mirror of
https://github.com/deepseek-ai/DeepEP.git
synced 2026-11-05 00:47:32 +00:00
[PR #438] [MERGED] Support CUDA Graph for internode dispatch normal kernel #570
Labels
No labels
pull-request
No milestone
No project
No assignees
1 participant
Notifications
Due date
No due date set.
Dependencies
No dependencies set.
Reference
deepseek-ai/DeepEP#570
Loading…
Add table
Add a link
Reference in a new issue
No description provided.
Delete branch "%!s()"
Deleting a branch is permanent. Although the deleted branch may continue to exist for a short time before it actually gets removed, it CANNOT be undone in most cases. Continue?
📋 Pull Request Information
Original PR: https://github.com/deepseek-ai/DeepEP/pull/438
Author: @yifeizhang-c
Created: 9/30/2025
Status: ✅ Merged
Merged: 11/5/2025
Merged by: @sphish
Base:
main← Head:enable-internode-cuda-graph📝 Commits (1)
d5e6717Enable CUDA Graph for internode dispatch📊 Changes
6 files changed (+93 additions, -37 deletions)
View changed files
📝
csrc/deep_ep.cpp(+30 -22)📝
csrc/deep_ep.hpp(+1 -0)📝
csrc/kernels/api.cuh(+2 -0)📝
csrc/kernels/internode.cu(+37 -9)📝
deep_ep/buffer.py(+5 -6)📝
tests/test_internode.py(+18 -0)📄 Description
Support CUDA Graph for internode dispatch kernels with the same logic as what has been done for intranode dispatch kernels.
🔄 This issue represents a GitHub Pull Request. It cannot be merged through Gitea due to API limitations.