FP8 tensorwise GEMM improvement#2585
Closed
jiawenliu64 wants to merge 1 commit into
Closed
Conversation
Contributor
|
This pull request was exported from Phabricator. Differential Revision: D57263833 |
❌ Deploy Preview for pytorch-fbgemm-docs failed.
|
Summary: This Diff improves FP8 tensorwise GEMM performance with scalar scale broadcasting along with EVT - **FP8 CUTLASS tensorwise is 15% faster than FP8 CUTLASS rowwise GEMM on average (up to 2.7x faster)** - Before this Diff, FP8 tensorwise CUTLASS GEMM is similar to FP8 rowwise - FP8 tensorwise would be useful in models that are not very sensitive to numeric variance, while require latency/throughput boost (e.g., LLM with 7B, LDM, etc) - More data can be found in [this data sheet](https://docs.google.com/spreadsheets/d/1SYSjYqWeESasl9LII-qHLHMvaNAXlV5wmCV9BWIrKBc/edit?usp=sharing) {F1636238658} TODO 1. Merge two FP8 tensorwise GEMMs into one 2. Support e5m2 for bwd and bias Differential Revision: D57263833
Contributor
|
This pull request was exported from Phabricator. Differential Revision: D57263833 |
Contributor
|
This pull request has been merged in 17a4e18. |
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
2 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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
As title