[LLVMGPU] Prefer column major MFMA to vectorize stores #19919
+229
−23
Conversation
This file contains 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
For MMA intrinsics with symmetric register layouts we have the freedom to choose between row major and column major output layouts by simply swapping the input operands. This adds a flag to `MMAAttr` that allows specifying whether the result should be column major. A follow up PR will turn the column major variant on by default for mfma as this enables vectorization of writes in non-transpose cases.
The layout of MFMA ops allows producing a column major result without changing the inputs. Because the layout of all MFMA results vectorize along the column, this is always better unless we are already trying to store transposed.
qedawkins
requested review from
MaheshRavishankar,
kuhar and
antiagainst
as code owners
bjacob
reviewed
Feb 7, 2025
| @@ -172,12 +173,40 @@ static std::optional<GPUMMASchedule> getMmaScheduleFromProblemAndTarget( | |||
| return schedule; | |||
| } | |||
|
|
|||
| /// A coarse approximation of when the given value |v| is consumed column | |||
There was a problem hiding this comment.
Can this "coarseness" be refined into something more precise in any useful way?
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
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Outdated suggestions cannot be applied.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.
The layout of MFMA ops allows producing a column major result without
changing the inputs. Because the layout of all MFMA results vectorize
along the column, this is always better unless we are already trying to
store transposed.
depends on #19860