fix(distributed): call destroy_process_group() on exit to prevent NCC…#1391
Open
mandanajignesh-byte wants to merge 1 commit into
Open
Conversation
…L resource leak In distributed runs (accelerate launch / multi-GPU), when evaluation hits an exception, process shutdown emits: WARNING: destroy_process_group() was not called before program exit, which can leak resources. Root cause: Accelerator registers no atexit handler, so distributed teardown never happens on error paths. Fix: wrap the evaluation loop and result printing in a try/finally block. The finally block calls accelerator.state.destroy_process_group() when Accelerator is active, or torch.distributed.destroy_process_group() directly when running in a pre-initialized distributed context. wandb logging runs inside try before finally to preserve correct ordering. Fixes EvolvingLMMs-Lab#1240
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
…L resource leak
In distributed runs (accelerate launch / multi-GPU), when evaluation hits an exception, process shutdown emits:
WARNING: destroy_process_group() was not called before program exit,
which can leak resources.
Root cause: Accelerator registers no atexit handler, so distributed teardown never happens on error paths.
Fix: wrap the evaluation loop and result printing in a try/finally block. The finally block calls accelerator.state.destroy_process_group() when Accelerator is active, or torch.distributed.destroy_process_group() directly when running in a pre-initialized distributed context. wandb logging runs inside try before finally to preserve correct ordering.
Fixes #1240
Summary
In scope
Out of scope
Validation
Risk / Compatibility
Type of Change