Fix TP + TransformerEngine issue. #1341
Open
+1
−1
Conversation
This file contains 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
wang2yn84
requested review from
gobbleturk,
khatwanimohit,
bvandermoon,
vipannalla,
RissyRan,
richjames0,
rni418 and
gagika
as code owners
|
Adding @aireen for review |
aireenmei
reviewed
Mar 6, 2025
There was a problem hiding this comment.
I'm not familiar with the impact of this change, seems orthoganol to any data input. I suggest to have @gobbleturk or @khatwanimohit to take a look.
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Outdated suggestions cannot be applied.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.
The data sharding applies to the data pipeline while TP applies to model parallelism. TP + TransformerEngine suffers from this and causes training hang due to much longer NCCL communication. It's verified by running llama 8B on 16 nodes and works fine.
FIXES: b/384441280
Checklist
Before submitting this PR, please make sure (put X in square brackets):