gpu::xetla::subgroup::elemwise_reduce_op_t< reduce_kind, dtype_in, arch_tag, enable > Struct Template Reference

Is the element-wise reduce op functor. More...

#include <tile_op_functor.hpp>

Detailed Description

template<reduce_op reduce_kind, typename dtype_in, gpu_arch arch_tag, class enable = void>
struct gpu::xetla::subgroup::elemwise_reduce_op_t< reduce_kind, dtype_in, arch_tag, enable >

Is the element-wise reduce op functor.

Load one buffer from memory and get another from matAcc, element-wise reduce and update the output in place. Used in epilogue::tile_op or chained_tile_op.

Template Parameters

reduce_kind	Is the reduce type, can be sum, prod, min and max.
dtype_in	Is the memory side buffer data type.
arch_tag	Is the hardware architecture tag.