Add blocked matmul (linalg.mmt4d) op support#265
Draft
MaheshRavishankar wants to merge 1 commit intoiree-org:mainfrom
Draft
Add blocked matmul (linalg.mmt4d) op support#265MaheshRavishankar wants to merge 1 commit intoiree-org:mainfrom
MaheshRavishankar wants to merge 1 commit intoiree-org:mainfrom
Conversation
Add BlockedMatmulNode for tiled matrix multiplication that lowers to linalg.mmt4d via torch_c casts: LHS logical [M0, K0, M1, K1] x RHS logical [K0, N0, K1, N1] -> OUT [M0, N0, M1, N1] RHS must be specified with transposed strides (physical [N0, K0, N1, K1]) matching linalg.mmt4d's expected layout. Non-transposed RHS returns a NotImplemented error. The emitter casts torch tensors to builtin tensors (torch_c), applies linalg.fill + linalg.mmt4d, and casts the result back. No permute ops are needed since the physical layout is used directly. New files: - BlockedMatmulAttr (attributes) - BlockedMatmulNode (node with validation and shape inference) - ASM emitter with getBuiltinTensorTypeAsm helper - Lit test verifying MLIR structure, flow compilation, single dispatch - E2E sample verifying numerical correctness on CPU Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com> Signed-off-by: MaheshRavishankar <mahesh.ravishankar@gmail.com>
MaheshRavishankar
force-pushed
the
users/MaheshRavishankar/blockedMatmul
branch
from
7966694 to
5cb5565
Compare
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
1 participant
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
Add BlockedMatmulNode for tiled matrix multiplication that lowers to linalg.mmt4d via torch_c casts:
LHS logical [M0, K0, M1, K1] x RHS logical [K0, N0, K1, N1]
-> OUT [M0, N0, M1, N1]
RHS must be specified with transposed strides (physical [N0, K0, N1, K1]) matching linalg.mmt4d's expected layout. Non-transposed RHS returns a NotImplemented error.
The emitter casts torch tensors to builtin tensors (torch_c), applies linalg.fill + linalg.mmt4d, and casts the result back. No permute ops are needed since the physical layout is used directly.
New files: