The SamplerDPMPP_3M_SDE node creates a DPM++ 3M SDE sampler for use in the sampling process. This sampler uses a third-order multistep stochastic differential equation method with configurable noise parameters. The node allows you to choose whether noise calculations are performed on the GPU or CPU.

## Inputs

| Parameter | Description | Data Type | Required | Range |
| --- | --- | --- | --- | --- |
| `eta` | Controls the stochasticity of the sampling process (default: 1.0) | FLOAT | Yes | 0.0 - 100.0 |
| `s_noise` | Controls the amount of noise added during sampling (default: 1.0) | FLOAT | Yes | 0.0 - 100.0 |
| `noise_device` | Selects the device for noise calculations, either GPU or CPU (default: "gpu") | COMBO | Yes | "gpu"<br>"cpu" |

## Outputs

| Output Name | Description | Data Type |
| --- | --- | --- |
| `sampler` | Returns a configured sampler object for use in sampling workflows | SAMPLER |

> This documentation was AI-generated. If you find any errors or have suggestions for improvement, please feel free to contribute! [Edit on GitHub](https://github.com/Comfy-Org/embedded-docs/blob/main/comfyui_embedded_docs/docs/SamplerDPMPP_3M_SDE/en.md)

---
**Source fingerprint (SHA-256):** `d25f8f69211ebca5aca8d1925d53f90a1bee3f883239635bffc107b5fdae8b36`