Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

[TL][BugFix] Disable Buffer Vectorization and Add OP Related TL Test Cases #197

Merged
merged 75 commits into from
Sep 28, 2024

Conversation

LeiWang1999
Copy link
Contributor

This pull request includes significant updates to the bitblas library, particularly around new matrix multiplication functionalities and their corresponding tests. The most important changes involve adding new matrix multiplication functions, updating existing ones, and enhancing testing.

New Features and Enhancements:

  • New Matrix Multiplication Functions:

    • Added matmul_macro_tensorcore and matmul_macro_tensorcore_weight_propagation_level_ldmatrix functions to bitblas/ops/general_matmul/tilelang/dense/matmul.py. These functions support tensor core operations with specific configurations for matrix transposition and weight propagation. [1] [2]
  • Function Parameter Updates:

    • Updated the matmul_macro_tensorcore function to include trans_A and trans_B parameters, along with assertions to ensure correct matrix transposition. [1] [2]

Codebase Simplification:

  • Import Statements:
    • Modified import statements in bitblas/ops/general_matmul/tilelang/dense/__init__.py and bitblas/ops/general_matmul/tilelang/dense/matmul.py to include new functions and clean up existing imports. [1] [2]

Testing Enhancements:

  • New Test Functions:

    • Added assert_matmul_macro_tensorcore_correctness and assert_tl_matmul_with_ladder_weight_only_transform_correctness functions in testing/python/operators/test_general_matmul_tilelang_impl.py to validate the new matrix multiplication functionalities.
  • Test Cases:

    • Updated test cases to include scenarios for the new matrix multiplication functions, ensuring comprehensive coverage. [1] [2]

Miscellaneous:

  • Submodule Update:
    • Updated the submodule commit for 3rdparty/tvm to point to a new commit.

These changes collectively enhance the functionality and robustness of the bitblas library, particularly in the context of matrix multiplication operations using tensor cores.

Refactor mma_layout.py and remove unused imports
Add matmul_macro_tensorcore and matmul_macro_tensorcore_weight_propagation_level_ldmatrix to dense/__init__.py
Refactor test_general_matmul_tilelang_impl.py to include additional matmul functions
@LeiWang1999 LeiWang1999 merged commit 033495e into microsoft:main Sep 28, 2024
6 checks passed
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
None yet
Projects
None yet
Development

Successfully merging this pull request may close these issues.

1 participant