Avoid FP32 scratch allocation in FP16 communication - #6293
Open
purnawirman wants to merge 1 commit into
Open
Conversation
Summary: The FP16 communication conversion creates an extra full-size FP32 copy, which can cause out-of-memory errors for large tensors. With the rollout flag enabled, convert to FP16 first and clamp in place to avoid that copy. Embeddings remain trainable through the existing custom backward pass. Keep the flag lookup compatible with compilation and test both flag settings. Differential Revision: D119536338
Contributor
|
@purnawirman has exported this pull request. If you are a Meta employee, you can view the originating Diff in D119536338. |
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
1 participant
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
Summary:
The FP16 communication conversion creates an extra full-size FP32 copy, which can cause out-of-memory errors for large tensors.
With the rollout flag enabled, convert to FP16 first and clamp in place to avoid that copy. Embeddings remain trainable through the existing custom backward pass.
Keep the flag lookup compatible with compilation and test both flag settings.
Differential Revision: D119536338