Open
Conversation
Contributor
Author
|
@d4l3k Could you help review this? |
d4l3k
reviewed
Feb 6, 2026
|
|
||
| std::shared_ptr<AllreduceSharedMemoryData> shmData; | ||
|
|
||
| std::mutex shmDataMutex; |
Member
There was a problem hiding this comment.
What happens if there's multiple gloo process groups? Does that cause issues at all?
Member
There was a problem hiding this comment.
also can we put this under shmData?
Contributor
Author
There was a problem hiding this comment.
- For multiple process groups scenario, I used gloo context's address to generate unique ID for shm name. In that case different group will use different shm buffer to do allreduce op. I've verified with a test with multiple process groups with pytorch and it passed.
- I think we could not put this under shmData. In the first run shmData is not initialized(nullptr), if there are multiple threads reaching this point, we need to ensure the initialization work is done only by one thread here.
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
2 participants
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
This is to fix CI failure of pytorch in bump PR pytorch/pytorch#172297.
In async mode shmData should be occupied exclusively. We added lock for shmData to make it thread safe and used unique tag to do synchronization among different ranks.