gpu::xetla::group::gemm_t< compute_policy, tile_shape, mem_desc_a, mem_desc_b, pre_processing, enable > Class Template Reference

Gemm functor. More...

#include <api.hpp>

Detailed Description

template<typename compute_policy, typename tile_shape, typename mem_desc_a, typename mem_desc_b, typename pre_processing = pre_processing_default_t<tile_shape, compute_policy::arch_tag>, class enable = void>
class gpu::xetla::group::gemm_t< compute_policy, tile_shape, mem_desc_a, mem_desc_b, pre_processing, enable >

Gemm functor.

Template Parameters

compute_policy	Is the compute algorithm of gemm implementation.
tile_shape	Is the workgroup-level tile shape.
mem_desc_a	Is the memory descriptor of matA.
mem_desc_b	Is the memory descriptor of matB.
pre_processing	Is the pre_processing functor.