[AICOMRCCL-355] Enable threshold-based p2p-batching #3000
Conversation
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
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Outdated suggestions cannot be applied.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.
Motivation
Address regression on MB/GB range for alltoall due when enabling p2p-batching on all message sizes.
Technical Details
P2p-batching is enabled only if sendBytes and recvBytes are below a per-rank size threshold message size which is 64KB by default.
Alltoallv hang was addressed by adding sendBytes == recvBytes requirement for p2p-batching and adding
batchP2Pentry in wipBatch to prevent batching together ops that are below and above the threshold.The user can specify a threshold through the
RCCL_P2P_BATCH_THRESHOLDenv. variable.JIRA ID
AICOMRCCL-355
Test Plan
Correctness tests on 4/8 node gfx950 and 4/8/16/32/33 node gfx942.
Test Result
gfx950 tests show improved performance.
Submission Checklist