[BACKEND] Add support converting MMAV3 accumulator layout to fp8 dot_operand#3370
Merged
ThomasRaoux merged 3 commits intotriton-lang:mainfrom Mar 14, 2024
Merged
Conversation
…operand This will allow us generating full fp8 attention kernel by having chain of dots in fp8 without going through shared memory.
Collaborator
|
Awesome! Maybe we should credit the https://research.colfax-intl.com/wp-content/uploads/2023/12/colfax-flashattention.pdf in the commit description? I see it's already in the comments of the code so no problem there :) |
Collaborator
Author
Good point, yes it should be added to the commit description as well. |
…6 as it was implicitly converted to fp32
ptillet
approved these changes
Mar 14, 2024
ThomasRaoux
pushed a commit
that referenced
this pull request
Mar 14, 2024
htyu
pushed a commit
to htyu/triton
that referenced
this pull request
Mar 20, 2024
…operand (triton-lang#3370) Implement layout conversion invented by Ganesh Bikshandi and Jay Shah described in the following paper: https://research.colfax-intl.com/wp-content/uploads/2023/12/colfax-flashattention.pdf This will allow us generating full fp8 attention kernel by having chain of dots in fp8 without going through shared memory.
htyu
pushed a commit
to htyu/triton
that referenced
this pull request
Mar 20, 2024
…ang#3372) Fix AMD backend failure caused by triton-lang#3362 and triton-lang#3370
karupayun
pushed a commit
to openxla/triton
that referenced
this pull request
Apr 3, 2024
…operand (triton-lang#3370) Implement layout conversion invented by Ganesh Bikshandi and Jay Shah described in the following paper: https://research.colfax-intl.com/wp-content/uploads/2023/12/colfax-flashattention.pdf This will allow us generating full fp8 attention kernel by having chain of dots in fp8 without going through shared memory.
karupayun
pushed a commit
to openxla/triton
that referenced
this pull request
Apr 3, 2024
…ang#3372) Fix AMD backend failure caused by triton-lang#3362 and triton-lang#3370
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.
Implement layout conversion invented by Ganesh Bikshandi and Jay Shah described in the following paper:
https://research.colfax-intl.com/wp-content/uploads/2023/12/colfax-flashattention.pdf
This will allow us generating full fp8 attention kernel by having chain of dots in fp8 without going through shared memory.