsumerReleaseFrame`, :py:obj:`~.cuEGLStreamConsumerConnectWithFlags` """ cdef cyruntime.EGLStreamKHR cyeglStream if eglStream is None: peglStream = 0 elif isinstance(eglStream, (EGLStreamKHR,)): peglStream = int(eglStream) else: peglStream = int(EGLStreamKHR(eglStream)) cyeglStream = peglStream cdef cudaEglStreamConnection conn = cudaEglStreamConnection() with nogil: err = cyruntime.cudaEGLStreamConsumerConnectWithFlags(conn._pvt_ptr, cyeglStream, flags) if err != cyruntime.cudaSuccess: return (_dict_cudaError_t[err], None) return (_dict_cudaError_t[err], conn) @cython.embedsignature(True) def cudaEGLStreamConsumerDisconnect(conn): """ Disconnect CUDA as a consumer to EGLStream . Disconnect CUDA as a consumer to EGLStreamKHR. Parameters ---------- conn : :py:obj:`~.cudaEglStreamConnection` Conection to disconnect. Returns ------- cudaError_t :py:obj:`~.cudaSuccess`, :py:obj:`~.cudaErrorInvalidValue`, :py:obj:`~.cudaErrorUnknown` See Also -------- :py:obj:`~.cudaEGLStreamConsumerConnect`, :py:obj:`~.cudaEGLStreamConsumerAcquireFrame`, :py:obj:`~.cudaEGLStreamConsumerReleaseFrame`, :py:obj:`~.cuEGLStreamConsumerDisconnect` """ cdef cyruntime.cudaEglStreamConnection *cyconn if conn is None: cyconn = NULL elif isinstance(conn, (cudaEglStreamConnection,driver.CUeglStreamConnection)): pconn = conn.getPtr() cyconn = pconn elif isinstance(conn, (int)): cyconn = conn else: raise TypeError("Argument 'conn' is not instance of type (expected , found " + str(type(conn))) with nogil: err = cyruntime.cudaEGLStreamConsumerDisconnect(cyconn) return (_dict_cudaError_t[err],) @cython.embedsignature(True) def cudaEGLStreamConsumerAcquireFrame(conn, pCudaResource, pStream, unsigned int timeout): """ Acquire an image frame from the EGLStream with CUDA as a consumer. Acquire an image frame from EGLStreamKHR. :py:obj:`~.cudaGraphicsResourceGetMappedEglFrame` can be called on `pCudaResource` to get :py:obj:`~.cudaEglFrame`. Parameters ---------- conn : :py:obj:`~.cudaEglStreamConnection` Connection on which to acquire pCudaResource : :py:obj:`~.cudaGraphicsResource_t` CUDA resource on which the EGLStream frame will be mapped for use. pStream : :py:obj:`~.cudaStream_t` CUDA stream for synchronization and any data migrations implied by :py:obj:`~.cudaEglResourceLocationFlags`. timeout : unsigned int Desired timeout in usec. Returns ------- cudaError_t :py:obj:`~.cudaSuccess`, :py:obj:`~.cudaErrorInvalidValue`, :py:obj:`~.cudaErrorUnknown`, :py:obj:`~.cudaErrorLaunchTimeout` See Also -------- :py:obj:`~.cudaEGLStreamConsumerConnect`, :py:obj:`~.cudaEGLStreamConsumerDisconnect`, :py:obj:`~.cudaEGLStreamConsumerReleaseFrame`, :py:obj:`~.cuEGLStreamConsumerAcquireFrame` """ cdef cyruntime.cudaStream_t *cypStream if pStream is None: cypStream = NULL elif isinstance(pStream, (cudaStream_t,driver.CUstream)): ppStream = pStream.getPtr() cypStream = ppStream elif isinstance(pStream, (int)): cypStream = pStream else: raise TypeError("Argument 'pStream' is not instance of type (expected , found " + str(type(pStream))) cdef cyruntime.cudaGraphicsResource_t *cypCudaResource if pCudaResource is None: cypCudaResource = NULL elif isinstance(pCudaResource, (cudaGraphicsResource_t,)): ppCudaResource = pCudaResource.getPtr() cypCudaResource = ppCudaResource elif isinstance(pCudaResource, (int)): cypCudaResource = pCudaResource else: raise TypeError("Argument 'pCudaResource' is not instance of type (expected , found " + str(type(pCudaResource))) cdef cyruntime.cudaEglStreamConnection *cyconn if conn is None: cyconn = NULL elif isinstance(conn, (cudaEglStreamConnection,driver.CUeglStreamConnection)): pconn = conn.getPtr() cyconn = pconn elif isinstance(conn, (int)): cyconn = conn else: raise TypeError("Argument 'conn' is not instance of type (expected , found " + str(type(conn))) with nogil: err = cyruntime.cudaEGLStreamConsumerAcquireFrame(cyconn, cypCudaResource, cypStream, timeout) return (_dict_cudaError_t[err],) @cython.embedsignature(True) def cudaEGLStreamConsumerReleaseFrame(conn, pCudaResource, pStream): """ Releases the last frame acquired from the EGLStream. Release the acquired image frame specified by `pCudaResource` to EGLStreamKHR. Parameters ---------- conn : :py:obj:`~.cudaEglStreamConnection` Connection on which to release pCudaResource : :py:obj:`~.cudaGraphicsResource_t` CUDA resource whose corresponding frame is to be released pStream : :py:obj:`~.cudaStream_t` CUDA stream on which release will be done. Returns ------- cudaError_t :py:obj:`~.cudaSuccess`, :py:obj:`~.cudaErrorInvalidValue`, :py:obj:`~.cudaErrorUnknown` See Also -------- :py:obj:`~.cudaEGLStreamConsumerConnect`, :py:obj:`~.cudaEGLStreamConsumerDisconnect`, :py:obj:`~.cudaEGLStreamConsumerAcquireFrame`, :py:obj:`~.cuEGLStreamConsumerReleaseFrame` """ cdef cyruntime.cudaStream_t *cypStream if pStream is None: cypStream = NULL elif isinstance(pStream, (cudaStream_t,driver.CUstream)): ppStream = pStream.getPtr() cypStream = ppStream elif isinstance(pStream, (int)): cypStream = pStream else: raise TypeError("Argument 'pStream' is not instance of type (expected , found " + str(type(pStream))) cdef cyruntime.cudaGraphicsResource_t cypCudaResource if pCudaResource is None: ppCudaResource = 0 elif isinstance(pCudaResource, (cudaGraphicsResource_t,)): ppCudaResource = int(pCudaResource) else: ppCudaResource = int(cudaGraphicsResource_t(pCudaResource)) cypCudaResource = ppCudaResource cdef cyruntime.cudaEglStreamConnection *cyconn if conn is None: cyconn = NULL elif isinstance(conn, (cudaEglStreamConnection,driver.CUeglStreamConnection)): pconn = conn.getPtr() cyconn = pconn elif isinstance(conn, (int)): cyconn = conn else: raise TypeError("Argument 'conn' is not instance of type (expected , found " + str(type(conn))) with nogil: err = cyruntime.cudaEGLStreamConsumerReleaseFrame(cyconn, cypCudaResource, cypStream) return (_dict_cudaError_t[err],) @cython.embedsignature(True) def cudaEGLStreamProducerConnect(eglStream, width, height): """ Connect CUDA to EGLStream as a producer. Connect CUDA as a producer to EGLStreamKHR specified by `stream`. The EGLStreamKHR is an EGL object that transfers a sequence of image frames from one API to another. Parameters ---------- eglStream : :py:obj:`~.EGLStreamKHR` EGLStreamKHR handle width : :py:obj:`~.EGLint` width of the image to be submitted to the stream height : :py:obj:`~.EGLint` height of the image to be submitted to the stream Returns ------- cudaError_t :py:obj:`~.cudaSuccess`, :py:obj:`~.cudaErrorInvalidValue`, :py:obj:`~.cudaErrorUnknown` conn : :py:obj:`~.cudaEglStreamConnection` Pointer to the returned connection handle See Also -------- :py:obj:`~.cudaEGLStreamProducerDisconnect`, :py:obj:`~.cudaEGLStreamProducerPresentFrame`, :py:obj:`~.cudaEGLStreamProducerReturnFrame`, :py:obj:`~.cuEGLStreamProducerConnect` """ cdef cyruntime.EGLint cyheight if height is None: pheight = 0 elif isinstance(height, (EGLint,)): pheight = int(height) else: pheight = int(EGLint(height)) cyheight = pheight cdef cyruntime.EGLint cywidth if width is None: pwidth = 0 elif isinstance(width, (EGLint,)): pwidth = int(width) else: pwidth = int(EGLint(width)) cywidth = pwidth cdef cyruntime.EGLStreamKHR cyeglStream if eglStream is None: peglStream = 0 elif isinstance(eglStream, (EGLStreamKHR,)): peglStream = int(eglStream) else: peglStream = int(EGLStreamKHR(eglStream)) cyeglStream = peglStream cdef cudaEglStreamConnection conn = cudaEglStreamConnection() with nogil: err = cyruntime.cudaEGLStreamProducerConnect(conn._pvt_ptr, cyeglStream, cywidth, cyheight) if err != cyruntime.cudaSuccess: return (_dict_cudaError_t[err], None) return (_dict_cudaError_t[err], conn) @cython.embedsignature(True) def cudaEGLStreamProducerDisconnect(conn): """ Disconnect CUDA as a producer to EGLStream . Disconnect CUDA as a producer to EGLStreamKHR. Parameters ---------- conn : :py:obj:`~.cudaEglStreamConnection` Conection to disconnect. Returns ------- cudaError_t :py:obj:`~.cudaSuccess`, :py:obj:`~.cudaErrorInvalidValue`, :py:obj:`~.cudaErrorUnknown` See Also -------- :py:obj:`~.cudaEGLStreamProducerConnect`, :py:obj:`~.cudaEGLStreamProducerPresentFrame`, :py:obj:`~.cudaEGLStreamProducerReturnFrame`, :py:obj:`~.cuEGLStreamProducerDisconnect` """ cdef cyruntime.cudaEglStreamConnection *cyconn if conn is None: cyconn = NULL elif isinstance(conn, (cudaEglStreamConnection,driver.CUeglStreamConnection)): pconn = conn.getPtr() cyconn = pconn elif isinstance(conn, (int)): cyconn = conn else: raise TypeError("Argument 'conn' is not instance of type (expected , found " + str(type(conn))) with nogil: err = cyruntime.cudaEGLStreamProducerDisconnect(cyconn) return (_dict_cudaError_t[err],) @cython.embedsignature(True) def cudaEGLStreamProducerPresentFrame(conn, eglframe not None : cudaEglFrame, pStream): """ Present a CUDA eglFrame to the EGLStream with CUDA as a producer. The :py:obj:`~.cudaEglFrame` is defined as: **View CUDA Toolkit Documentation for a C++ code example** For :py:obj:`~.cudaEglFrame` of type :py:obj:`~.cudaEglFrameTypePitch`, the application may present sub-region of a memory allocation. In that case, :py:obj:`~.cudaPitchedPtr.ptr` will specify the start address of the sub-region in the allocation and :py:obj:`~.cudaEglPlaneDesc` will specify the dimensions of the sub-region. Parameters ---------- conn : :py:obj:`~.cudaEglStreamConnection` Connection on which to present the CUDA array eglframe : :py:obj:`~.cudaEglFrame` CUDA Eglstream Proucer Frame handle to be sent to the consumer over EglStream. pStream : :py:obj:`~.cudaStream_t` CUDA stream on which to present the frame. Returns ------- cudaError_t :py:obj:`~.cudaSuccess`, :py:obj:`~.cudaErrorInvalidValue`, :py:obj:`~.cudaErrorUnknown` See Also -------- :py:obj:`~.cudaEGLStreamProducerConnect`, :py:obj:`~.cudaEGLStreamProducerDisconnect`, :py:obj:`~.cudaEGLStreamProducerReturnFrame`, :py:obj:`~.cuEGLStreamProducerPresentFrame` """ cdef cyruntime.cudaStream_t *cypStream if pStream is None: cypStream = NULL elif isinstance(pStream, (cudaStream_t,driver.CUstream)): ppStream = pStream.getPtr() cypStream = ppStream elif isinstance(pStream, (int)): cypStream = pStream else: raise TypeError("Argument 'pStream' is not instance of type (expected , found " + str(type(pStream))) cdef cyruntime.cudaEglStreamConnection *cyconn if conn is None: cyconn = NULL elif isinstance(conn, (cudaEglStreamConnection,driver.CUeglStreamConnection)): pconn = conn.getPtr() cyconn = pconn elif isinstance(conn, (int)): cyconn = conn else: raise TypeError("Argument 'conn' is not instance of type (expected , found " + str(type(conn))) with nogil: err = cyruntime.cudaEGLStreamProducerPresentFrame(cyconn, eglframe._pvt_ptr[0], cypStream) return (_dict_cudaError_t[err],) @cython.embedsignature(True) def cudaEGLStreamProducerReturnFrame(conn, eglframe : Optional[cudaEglFrame], pStream): """ Return the CUDA eglFrame to the EGLStream last released by the consumer. This API can potentially return cudaErrorLaunchTimeout if the consumer has not returned a frame to EGL stream. If timeout is returned the application can retry. Parameters ---------- conn : :py:obj:`~.cudaEglStreamConnection` Connection on which to present the CUDA array eglframe : :py:obj:`~.cudaEglFrame` CUDA Eglstream Proucer Frame handle returned from the consumer over EglStream. pStream : :py:obj:`~.cudaStream_t` CUDA stream on which to return the frame. Returns ------- cudaError_t :py:obj:`~.cudaSuccess`, :py:obj:`~.cudaErrorLaunchTimeout`, :py:obj:`~.cudaErrorInvalidValue`, :py:obj:`~.cudaErrorUnknown` See Also -------- :py:obj:`~.cudaEGLStreamProducerConnect`, :py:obj:`~.cudaEGLStreamProducerDisconnect`, :py:obj:`~.cudaEGLStreamProducerPresentFrame`, :py:obj:`~.cuEGLStreamProducerReturnFrame` """ cdef cyruntime.cudaStream_t *cypStream if pStream is None: cypStream = NULL elif isinstance(pStream, (cudaStream_t,driver.CUstream)): ppStream = pStream.getPtr() cypStream = ppStream elif isinstance(pStream, (int)): cypStream = pStream else: raise TypeError("Argument 'pStream' is not instance of type (expected , found " + str(type(pStream))) cdef cyruntime.cudaEglStreamConnection *cyconn if conn is None: cyconn = NULL elif isinstance(conn, (cudaEglStreamConnection,driver.CUeglStreamConnection)): pconn = conn.getPtr() cyconn = pconn elif isinstance(conn, (int)): cyconn = conn else: raise TypeError("Argument 'conn' is not instance of type (expected , found " + str(type(conn))) cdef cyruntime.cudaEglFrame* cyeglframe_ptr = eglframe._pvt_ptr if eglframe is not None else NULL with nogil: err = cyruntime.cudaEGLStreamProducerReturnFrame(cyconn, cyeglframe_ptr, cypStream) return (_dict_cudaError_t[err],) @cython.embedsignature(True) def cudaGraphicsResourceGetMappedEglFrame(resource, unsigned int index, unsigned int mipLevel): """ Get an eglFrame through which to access a registered EGL graphics resource. Returns in `*eglFrame` an eglFrame pointer through which the registered graphics resource `resource` may be accessed. This API can only be called for EGL graphics resources. The :py:obj:`~.cudaEglFrame` is defined as **View CUDA Toolkit Documentation for a C++ code example** Parameters ---------- resource : :py:obj:`~.cudaGraphicsResource_t` Registered resource to access. index : unsigned int Index for cubemap surfaces. mipLevel : unsigned int Mipmap level for the subresource to access. Returns ------- cudaError_t :py:obj:`~.cudaSuccess`, :py:obj:`~.cudaErrorInvalidValue`, :py:obj:`~.cudaErrorUnknown` eglFrame : :py:obj:`~.cudaEglFrame` Returned eglFrame. See Also -------- :py:obj:`~.cudaGraphicsSubResourceGetMappedArray`, :py:obj:`~.cudaGraphicsResourceGetMappedPointer`, :py:obj:`~.cuGraphicsResourceGetMappedEglFrame` Notes ----- Note that in case of multiplanar `*eglFrame`, pitch of only first plane (unsigned int :py:obj:`~.cudaEglPlaneDesc.pitch`) is to be considered by the application. """ cdef cyruntime.cudaGraphicsResource_t cyresource if resource is None: presource = 0 elif isinstance(resource, (cudaGraphicsResource_t,)): presource = int(resource) else: presource = int(cudaGraphicsResource_t(resource)) cyresource = presource cdef cudaEglFrame eglFrame = cudaEglFrame() with nogil: err = cyruntime.cudaGraphicsResourceGetMappedEglFrame(eglFrame._pvt_ptr, cyresource, index, mipLevel) if err != cyruntime.cudaSuccess: return (_dict_cudaError_t[err], None) return (_dict_cudaError_t[err], eglFrame) @cython.embedsignature(True) def cudaEventCreateFromEGLSync(eglSync, unsigned int flags): """ Creates an event from EGLSync object. Creates an event *phEvent from an EGLSyncKHR eglSync with the flages specified via `flags`. Valid flags include: - :py:obj:`~.cudaEventDefault`: Default event creation flag. - :py:obj:`~.cudaEventBlockingSync`: Specifies that the created event should use blocking synchronization. A CPU thread that uses :py:obj:`~.cudaEventSynchronize()` to wait on an event created with this flag will block until the event has actually been completed. :py:obj:`~.cudaEventRecord` and TimingData are not supported for events created from EGLSync. The EGLSyncKHR is an opaque handle to an EGL sync object. typedef void* EGLSyncKHR Parameters ---------- eglSync : :py:obj:`~.EGLSyncKHR` Opaque handle to EGLSync object flags : unsigned int Event creation flags Returns ------- cudaError_t :py:obj:`~.cudaSuccess`, :py:obj:`~.cudaErrorInitializationError`, :py:obj:`~.cudaErrorInvalidValue`, :py:obj:`~.cudaErrorLaunchFailure`, :py:obj:`~.cudaErrorMemoryAllocation` phEvent : :py:obj:`~.cudaEvent_t` Returns newly created event See Also -------- :py:obj:`~.cudaEventQuery`, :py:obj:`~.cudaEventSynchronize`, :py:obj:`~.cudaEventDestroy` """ cdef cyruntime.EGLSyncKHR cyeglSync if eglSync is None: peglSync = 0 elif isinstance(eglSync, (EGLSyncKHR,)): peglSync = int(eglSync) else: peglSync = int(EGLSyncKHR(eglSync)) cyeglSync = peglSync cdef cudaEvent_t phEvent = cudaEvent_t() with nogil: err = cyruntime.cudaEventCreateFromEGLSync(phEvent._pvt_ptr, cyeglSync, flags) if err != cyruntime.cudaSuccess: return (_dict_cudaError_t[err], None) return (_dict_cudaError_t[err], phEvent) @cython.embedsignature(True) def cudaProfilerStart(): """ Enable profiling. Enables profile collection by the active profiling tool for the current context. If profiling is already enabled, then :py:obj:`~.cudaProfilerStart()` has no effect. cudaProfilerStart and cudaProfilerStop APIs are used to programmatically control the profiling granularity by allowing profiling to be done only on selective pieces of code. Returns ------- cudaError_t :py:obj:`~.cudaSuccess` See Also -------- :py:obj:`~.cudaProfilerStop`, :py:obj:`~.cuProfilerStart` """ with nogil: err = cyruntime.cudaProfilerStart() return (_dict_cudaError_t[err],) @cython.embedsignature(True) def cudaProfilerStop(): """ Disable profiling. Disables profile collection by the active profiling tool for the current context. If profiling is already disabled, then :py:obj:`~.cudaProfilerStop()` has no effect. cudaProfilerStart and cudaProfilerStop APIs are used to programmatically control the profiling granularity by allowing profiling to be done only on selective pieces of code. Returns ------- cudaError_t :py:obj:`~.cudaSuccess` See Also -------- :py:obj:`~.cudaProfilerStart`, :py:obj:`~.cuProfilerStop` """ with nogil: err = cyruntime.cudaProfilerStop() return (_dict_cudaError_t[err],) @cython.embedsignature(True) def cudaGLGetDevices(unsigned int cudaDeviceCount, deviceList not None : cudaGLDeviceList): """ Gets the CUDA devices associated with the current OpenGL context. Returns in `*pCudaDeviceCount` the number of CUDA-compatible devices corresponding to the current OpenGL context. Also returns in `*pCudaDevices` at most `cudaDeviceCount` of the CUDA-compatible devices corresponding to the current OpenGL context. If any of the GPUs being used by the current OpenGL context are not CUDA capable then the call will return cudaErrorNoDevice. Parameters ---------- cudaDeviceCount : unsigned int The size of the output device array `pCudaDevices` deviceList : cudaGLDeviceList The set of devices to return. This set may be cudaGLDeviceListAll for all devices, cudaGLDeviceListCurrentFrame for the devices used to render the current frame (in SLI), or cudaGLDeviceListNextFrame for the devices used to render the next frame (in SLI). Returns ------- cudaError_t cudaSuccess cudaErrorNoDevice cudaErrorInvalidGraphicsContext cudaErrorUnknown pCudaDeviceCount : unsigned int Returned number of CUDA devices corresponding to the current OpenGL context pCudaDevices : list[int] Returned CUDA devices corresponding to the current OpenGL context See Also -------- ~.cudaGraphicsUnregisterResource ~.cudaGraphicsMapResources ~.cudaGraphicsSubResourceGetMappedArray ~.cudaGraphicsResourceGetMappedPointer ~.cuGLGetDevices Notes ----- This function is not supported on Mac OS X. """ cdef unsigned int pCudaDeviceCount = 0 cdef int* cypCudaDevices = NULL pypCudaDevices = [] if cudaDeviceCount != 0: cypCudaDevices = calloc(cudaDeviceCount, sizeof(int)) if cypCudaDevices is NULL: raise MemoryError('Failed to allocate length x size memory: ' + str(cudaDeviceCount) + 'x' + str(sizeof(int))) cdef cyruntime.cudaGLDeviceList cydeviceList = deviceList.value with nogil: err = cyruntime.cudaGLGetDevices(&pCudaDeviceCount, cypCudaDevices, cudaDeviceCount, cydeviceList) if cudaError_t(err) == cudaError_t(0): pypCudaDevices = [cypCudaDevices[idx] for idx in range(cudaDeviceCount)] if cypCudaDevices is not NULL: free(cypCudaDevices) if err != cyruntime.cudaSuccess: return (_dict_cudaError_t[err], None, None) return (_dict_cudaError_t[err], pCudaDeviceCount, pypCudaDevices) @cython.embedsignature(True) def cudaGraphicsGLRegisterImage(image, target, unsigned int flags): """ Register an OpenGL texture or renderbuffer object. Registers the texture or renderbuffer object specified by `image` for access by CUDA. A handle to the registered object is returned as `resource`. `target` must match the type of the object, and must be one of :py:obj:`~.GL_TEXTURE_2D`, :py:obj:`~.GL_TEXTURE_RECTANGLE`, :py:obj:`~.GL_TEXTURE_CUBE_MAP`, :py:obj:`~.GL_TEXTURE_3D`, :py:obj:`~.GL_TEXTURE_2D_ARRAY`, or :py:obj:`~.GL_RENDERBUFFER`. The register flags `flags` specify the intended usage, as follows: - :py:obj:`~.cudaGraphicsRegisterFlagsNone`: Specifies no hints about how this resource will be used. It is therefore assumed that this resource will be read from and written to by CUDA. This is the default value. - :py:obj:`~.cudaGraphicsRegisterFlagsReadOnly`: Specifies that CUDA will not write to this resource. - :py:obj:`~.cudaGraphicsRegisterFlagsWriteDiscard`: Specifies that CUDA will not read from this resource and will write over the entire contents of the resource, so none of the data previously stored in the resource will be preserved. - :py:obj:`~.cudaGraphicsRegisterFlagsSurfaceLoadStore`: Specifies that CUDA will bind this resource to a surface reference. - :py:obj:`~.cudaGraphicsRegisterFlagsTextureGather`: Specifies that CUDA will perform texture gather operations on this resource. The following image formats are supported. For brevity's sake, the list is abbreviated. For ex., {GL_R, GL_RG} X {8, 16} would expand to the following 4 formats {GL_R8, GL_R16, GL_RG8, GL_RG16} : - GL_RED, GL_RG, GL_RGBA, GL_LUMINANCE, GL_ALPHA, GL_LUMINANCE_ALPHA, GL_INTENSITY - {GL_R, GL_RG, GL_RGBA} X {8, 16, 16F, 32F, 8UI, 16UI, 32UI, 8I, 16I, 32I} - {GL_LUMINANCE, GL_ALPHA, GL_LUMINANCE_ALPHA, GL_INTENSITY} X {8, 16, 16F_ARB, 32F_ARB, 8UI_EXT, 16UI_EXT, 32UI_EXT, 8I_EXT, 16I_EXT, 32I_EXT} The following image classes are currently disallowed: - Textures with borders - Multisampled renderbuffers Parameters ---------- image : :py:obj:`~.GLuint` name of texture or renderbuffer object to be registered target : :py:obj:`~.GLenum` Identifies the type of object specified by `image` flags : unsigned int Register flags Returns ------- cudaError_t :py:obj:`~.cudaSuccess`, :py:obj:`~.cudaErrorInvalidDevice`, :py:obj:`~.cudaErrorInvalidValue`, :py:obj:`~.cudaErrorInvalidResourceHandle`, :py:obj:`~.cudaErrorOperatingSystem`, :py:obj:`~.cudaErrorUnknown` resource : :py:obj:`~.cudaGraphicsResource` Pointer to the returned object handle See Also -------- :py:obj:`~.cudaGraphicsUnregisterResource`, :py:obj:`~.cudaGraphicsMapResources`, :py:obj:`~.cudaGraphicsSubResourceGetMappedArray`, :py:obj:`~.cuGraphicsGLRegisterImage` """ cdef cyruntime.GLenum cytarget if target is None: ptarget = 0 elif isinstance(target, (GLenum,)): ptarget = int(target) else: ptarget = int(GLenum(target)) cytarget = ptarget cdef cyruntime.GLuint cyimage if image is None: pimage = 0 elif isinstance(image, (GLuint,)): pimage = int(image) else: pimage = int(GLuint(image)) cyimage = pimage cdef cudaGraphicsResource_t resource = cudaGraphicsResource_t() with nogil: err = cyruntime.cudaGraphicsGLRegisterImage(resource._pvt_ptr, cyimage, cytarget, flags) if err != cyruntime.cudaSuccess: return (_dict_cudaError_t[err], None) return (_dict_cudaError_t[err], resource) @cython.embedsignature(True) def cudaGraphicsGLRegisterBuffer(buffer, unsigned int flags): """ Registers an OpenGL buffer object. Registers the buffer object specified by `buffer` for access by CUDA. A handle to the registered object is returned as `resource`. The register flags `flags` specify the intended usage, as follows: - :py:obj:`~.cudaGraphicsRegisterFlagsNone`: Specifies no hints about how this resource will be used. It is therefore assumed that this resource