This client example demonstrates grouped GEMM fused with FastGELU activation. Grouped GEMM performs multiple independent GEMM operations (with potentially different shapes) in a single kernel launch, and FastGELU is a fast approximation of the GELU activation used in transformers and MLPs.
Mathematical Formulation:
For
- GEMM:
$Y_g = A_g \times B_g$ - FastGELU:
$E_g = \text{FastGELU}(Y_g)$
FastGELU is defined as:
$$
\text{FastGELU}(x) = x \cdot \sigma(1.702 \cdot x)
$$
where
Algorithmic Background:
- Each group can have different matrix sizes and strides.
- The kernel launches a grid covering all groups, with each block assigned to a group.
- FastGELU is applied in the epilogue for each group.
Please follow the instructions in the main Build Guide section as a prerequisite to building and running this example.
cd composable_kernel/client_example/17_grouped_gemm_fastgelu
mkdir build && cd build
cmake -DCMAKE_CXX_COMPILER=/opt/rocm/bin/hipcc ..
make -j
# Example run
./grouped_gemm_fastgeluclient_example/17_grouped_gemm_fastgelu/
├── grouped_gemm_fastgelu.cpp # Main client example: grouped GEMM + FastGELU
├── CMakeLists.txt # Build configuration for the example
- main() (in
grouped_gemm_fastgelu.cpp):
Sets up input matrices for each group, configures GEMM and epilogue parameters, launches the grouped kernel, and verifies the result. - Grouped GEMM kernel invocation:
Uses the Composable Kernel device API to launch grouped GEMM with FastGELU activation.
- Supports multiple groups with different matrix shapes.
- Example parameters can be adjusted in the source for different workloads.
- 15_grouped_gemm: Grouped GEMM in the main example directory
- 04_gemm_add_add_fastgelu: GEMM with FastGELU fusion