Open
Conversation
This would happen with memoryless NUMA domains, which would cause the received cuMem SHM descriptor to be interpreted as a file path.
Member
|
Thanks for the report. I have the feeling that we could have fixed the issue though, rather than failing. |
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
Description
With memoryless NUMA domains (e.g. if you don't populate all of the memory channels on an EPYC CPU), NCCL IPC has a type confusion that results in this error:
This happens with memoryless NUMA domains, which would cause the received cuMem SHM descriptor union to be interpreted as a /dev/shm file path. This PR checks the union tag and turns it into a clear warning.
Related Issues
I have a more detailed writeup here: pytorch/pytorch#152302 (comment)
Changes & Impact
N/A
Performance Impact
N/A