[PR #711] [CLOSED] Adapt nvshmem37 #724

Closed
opened 2026-10-27 21:05:37 +00:00 by bsakaguchi · 0 comments
Owner

📋 Pull Request Information

Original PR: https://github.com/deepseek-ai/DeepEP/pull/711
Author: @MARD1NO
Created: 8/1/2026
Status: ❌ Closed

Base: main ← Head: adapt_nvshmem37


📝 Commits (4)

  • 389e8ff compatible with cu132, since cuStreamBatchMemOp_v1 has been deprecated
  • 0b32c55 support kimi k3 config
  • 991b786 Merge pull request #3 from MARD1NO/support_kimi_k3
  • efa8d3d apply patch

📊 Changes

6 files changed (+68 additions, -18 deletions)

View changed files

📝 csrc/kernels/backend/cuda_driver.cu (+3 -3)
📝 csrc/kernels/backend/nvshmem.cu (+42 -5)
📝 csrc/kernels/legacy/ibgda_device.cuh (+18 -7)
📝 csrc/kernels/legacy/internode_ll.cu (+2 -2)
📝 csrc/kernels/legacy/launch.cuh (+2 -0)
📝 csrc/utils/lazy_driver.hpp (+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/711 **Author:** [@MARD1NO](https://github.com/MARD1NO) **Created:** 8/1/2026 **Status:** ❌ Closed **Base:** `main` ← **Head:** `adapt_nvshmem37` --- ### 📝 Commits (4) - [`389e8ff`](https://github.com/deepseek-ai/DeepEP/commit/389e8ffa967c37e0989c19c57eebe955a5d44628) compatible with cu132, since cuStreamBatchMemOp_v1 has been deprecated - [`0b32c55`](https://github.com/deepseek-ai/DeepEP/commit/0b32c55a5f9612af6918d093e6514738f3bebd36) support kimi k3 config - [`991b786`](https://github.com/deepseek-ai/DeepEP/commit/991b786a5b68903f5abf479f3d145b1b901705e6) Merge pull request #3 from MARD1NO/support_kimi_k3 - [`efa8d3d`](https://github.com/deepseek-ai/DeepEP/commit/efa8d3dd579ceebdfd4b3a3ab81780f4f2ed7d7f) apply patch ### 📊 Changes **6 files changed** (+68 additions, -18 deletions) <details> <summary>View changed files</summary> 📝 `csrc/kernels/backend/cuda_driver.cu` (+3 -3) 📝 `csrc/kernels/backend/nvshmem.cu` (+42 -5) 📝 `csrc/kernels/legacy/ibgda_device.cuh` (+18 -7) 📝 `csrc/kernels/legacy/internode_ll.cu` (+2 -2) 📝 `csrc/kernels/legacy/launch.cuh` (+2 -0) 📝 `csrc/utils/lazy_driver.hpp` (+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 2026-10-27 21:05:37 +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#724
No description provided.