Avoid aten::select calls in fbgemm::permute_multi_embedding #5017
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
Summary: Traces for ads models show an absurd amount of show an absurd amount of aten::select calls inside fbgemm::permute_multi_embedding. Use dataptr with the stride to avoid these select calls (indexing into a tensor turns into an aten::select call). Testing with the model shows decrease in trace time from 6MS to 70 uS Differential Revision: D83609157
✅ Deploy Preview for pytorch-fbgemm-docs ready!
To edit notification comments on pull requests, go to your Netlify project configuration. |
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Outdated suggestions cannot be applied.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:
Traces for ads models show an absurd amount of show an absurd amount of aten::select calls inside fbgemm::permute_multi_embedding. Use dataptr with the stride to avoid these select calls (indexing into a tensor turns into an aten::select call).
Testing with the model shows decrease in trace time from 6MS to 70 uS
Differential Revision: D83609157