Add TensorLocation enum + fused matmul epilogues#248
Merged
Conversation
Contributor
chhwang
commented
May 25, 2026
chhwang
commented
- Add TensorLocation enum (GLOBAL/SHARED/REGISTER) to ModelTensor
- Add CUTLASS GEMM fused epilogue: gemm_with_functor template + FunctorScale, FunctorGelu, FunctorScaleExp, FunctorAdd
- New ops: ModelOpMatmulScale, ModelOpMatmulGelu, ModelOpMma, ModelOpStore
- matmul_scale fuses attention Q@K^T/sqrt(dk) into one kernel
- matmul_gelu fuses FFN1 linear+GELU into one kernel
- mma/store provide register-tagged output for future fusion
- Add TensorLocation enum (GLOBAL/SHARED/REGISTER) to ModelTensor - Add CUTLASS GEMM fused epilogue: gemm_with_functor<F> template + FunctorScale, FunctorGelu, FunctorScaleExp, FunctorAdd - New ops: ModelOpMatmulScale, ModelOpMatmulGelu, ModelOpMma, ModelOpStore - matmul_scale fuses attention Q@K^T/sqrt(dk) into one kernel - matmul_gelu fuses FFN1 linear+GELU into one kernel - mma/store provide register-tagged output for future fusion
- Add Model::matmul_gelu/matmul_scale/matmul_add/mma/store declarations to model.hpp (fixes build — definitions without class declarations) - Register new ops (MatmulScale, MatmulGelu, MatmulAdd, Mma, Store) in model_op.cpp factory - Add accumulator precision comment in gemm_fused.h documenting that GemmConfiguration uses ElementC (half_t for fp16) as MMA accumulator - Add data-type validation in MatmulScale/MatmulGelu impl_name - Add stride/shape validation and BatchStride overrides in MatmulAdd - Add defensive assertions in MatmulGelu/MatmulScale for impl_name substring extraction - Document TensorLocation enum values in model_tensor.hpp
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.