fix(CuTeDSL): restore trailing Int<1> dimension in SM90 MMA atom TV L…#3258
Open
zhils wants to merge 1 commit into
Open
fix(CuTeDSL): restore trailing Int<1> dimension in SM90 MMA atom TV L…#3258zhils wants to merge 1 commit into
zhils wants to merge 1 commit into
Conversation
…ayout C when N=8 When N=8, the SM90 WGMMA CLayout_64xN has a trailing Int<N/8> = Int<1> dimension in the C++ CuTe reference (mma_traits_sm90_gmma.hpp L432-L435). The MLIR layout canonicalization was incorrectly dropping this size-1 dimension, causing tv_layout_C to return ((4,8,4),(2,2)) instead of ((4,8,4),(2,2,1)). This fix adds a post-processing check in MmaAtom.tv_layout_C to restore the trailing Int<1> dimension when N=8, matching the C++ implementation and preserving layout structural invariants. Fixes NVIDIA#3254
Contributor
|
There is no invariant that requires This appears to be wasted logic in pursuit of matching (unnecessary) C++ convention. |
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
…ayout C when N=8
When N=8, the SM90 WGMMA CLayout_64xN has a trailing Int<N/8> = Int<1> dimension in the C++ CuTe reference (mma_traits_sm90_gmma.hpp L432-L435). The MLIR layout canonicalization was incorrectly dropping this size-1 dimension, causing tv_layout_C to return ((4,8,4),(2,2)) instead of ((4,8,4),(2,2,1)).
This fix adds a post-processing check in MmaAtom.tv_layout_C to restore the trailing Int<1> dimension when N=8, matching the C++ implementation and preserving layout structural invariants.
Fixes #3254