will be read from and written to by CUDA. This is the default value. - :py:obj:`~.cudaGraphicsRegisterFlagsReadOnly`: Specifies that CUDA will not write to this resource. - :py:obj:`~.cudaGraphicsRegisterFlagsWriteDiscard`: Specifies that CUDA will not read from this resource and will write over the entire contents of the resource, so none of the data previously stored in the resource will be preserved. Parameters ---------- buffer : :py:obj:`~.GLuint` name of buffer object to be registered flags : unsigned int Register flags Returns ------- cudaError_t :py:obj:`~.cudaSuccess`, :py:obj:`~.cudaErrorInvalidDevice`, :py:obj:`~.cudaErrorInvalidValue`, :py:obj:`~.cudaErrorInvalidResourceHandle`, :py:obj:`~.cudaErrorOperatingSystem`, :py:obj:`~.cudaErrorUnknown` resource : :py:obj:`~.cudaGraphicsResource` Pointer to the returned object handle See Also -------- :py:obj:`~.cudaGraphicsUnregisterResource`, :py:obj:`~.cudaGraphicsMapResources`, :py:obj:`~.cudaGraphicsResourceGetMappedPointer`, :py:obj:`~.cuGraphicsGLRegisterBuffer` """ cdef cyruntime.GLuint cybuffer if buffer is None: pbuffer = 0 elif isinstance(buffer, (GLuint,)): pbuffer = int(buffer) else: pbuffer = int(GLuint(buffer)) cybuffer = pbuffer cdef cudaGraphicsResource_t resource = cudaGraphicsResource_t() with nogil: err = cyruntime.cudaGraphicsGLRegisterBuffer(resource._pvt_ptr, cybuffer, flags) if err != cyruntime.cudaSuccess: return (_dict_cudaError_t[err], None) return (_dict_cudaError_t[err], resource) @cython.embedsignature(True) def cudaVDPAUGetDevice(vdpDevice, vdpGetProcAddress): """ Gets the CUDA device associated with a VdpDevice. Returns the CUDA device associated with a VdpDevice, if applicable. Parameters ---------- vdpDevice : :py:obj:`~.VdpDevice` A VdpDevice handle vdpGetProcAddress : :py:obj:`~.VdpGetProcAddress` VDPAU's VdpGetProcAddress function pointer Returns ------- cudaError_t :py:obj:`~.cudaSuccess` device : int Returns the device associated with vdpDevice, or -1 if the device associated with vdpDevice is not a compute device. See Also -------- :py:obj:`~.cudaVDPAUSetVDPAUDevice`, :py:obj:`~.cuVDPAUGetDevice` """ cdef cyruntime.VdpGetProcAddress *cyvdpGetProcAddress if vdpGetProcAddress is None: cyvdpGetProcAddress = NULL elif isinstance(vdpGetProcAddress, (VdpGetProcAddress,)): pvdpGetProcAddress = vdpGetProcAddress.getPtr() cyvdpGetProcAddress = pvdpGetProcAddress elif isinstance(vdpGetProcAddress, (int)): cyvdpGetProcAddress = vdpGetProcAddress else: raise TypeError("Argument 'vdpGetProcAddress' is not instance of type (expected , found " + str(type(vdpGetProcAddress))) cdef cyruntime.VdpDevice cyvdpDevice if vdpDevice is None: pvdpDevice = 0 elif isinstance(vdpDevice, (VdpDevice,)): pvdpDevice = int(vdpDevice) else: pvdpDevice = int(VdpDevice(vdpDevice)) cyvdpDevice = pvdpDevice cdef int device = 0 with nogil: err = cyruntime.cudaVDPAUGetDevice(&device, cyvdpDevice, cyvdpGetProcAddress) if err != cyruntime.cudaSuccess: return (_dict_cudaError_t[err], None) return (_dict_cudaError_t[err], device) @cython.embedsignature(True) def cudaVDPAUSetVDPAUDevice(int device, vdpDevice, vdpGetProcAddress): """ Sets a CUDA device to use VDPAU interoperability. Records `vdpDevice` as the VdpDevice for VDPAU interoperability with the CUDA device `device` and sets `device` as the current device for the calling host thread. This function will immediately initialize the primary context on `device` if needed. If `device` has already been initialized then this call will fail with the error :py:obj:`~.cudaErrorSetOnActiveProcess`. In this case it is necessary to reset `device` using :py:obj:`~.cudaDeviceReset()` before VDPAU interoperability on `device` may be enabled. Parameters ---------- device : int Device to use for VDPAU interoperability vdpDevice : :py:obj:`~.VdpDevice` The VdpDevice to interoperate with vdpGetProcAddress : :py:obj:`~.VdpGetProcAddress` VDPAU's VdpGetProcAddress function pointer Returns ------- cudaError_t :py:obj:`~.cudaSuccess`, :py:obj:`~.cudaErrorInvalidDevice`, :py:obj:`~.cudaErrorSetOnActiveProcess` See Also -------- :py:obj:`~.cudaGraphicsVDPAURegisterVideoSurface`, :py:obj:`~.cudaGraphicsVDPAURegisterOutputSurface`, :py:obj:`~.cudaDeviceReset` """ cdef cyruntime.VdpGetProcAddress *cyvdpGetProcAddress if vdpGetProcAddress is None: cyvdpGetProcAddress = NULL elif isinstance(vdpGetProcAddress, (VdpGetProcAddress,)): pvdpGetProcAddress = vdpGetProcAddress.getPtr() cyvdpGetProcAddress = pvdpGetProcAddress elif isinstance(vdpGetProcAddress, (int)): cyvdpGetProcAddress = vdpGetProcAddress else: raise TypeError("Argument 'vdpGetProcAddress' is not instance of type (expected , found " + str(type(vdpGetProcAddress))) cdef cyruntime.VdpDevice cyvdpDevice if vdpDevice is None: pvdpDevice = 0 elif isinstance(vdpDevice, (VdpDevice,)): pvdpDevice = int(vdpDevice) else: pvdpDevice = int(VdpDevice(vdpDevice)) cyvdpDevice = pvdpDevice with nogil: err = cyruntime.cudaVDPAUSetVDPAUDevice(device, cyvdpDevice, cyvdpGetProcAddress) return (_dict_cudaError_t[err],) @cython.embedsignature(True) def cudaGraphicsVDPAURegisterVideoSurface(vdpSurface, unsigned int flags): """ Register a VdpVideoSurface object. Registers the VdpVideoSurface specified by `vdpSurface` for access by CUDA. A handle to the registered object is returned as `resource`. The surface's intended usage is specified using `flags`, as follows: - :py:obj:`~.cudaGraphicsMapFlagsNone`: Specifies no hints about how this resource will be used. It is therefore assumed that this resource will be read from and written to by CUDA. This is the default value. - :py:obj:`~.cudaGraphicsMapFlagsReadOnly`: Specifies that CUDA will not write to this resource. - :py:obj:`~.cudaGraphicsMapFlagsWriteDiscard`: Specifies that CUDA will not read from this resource and will write over the entire contents of the resource, so none of the data