custom-operator

Here is 1 public repository matching this topic...

triple-mu / fast-ulysses

Ulysses sequence-parallel all-to-all as a torch custom op, moved by the GPU copy engines over NVSHMEM symmetric memory. Zero SM usage; 1.7-2.2x over torch.distributed on NVLink.

cuda ulysses pytorch diffusion-models collective-communication nvshmem sequence-parallelism custom-operator

Updated Aug 7, 2026
Python

Improve this page

Add a description, image, and links to the custom-operator topic page so that developers can more easily learn about it.

Curate this topic

Add this topic to your repo

To associate your repository with the custom-operator topic, visit your repo's landing page and select "manage topics."

Learn more