Example codes of matrix multiplication with Intel AMX
File list:
- int8-gemm-small.cpp: compute int8 matrix multiplication in small sizes
- int8-gemm-large.cpp: compute int8 matrix multiplication in large sizes
- bf16-gemm-small.cpp: compute bf16 matrix multiplication in small sizes
- bf16-gemm-large.cpp: compute bf16 matrix multiplication in large sizes
The examples with small shapes show how to manipulate with tile registers to compute a tiny GEMM. The examples with large shapes show how to make full use of all tile registers and accumulate results of each small block of GEMM.
How to build:
- run
make
How to clean build:
- run
make clean