# Reduce的并行加速

### CUDA

1\. 采用Divergence的支持和Block同步来支持

[![image.png](https://agix.host/uploads/images/gallery/2026-04/mjlimage.png)](Reduce的并行加速/MjLimage.png)[![image.png](https://agix.host/uploads/images/gallery/2026-04/a8kimage.png)](Reduce的并行加速/A8Kimage.png)

2\. 其他的深度优化：https://developer.download.nvidia.com/assets/cuda/files/reduction.pdf

### DSA/ASIC