previously stored in the resource will be preserved. Parameters ---------- vdpSurface : :py:obj:`~.VdpVideoSurface` VDPAU object to be registered flags : unsigned int Map flags Returns ------- cudaError_t :py:obj:`~.cudaSuccess`, :py:obj:`~.cudaErrorInvalidDevice`, :py:obj:`~.cudaErrorInvalidValue`, :py:obj:`~.cudaErrorInvalidResourceHandle`, :py:obj:`~.cudaErrorUnknown` resource : :py:obj:`~.cudaGraphicsResource` Pointer to the returned object handle See Also -------- :py:obj:`~.cudaVDPAUSetVDPAUDevice`, :py:obj:`~.cudaGraphicsUnregisterResource`, :py:obj:`~.cudaGraphicsSubResourceGetMappedArray`, :py:obj:`~.cuGraphicsVDPAURegisterVideoSurface` """ cdef cyruntime.VdpVideoSurface cyvdpSurface if vdpSurface is None: pvdpSurface = 0 elif isinstance(vdpSurface, (VdpVideoSurface,)): pvdpSurface = int(vdpSurface) else: pvdpSurface = int(VdpVideoSurface(vdpSurface)) cyvdpSurface = pvdpSurface cdef cudaGraphicsResource_t resource = cudaGraphicsResource_t() with nogil: err = cyruntime.cudaGraphicsVDPAURegisterVideoSurface(resource._pvt_ptr, cyvdpSurface, flags) if err != cyruntime.cudaSuccess: return (_dict_cudaError_t[err], None) return (_dict_cudaError_t[err], resource) @cython.embedsignature(True) def cudaGraphicsVDPAURegisterOutputSurface(vdpSurface, unsigned int flags): """ Register a VdpOutputSurface object. Registers the VdpOutputSurface specified by `vdpSurface` for access by CUDA. A handle to the registered object is returned as `resource`. The surface's intended usage is specified using `flags`, as follows: - :py:obj:`~.cudaGraphicsMapFlagsNone`: Specifies no hints about how this resource will be used. It is therefore assumed that this resource will be read from and written to by CUDA. This is the default value. - :py:obj:`~.cudaGraphicsMapFlagsReadOnly`: Specifies that CUDA will not write to this resource. - :py:obj:`~.cudaGraphicsMapFlagsWriteDiscard`: Specifies that CUDA will not read from this resource and will write over the entire contents of the resource, so none of the data previously stored in the resource will be preserved. Parameters ---------- vdpSurface : :py:obj:`~.VdpOutputSurface` VDPAU object to be registered flags : unsigned int Map flags Returns ------- cudaError_t :py:obj:`~.cudaSuccess`, :py:obj:`~.cudaErrorInvalidDevice`, :py:obj:`~.cudaErrorInvalidValue`, :py:obj:`~.cudaErrorInvalidResourceHandle`, :py:obj:`~.cudaErrorUnknown` resource : :py:obj:`~.cudaGraphicsResource` Pointer to the returned object handle See Also -------- :py:obj:`~.cudaVDPAUSetVDPAUDevice`, :py:obj:`~.cudaGraphicsUnregisterResource`, :py:obj:`~.cudaGraphicsSubResourceGetMappedArray`, :py:obj:`~.cuGraphicsVDPAURegisterOutputSurface` """ cdef cyruntime.VdpOutputSurface cyvdpSurface if vdpSurface is None: pvdpSurface = 0 elif isinstance(vdpSurface, (VdpOutputSurface,)): pvdpSurface = int(vdpSurface) else: pvdpSurface = int(VdpOutputSurface(vdpSurface)) cyvdpSurface = pvdpSurface cdef cudaGraphicsResource_t resource = cudaGraphicsResource_t() with nogil: err = cyruntime.cudaGraphicsVDPAURegisterOutputSurface(resource._pvt_ptr, cyvdpSurface, flags) if err != cyruntime.cudaSuccess: return (_dict_cudaError_t[err], None) return (_dict_cudaError_t[err], resource) @cython.embedsignature(True) def getLocalRuntimeVersion(): """ Returns the CUDA Runtime version of local shared library. Returns in `*runtimeVersion` the version number of the current CUDA Runtime instance. The version is returned as (1000 * major + 10 * minor). For example, CUDA 9.2 would be represented by 9020. As of CUDA 12.0, this function no longer initializes CUDA. The purpose of this API is solely to return a compile-time constant stating the CUDA Toolkit version in the above format. This function automatically returns :py:obj:`~.cudaErrorInvalidValue` if the `runtimeVersion` argument is NULL. Returns ------- cudaError_t :py:obj:`~.cudaSuccess`, :py:obj:`~.cudaErrorInvalidValue` runtimeVersion : int Returns the CUDA Runtime version. See Also -------- :py:obj:`~.cudaDriverGetVersion`, :py:obj:`~.cuDriverGetVersion` """ cdef int runtimeVersion = 0 err = cyruntime.getLocalRuntimeVersion(&runtimeVersion) return (cudaError_t(err), runtimeVersion) cdef class cudaBindingsRuntimeGlobal: cdef map[void_ptr, void*] _allocated def __dealloc__(self): for item in self._allocated: free(item.second) self._allocated.clear() cdef cudaBindingsRuntimeGlobal m_global = cudaBindingsRuntimeGlobal() @cython.embedsignature(True) def sizeof(objType): """ Returns the size of provided CUDA Python structure in bytes Parameters ---------- objType : Any CUDA Python object Returns ------- lowered_name : int The size of `objType` in bytes """ if objType == dim3: return sizeof(cyruntime.dim3) if objType == cudaChannelFormatDesc: return sizeof(cyruntime.cudaChannelFormatDesc) if objType == cudaArray_t: return sizeof(cyruntime.cudaArray_t) if objType == cudaArray_const_t: return sizeof(cyruntime.cudaArray_const_t) if objType == cudaMipmappedArray_t: return sizeof(cyruntime.cudaMipmappedArray_t) if objType == cudaMipmappedArray_const_t: return sizeof(cyruntime.cudaMipmappedArray_const_t) if objType == cudaArraySparseProperties: return sizeof(cyruntime.cudaArraySparseProperties) if objType == cudaArrayMemoryRequirements: return sizeof(cyruntime.cudaArrayMemoryRequirements) if objType == cudaPitchedPtr: return sizeof(cyruntime.cudaPitchedPtr) if objType == cudaExtent: return sizeof(cyruntime.cudaExtent) if objType == cudaPos: return sizeof(cyruntime.cudaPos) if objType == cudaMemcpy3DParms: return sizeof(cyruntime.cudaMemcpy3DParms) if objType == cudaMemcpyNodeParams: return sizeof(cyruntime.cudaMemcpyNodeParams) if objType == cudaMemcpy3DPeerParms: return sizeof(cyruntime.cudaMemcpy3DPeerParms) if objType == cudaMemsetParams: return sizeof(cyruntime.cudaMemsetParams) if objType == cudaMemsetParamsV2: return sizeof(cyruntime.cudaMemsetParamsV2) if