[PR #247] [MERGED] Optimize intranode combine. #481

Closed
opened 2026-10-26 10:01:16 +00:00 by bsakaguchi · 0 comments
Owner

📋 Pull Request Information

Original PR: https://github.com/deepseek-ai/DeepEP/pull/247
Author: @sphish
Created: 6/23/2025
Status: ✅ Merged
Merged: 6/24/2025
Merged by: @LyricZhao

Base: main ← Head: intra-combine


📝 Commits (4)

  • e7dd988 Increase the test round.
  • a2db660 Add warp synchronization.
  • 2cbd3a6 Shuffle the send warps.
  • e3f2539 Add time elapsed into bench result.

📊 Changes

3 files changed (+8 additions, -8 deletions)

View changed files

📝 csrc/kernels/intranode.cu (+3 -3)
📝 tests/test_intranode.py (+4 -4)
📝 tests/utils.py (+1 -1)

📄 Description

No description provided


🔄 This issue represents a GitHub Pull Request. It cannot be merged through Gitea due to API limitations.

## 📋 Pull Request Information **Original PR:** https://github.com/deepseek-ai/DeepEP/pull/247 **Author:** [@sphish](https://github.com/sphish) **Created:** 6/23/2025 **Status:** ✅ Merged **Merged:** 6/24/2025 **Merged by:** [@LyricZhao](https://github.com/LyricZhao) **Base:** `main` ← **Head:** `intra-combine` --- ### 📝 Commits (4) - [`e7dd988`](https://github.com/deepseek-ai/DeepEP/commit/e7dd9884de192d65d4d27693843f562466abb2f4) Increase the test round. - [`a2db660`](https://github.com/deepseek-ai/DeepEP/commit/a2db660caf2a08dd9574691a0e1a193e520022b7) Add warp synchronization. - [`2cbd3a6`](https://github.com/deepseek-ai/DeepEP/commit/2cbd3a652eec7ae523383971f5c6c833ccf5c01d) Shuffle the send warps. - [`e3f2539`](https://github.com/deepseek-ai/DeepEP/commit/e3f25394aee9c50cfafd37f4ed1c048b732b7831) Add time elapsed into bench result. ### 📊 Changes **3 files changed** (+8 additions, -8 deletions) <details> <summary>View changed files</summary> 📝 `csrc/kernels/intranode.cu` (+3 -3) 📝 `tests/test_intranode.py` (+4 -4) 📝 `tests/utils.py` (+1 -1) </details> ### 📄 Description _No description provided_ --- <sub>🔄 This issue represents a GitHub Pull Request. It cannot be merged through Gitea due to API limitations.</sub>
bsakaguchi 2026-10-26 10:01:16 +00:00
Sign in to join this conversation.
No labels
pull-request
No milestone
No project
No assignees
1 participant
Notifications
Due date
The due date is invalid or out of range. Please use the format "yyyy-mm-dd".

No due date set.

Dependencies

No dependencies set.

Reference
deepseek-ai/DeepEP#481
No description provided.