deepseek-ai/DeepEP
1
0
Fork 0
mirror of https://github.com/deepseek-ai/DeepEP.git synced 2026-11-05 00:47:32 +00:00
Code Issues 308 Projects Packages Wiki Activity

[PR #752] [MERGED] Ensure that the last get rings the doorbell #756

New issue
Closed
opened 2026-10-27 21:05:59 +00:00 by bsakaguchi · 0 comments
bsakaguchi commented 2026-10-27 21:05:59 +00:00
Owner
Copy link

📋 Pull Request Information

Original PR: https://github.com/deepseek-ai/DeepEP/pull/752
Author: @tianr22
Created: 9/16/2026
Status: ✅ Merged
Merged: 9/16/2026
Merged by: @LyricZhao

Base: main ← Head: engram-db

📝 Commits (3)

📊 Changes

1 file changed (+14 additions, -7 deletions)

View changed files

📝 deep_ep/include/deep_ep/impls/engram_fetch.cuh (+14 -7)

📄 Description

NCCL v2.32 对于某些自带mcst的platform，会跳过flush_async中发送DUMP WQE
之前Engram fetch的实现依赖flush_async ring doobell, 在nccl v2.32中可能永远不会ring doorbell导致卡死，修复该问题

🔄 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/752 **Author:** [@tianr22](https://github.com/tianr22) **Created:** 9/16/2026 **Status:** ✅ Merged **Merged:** 9/16/2026 **Merged by:** [@LyricZhao](https://github.com/LyricZhao) **Base:** `main` ← **Head:** `engram-db` --- ### 📝 Commits (3) - [`53820ed`](https://github.com/deepseek-ai/DeepEP/commit/53820edcb043df891397a4023828693bdd59a120) Is last get - [`22a574f`](https://github.com/deepseek-ai/DeepEP/commit/22a574f12d58c20f2b990e4d8c6cb0b5b871db1d) ring DB before flush_async - [`ed9bb37`](https://github.com/deepseek-ai/DeepEP/commit/ed9bb3741e67f176e07ce75803ac6b26d9a84db8) update comment ### 📊 Changes **1 file changed** (+14 additions, -7 deletions) <details> <summary>View changed files</summary> 📝 `deep_ep/include/deep_ep/impls/engram_fetch.cuh` (+14 -7) </details> ### 📄 Description NCCL v2.32 对于某些自带mcst的platform，会跳过flush_async中发送DUMP WQE 之前Engram fetch的实现依赖flush_async ring doobell, 在nccl v2.32中可能永远不会ring doorbell导致卡死，修复该问题 --- <sub>🔄 This issue represents a GitHub Pull Request. It cannot be merged through Gitea due to API limitations.</sub>
bsakaguchi 2026-10-27 21:05:59 +00:00
Sign in to join this conversation.
No Branch/Tag specified
Branches Tags
main
hybrid-ep
chore/remove-clang-format-yapf
v2.5
nvDev
try-fix-cu128
mori-ep
antgroup-opt
compat-old-torch
add-linter
eager-rdma
v1.2.1
Labels
Clear labels   
pull-request
Mirrored from GitHub Pull Request

No labels
pull-request
Milestone
Clear milestone
No items
No milestone
Projects
Clear projects
No items
No project
Assignees
Clear assignees
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#756
No description provided.