[fix] MD-TRT: rebind NCCL communicator after execution context invalidation - #4464
Open
apbose wants to merge 1 commit into
Open
[fix] MD-TRT: rebind NCCL communicator after execution context invalidation#4464apbose wants to merge 1 commit into
apbose wants to merge 1 commit into
Conversation
bind_nccl_comm() attaches the communicator to the IExecutionContext. disable_profiling() and set_resource_allocation_strategy() dropped that context without clearing nccl_initialized, so the lazy re-bind in execute_engine.cpp never fired and the replacement context ran collectives with no communicator (~98% output mismatch). Mirrors runtime_settings(). Verified on 2 GPUs: both new tests fail before the fix and pass after.
apbose
changed the title
Collaborator
Author
|
/nightly |
Collaborator
Author
|
/test nightly |
Collaborator
Author
|
/test full |
lanluo-nvidia
added
ci: full
ci: nightly
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
bind_nccl_comm() attaches the communicator to the IExecutionContext. disable_profiling() and set_resource_allocation_strategy() dropped that context without clearing nccl_initialized, so the lazy re-bind in execute_engine.cpp never fired and the replacement context ran collectives with no communicator (~98% output mismatch). Mirrors runtime_settings().