add triton based Softmax and unit tests#475
Open
lessw2020 wants to merge 1 commit intofacebookresearch:mainfrom
Open
add triton based Softmax and unit tests#475lessw2020 wants to merge 1 commit intofacebookresearch:mainfrom
lessw2020 wants to merge 1 commit intofacebookresearch:mainfrom
Conversation
facebook-github-bot
added
the
CLA Signed
rohan-varma
reviewed
Sep 26, 2023
Contributor
rohan-varma
left a comment
rohan-varma
left a comment
There was a problem hiding this comment.
Curious, when would users have to author this sort of performant triton-based softmax, versus getting these perf wins ootb with torch.compile?
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
3 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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
Summary:
This PR is to add a Triton based softmax, which supports both fwd and bwd as well as multi-dim inputs.
This kernel outperforms the PT native softmax in the fwd pass and thus might be favored esp. for inference situations.
It's also a first opportunity to consider expanding the TorchMM offerings to include triton kernels.
Thus, I have listed this under a separate Triton directory in order to separate it from the rest of the codebase which is PyTorch base.
Test plan:
Tested it in end to end training (short run).
Added a number of unit tests which compare it directly with the torch.softmax and tests both forward and backward passes, as well as 2d and multi-dim inputs and fp32 and bf16 dtypes.