objType == cudaAccessPolicyWindow: return sizeof(cyruntime.cudaAccessPolicyWindow) if objType == cudaHostFn_t: return sizeof(cyruntime.cudaHostFn_t) if objType == cudaHostNodeParams: return sizeof(cyruntime.cudaHostNodeParams) if objType == cudaHostNodeParamsV2: return sizeof(cyruntime.cudaHostNodeParamsV2) if objType == cudaResourceDesc: return sizeof(cyruntime.cudaResourceDesc) if objType == cudaResourceViewDesc: return sizeof(cyruntime.cudaResourceViewDesc) if objType == cudaPointerAttributes: return sizeof(cyruntime.cudaPointerAttributes) if objType == cudaFuncAttributes: return sizeof(cyruntime.cudaFuncAttributes) if objType == cudaMemLocation: return sizeof(cyruntime.cudaMemLocation) if objType == cudaMemAccessDesc: return sizeof(cyruntime.cudaMemAccessDesc) if objType == cudaMemPoolProps: return sizeof(cyruntime.cudaMemPoolProps) if objType == cudaMemPoolPtrExportData: return sizeof(cyruntime.cudaMemPoolPtrExportData) if objType == cudaMemAllocNodeParams: return sizeof(cyruntime.cudaMemAllocNodeParams) if objType == cudaMemAllocNodeParamsV2: return sizeof(cyruntime.cudaMemAllocNodeParamsV2) if objType == cudaMemFreeNodeParams: return sizeof(cyruntime.cudaMemFreeNodeParams) if objType == cudaMemcpyAttributes: return sizeof(cyruntime.cudaMemcpyAttributes) if objType == cudaOffset3D: return sizeof(cyruntime.cudaOffset3D) if objType == cudaMemcpy3DOperand: return sizeof(cyruntime.cudaMemcpy3DOperand) if objType == cudaMemcpy3DBatchOp: return sizeof(cyruntime.cudaMemcpy3DBatchOp) if objType == CUuuid_st: return sizeof(cyruntime.CUuuid_st) if objType == CUuuid: return sizeof(cyruntime.CUuuid) if objType == cudaUUID_t: return sizeof(cyruntime.cudaUUID_t) if objType == cudaDeviceProp: return sizeof(cyruntime.cudaDeviceProp) if objType == cudaIpcEventHandle_st: return sizeof(cyruntime.cudaIpcEventHandle_st) if objType == cudaIpcEventHandle_t: return sizeof(cyruntime.cudaIpcEventHandle_t) if objType == cudaIpcMemHandle_st: return sizeof(cyruntime.cudaIpcMemHandle_st) if objType == cudaIpcMemHandle_t: return sizeof(cyruntime.cudaIpcMemHandle_t) if objType == cudaMemFabricHandle_st: return sizeof(cyruntime.cudaMemFabricHandle_st) if objType == cudaMemFabricHandle_t: return sizeof(cyruntime.cudaMemFabricHandle_t) if objType == cudaExternalMemoryHandleDesc: return sizeof(cyruntime.cudaExternalMemoryHandleDesc) if objType == cudaExternalMemoryBufferDesc: return sizeof(cyruntime.cudaExternalMemoryBufferDesc) if objType == cudaExternalMemoryMipmappedArrayDesc: return sizeof(cyruntime.cudaExternalMemoryMipmappedArrayDesc) if objType == cudaExternalSemaphoreHandleDesc: return sizeof(cyruntime.cudaExternalSemaphoreHandleDesc) if objType == cudaExternalSemaphoreSignalParams: return sizeof(cyruntime.cudaExternalSemaphoreSignalParams) if objType == cudaExternalSemaphoreWaitParams: return sizeof(cyruntime.cudaExternalSemaphoreWaitParams) if objType == cudaStream_t: return sizeof(cyruntime.cudaStream_t) if objType == cudaEvent_t: return sizeof(cyruntime.cudaEvent_t) if objType == cudaGraphicsResource_t: return sizeof(cyruntime.cudaGraphicsResource_t) if objType == cudaExternalMemory_t: return sizeof(cyruntime.cudaExternalMemory_t) if objType == cudaExternalSemaphore_t: return sizeof(cyruntime.cudaExternalSemaphore_t) if objType == cudaGraph_t: return sizeof(cyruntime.cudaGraph_t) if objType == cudaGraphNode_t: return sizeof(cyruntime.cudaGraphNode_t) if objType == cudaUserObject_t: return sizeof(cyruntime.cudaUserObject_t) if objType == cudaGraphConditionalHandle: return sizeof(cyruntime.cudaGraphConditionalHandle) if objType == cudaFunction_t: return sizeof(cyruntime.cudaFunction_t) if objType == cudaKernel_t: return sizeof(cyruntime.cudaKernel_t) if objType == cudalibraryHostUniversalFunctionAndDataTable: return sizeof(cyruntime.cudalibraryHostUniversalFunctionAndDataTable) if objType == cudaLibrary_t: return sizeof(cyruntime.cudaLibrary_t) if objType == cudaMemPool_t: return sizeof(cyruntime.cudaMemPool_t) if objType == cudaKernelNodeParams: return sizeof(cyruntime.cudaKernelNodeParams) if objType == cudaKernelNodeParamsV2: return sizeof(cyruntime.cudaKernelNodeParamsV2) if objType == cudaExternalSemaphoreSignalNodeParams: return sizeof(cyruntime.cudaExternalSemaphoreSignalNodeParams) if objType == cudaExternalSemaphoreSignalNodeParamsV2: return sizeof(cyruntime.cudaExternalSemaphoreSignalNodeParamsV2) if objType == cudaExternalSemaphoreWaitNodeParams: return sizeof(cyruntime.cudaExternalSemaphoreWaitNodeParams) if objType == cudaExternalSemaphoreWaitNodeParamsV2: return sizeof(cyruntime.cudaExternalSemaphoreWaitNodeParamsV2) if objType == cudaConditionalNodeParams: return sizeof(cyruntime.cudaConditionalNodeParams) if objType == cudaChildGraphNodeParams: return sizeof(cyruntime.cudaChildGraphNodeParams) if objType == cudaEventRecordNodeParams: return sizeof(cyruntime.cudaEventRecordNodeParams) if objType == cudaEventWaitNodeParams: return sizeof(cyruntime.cudaEventWaitNodeParams) if objType == cudaGraphNodeParams: return sizeof(cyruntime.cudaGraphNodeParams) if objType == cudaGraphEdgeData_st: return sizeof(cyruntime.cudaGraphEdgeData_st) if objType == cudaGraphEdgeData: return sizeof(cyruntime.cudaGraphEdgeData) if objType == cudaGraphExec_t: return sizeof(cyruntime.cudaGraphExec_t) if objType == cudaGraphInstantiateParams_st: return sizeof(cyruntime.cudaGraphInstantiateParams_st) if objType == cudaGraphInstantiateParams: return sizeof(cyruntime.cudaGraphInstantiateParams) if objType == cudaGraphExecUpdateResultInfo_st: return sizeof(cyruntime.cudaGraphExecUpdateResultInfo_st) if objType == cudaGraphExecUpdateResultInfo: return sizeof(cyruntime.cudaGraphExecUpdateResultInfo) if objType == cudaGraphDeviceNode_t: return sizeof(cyruntime.cudaGraphDeviceNode_t) if objType == cudaGraphKernelNodeUpdate: return sizeof(cyruntime.cudaGraphKernelNodeUpdate) if objType == cudaLaunchMemSyncDomainMap_st: return sizeof(cyruntime.cudaLaunchMemSyncDomainMap_st) if objType == cudaLaunchMemSyncDomainMap: return sizeof(cyruntime.cudaLaunchMemSyncDomainMap) if objType == cudaLaunchAttributeValue: return sizeof(cyruntime.cudaLaunchAttributeValue) if objType == cudaLaunchAttribute_st: return sizeof(cyruntime.cudaLaunchAttribute_st) if objType == cudaLaunchAttribute: return sizeof(cyruntime.cudaLaunchAttribute) if objType == cudaAsyncCallbackHandle_t: return sizeof(cyruntime.cudaAsyncCallbackHandle_t) if objType == cudaAsyncNotificationInfo: return sizeof(cyruntime.cudaAsyncNotificationInfo) if objType == cudaAsyncNotificationInfo_t: return sizeof(cyruntime.cudaAsyncNotificationInfo_t) if objType == cudaAsyncCallback: return sizeof(cyruntime.cudaAsyncCallback) if objType == cudaSurfaceObject_t: return sizeof(cyruntime.cudaSurfaceObject_t) if objType == cudaTextureDesc: return sizeof(cyruntime.cudaTextureDesc) if objType == cudaTextureObject_t: return sizeof(cyruntime.cudaTextureObject_t) if objType == cudaStreamCallback_t: return sizeof(cyruntime.cudaStreamCallback_t) if objType == GLenum: return sizeof(cyruntime.GLenum) if objType == GLuint: return sizeof(cyruntime.GLuint) if objType == EGLImageKHR: return sizeof(cyruntime.EGLImageKHR) if objType == EGLStreamKHR: return sizeof(cyruntime.EGLStreamKHR) if objType == EGLint: return sizeof(cyruntime.EGLint) if objType == EGLSyncKHR: return sizeof(cyruntime.EGLSyncKHR) if objType == VdpDevice: return sizeof(cyruntime.VdpDevice) if objType == VdpGetProcAddress: return sizeof(cyruntime.VdpGetProcAddress) if objType == VdpVideoSurface: return sizeof(cyruntime.VdpVideoSurface) if objType == VdpOutputSurface: return sizeof(cyruntime.VdpOutputSurface) if objType == cudaStreamAttrValue: return sizeof(cyruntime.cudaStreamAttrValue) if objType == cudaKernelNodeAttrValue: return sizeof(cyruntime.cudaKernelNodeAttrValue) if objType == cudaEglPlaneDesc_st: return sizeof(cyruntime.cudaEglPlaneDesc_st) if objType == cudaEglPlaneDesc: return sizeof(cyruntime.cudaEglPlaneDesc) if objType == cudaEglFrame_st: return sizeof(cyruntime.cudaEglFrame_st) if objType == cudaEglFrame: return sizeof(cyruntime.cudaEglFrame) if objType == cudaEglStreamConnection: return sizeof(cyruntime.cudaEglStreamConnection) raise TypeError("Unknown type: " + str(objType)) cdef int _add_native_handle_getters() except?-1: from cuda.bindings.utils import _add_cuda_native_handle_getter def cudaArray_t_getter(cudaArray_t x): return (x._pvt_ptr[0]) _add_cuda_native_handle_getter(cudaArray_t, cudaArray_t_getter) def cudaArray_const_t_getter(cudaArray_const_t x): return (x._pvt_ptr[0]) _add_cuda_native_handle_getter(cudaArray_const_t, cudaArray_const_t_getter) def cudaMipmappedArray_t_getter(cudaMipmappedArray_t x): return (x._pvt_ptr[0]) _add_cuda_native_handle_getter(cudaMipmappedArray_t, cudaMipmappedArray_t_getter) def cudaMipmappedArray_const_t_getter(cudaMipmappedArray_const_t x): return (x._pvt_ptr[0]) _add_cuda_native_handle_getter(cudaMipmappedArray_const_t, cudaMipmappedArray_const_t_getter) def cudaStream_t_getter(cudaStream_t x): return (x._pvt_ptr[0]) _add_cuda_native_handle_getter(cudaStream_t, cudaStream_t_getter) def cudaEvent_t_getter(cudaEvent_t x): return (x._pvt_ptr[0]) _add_cuda_native_handle_getter(cudaEvent_t, cudaEvent_t_getter) def cudaGraphicsResource_t_getter(cudaGraphicsResource_t x): return (x._pvt_ptr[0]) _add_cuda_native_handle_getter(cudaGraphicsResource_t, cudaGraphicsResource_t_getter) def cudaExternalMemory_t_getter(cudaExternalMemory_t x): return (x._pvt_ptr[0]) _add_cuda_native_handle_getter(cudaExternalMemory_t, cudaExternalMemory_t_getter) def cudaExternalSemaphore_t_getter(cudaExternalSemaphore_t x): return (x._pvt_ptr[0]) _add_cuda_native_handle_getter(cudaExternalSemaphore_t, cudaExternalSemaphore_t_getter) def cudaGraph_t_getter(cudaGraph_t x): return (x._pvt_ptr[0]) _add_cuda_native_handle_getter(cudaGraph_t, cudaGraph_t_getter) def cudaGraphNode_t_getter(cudaGraphNode_t x): return (x._pvt_ptr[0]) _add_cuda_native_handle_getter(cudaGraphNode_t, cudaGraphNode_t_getter) def cudaUserObject_t_getter(cudaUserObject_t x): return (x._pvt_ptr[0]) _add_cuda_native_handle_getter(cudaUserObject_t, cudaUserObject_t_getter) def cudaFunction_t_getter(cudaFunction_t x): return (x._pvt_ptr[0]) _add_cuda_native_handle_getter(cudaFunction_t, cudaFunction_t_getter) def cudaKernel_t_getter(cudaKernel_t x): return (x._pvt_ptr[0]) _add_cuda_native_handle_getter(cudaKernel_t, cudaKernel_t_getter) def cudaLibrary_t_getter(cudaLibrary_t x): return (x._pvt_ptr[0]) _add_cuda_native_handle_getter(cudaLibrary_t, cudaLibrary_t_getter) def cudaMemPool_t_getter(cudaMemPool_t x): return (x._pvt_ptr[0]) _add_cuda_native_handle_getter(cudaMemPool_t, cudaMemPool_t_getter) def cudaGraphExec_t_getter(cudaGraphExec_t x): return (x._pvt_ptr[0]) _add_cuda_native_handle_getter(cudaGraphExec_t, cudaGraphExec_t_getter) def cudaGraphDeviceNode_t_getter(cudaGraphDeviceNode_t x): return (x._pvt_ptr[0]) _add_cuda_native_handle_getter(cudaGraphDeviceNode_t, cudaGraphDeviceNode_t_getter) def cudaAsyncCallbackHandle_t_getter(cudaAsyncCallbackHandle_t x): return (x._pvt_ptr[0]) _add_cuda_native_handle_getter(cudaAsyncCallbackHandle_t, cudaAsyncCallbackHandle_t_getter) def EGLImageKHR_getter(EGLImageKHR x): return (x._pvt_ptr[0]) _add_cuda_native_handle_getter(EGLImageKHR, EGLImageKHR_getter) def EGLStreamKHR_getter(EGLStreamKHR x): return (x._pvt_ptr[0]) _add_cuda_native_handle_getter(EGLStreamKHR, EGLStreamKHR_getter) def EGLSyncKHR_getter(EGLSyncKHR x): return (x._pvt_ptr[0]) _add_cuda_native_handle_getter(EGLSyncKHR, EGLSyncKHR_getter) def cudaEglStreamConnection_getter(cudaEglStreamConnection x): return (x._pvt_ptr[0]) _add_cuda_native_handle_getter(cudaEglStreamConnection, cudaEglStreamConnection_getter) return 0 _add_native_handle_getters()