default_mma_core_simt.h File Reference

Defines basic properties needed by CTA-level GEMMs assuming expectations about data layout of the global memory fragments, data types, and internal tile sizes. More...

#include "cutlass/cutlass.h"
#include "cutlass/array.h"
#include "cutlass/fast_math.h"
#include "cutlass/numeric_types.h"
#include "cutlass/matrix_shape.h"
#include "cutlass/transform/pitch_linear_thread_map.h"
#include "cutlass/transform/threadblock/regular_tile_iterator_pitch_linear.h"
#include "cutlass/transform/threadblock/regular_tile_iterator_pitch_linear_2dthreadtile.h"
#include "cutlass/gemm/warp/mma_simt_policy.h"
#include "cutlass/gemm/warp/mma_simt.h"
#include "cutlass/gemm/threadblock/default_mma_core.h"

Include dependency graph for default_mma_core_simt.h:

This graph shows which files directly or indirectly include this file:

Go to the source code of this file.

Classes
struct	cutlass::gemm::threadblock::DefaultMmaCore< Shape_, WarpShape_, GemmShape< 1, 1, 1 >, ElementA_, layout::ColumnMajor, ElementB_, layout::RowMajor, ElementC_, LayoutC_, arch::OpClassSimt, 2, Operator_ >
struct	cutlass::gemm::threadblock::DefaultMmaCore< Shape_, WarpShape_, GemmShape< 1, 1, 1 >, ElementA_, layout::RowMajor, ElementB_, layout::ColumnMajor, ElementC_, LayoutC_, arch::OpClassSimt, 2, