[PR #213] [MERGED] Add automatic warp count control for low-latency kernels #469

New issue

Closed

opened 2026-10-26 10:01:06 +00:00 by bsakaguchi · 0 comments

bsakaguchi commented

2026-10-26 10:01:06 +00:00

Owner

📋 Pull Request Information

Original PR: https://github.com/deepseek-ai/DeepEP/pull/213
Author: @LyricZhao
Created: 6/16/2025
Status: ✅ Merged
Merged: 6/16/2025
Merged by: @LyricZhao

Base: main ← Head: maximum-sms

📝 Commits (3)

632c81f Add automatic warp count control for low-latency dispatch
72beb15 Add automatic warp count control for low-latency combine
b09308b More assertions

📊 Changes

6 files changed (+83 additions, -65 deletions)

View changed files

📝 csrc/config.hpp (+2 -2)
📝 csrc/deep_ep.cpp (+6 -2)
📝 csrc/deep_ep.hpp (+1 -0)
📝 csrc/kernels/api.cuh (+4 -2)
📝 csrc/kernels/internode_ll.cu (+67 -56)
📝 csrc/kernels/utils.cuh (+3 -3)

📄 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/213 **Author:** [@LyricZhao](https://github.com/LyricZhao) **Created:** 6/16/2025 **Status:** ✅ Merged **Merged:** 6/16/2025 **Merged by:** [@LyricZhao](https://github.com/LyricZhao) **Base:** `main` ← **Head:** `maximum-sms` --- ### 📝 Commits (3) - [`632c81f`](https://github.com/deepseek-ai/DeepEP/commit/632c81f1d7f2c1434c9378794801397265046f42) Add automatic warp count control for low-latency dispatch - [`72beb15`](https://github.com/deepseek-ai/DeepEP/commit/72beb158271121545f4efc035994999a025bc648) Add automatic warp count control for low-latency combine - [`b09308b`](https://github.com/deepseek-ai/DeepEP/commit/b09308b731160f4c539ed36d12642a0d71e4096d) More assertions ### 📊 Changes **6 files changed** (+83 additions, -65 deletions) <details> <summary>View changed files</summary> 📝 `csrc/config.hpp` (+2 -2) 📝 `csrc/deep_ep.cpp` (+6 -2) 📝 `csrc/deep_ep.hpp` (+1 -0) 📝 `csrc/kernels/api.cuh` (+4 -2) 📝 `csrc/kernels/internode_ll.cu` (+67 -56) 📝 `csrc/kernels/utils.cuh` (+3 -3) </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 10:01:06 +00:00