[TL][BugFix] Add implementation of TL Gemm and Fix a bug for TL Jit #195
Conversation
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
…ability and maintainability
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Outdated suggestions cannot be applied.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.
This pull request introduces several new features and improvements to the
bitblaslibrary, including new matrix multiplication implementations, utility functions, and test cases. The most important changes include the addition of new matrix multiplication functions, utility functions for tensor core operations, and test cases to ensure correctness.New Matrix Multiplication Implementations:
bitblas/ops/general_matmul/tilelang/dense/matmul.py: Addedmatmul_blockedandmatmul_macro_tensorcorefunctions for efficient matrix multiplication using tensor cores.Utility Functions for Tensor Core Operations:
bitblas/tl/mma_layout.py: Introduced various layout transformation functions for tensor core operations.bitblas/tl/utils.py: Added utility functionsget_mma_micro_sizeandmake_swizzle_layoutto support tensor core operations. [1] [2]Test Cases:
testing/python/operators/test_general_matmul_tilelang_impl.py: Added test cases to validate the correctness of the new matrix multiplication functions.testing/python/tilelang/test_tilelang_dequantize_gemm.py: Updated test cases to include new tensor core operations and layout transformations. [1] [2] [3]Other Changes:
bitblas/ops/general_matmul/tilelang/dense/__init__.py: Added an import statement for the new matrix multiplication function.3rdparty/tvm: Updated the submodule commit reference.