Skip to content

Conversation

@chhwang
Copy link
Contributor

@chhwang chhwang commented Jul 31, 2025

  • Fix an allgather bug
  • Fix the nccl.h installation path (now goes under mscclpp/)

@chhwang chhwang requested review from Binyang2014 and Copilot July 31, 2025 07:17
Copy link
Contributor

Copilot AI left a comment

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Pull Request Overview

This PR addresses NCCL fixes including an allgather bug fix and correcting the installation path for nccl.h to go under the mscclpp/ directory. The changes also remove test code that appears to be example code borrowed from NVIDIA documentation.

  • Fix allgather synchronization bug by changing from relaxedSignal() to signal()
  • Update header installation path to place files under mscclpp/ directory
  • Remove NCCL API test code and related CMake configuration

Reviewed Changes

Copilot reviewed 5 out of 6 changed files in this pull request and generated 1 comment.

Show a summary per file
File Description
apps/nccl/test/nccl_api_test.cc Complete removal of NCCL API test file
apps/nccl/test/CMakeLists.txt Complete removal of test CMake configuration
apps/nccl/src/nccl.cu Update header include path to use mscclpp/nccl.h
apps/nccl/src/allgather.hpp Fix synchronization bug by changing relaxedSignal() to signal()
apps/nccl/CMakeLists.txt Update header installation path and remove test subdirectory

@chhwang chhwang enabled auto-merge (squash) August 1, 2025 04:51
@chhwang chhwang merged commit af2098b into main Aug 1, 2025
14 checks passed
@chhwang chhwang deleted the chhwang/fix-nccl branch August 1, 2025 04:56
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

3 participants

SYSTEM_READY >> ...MS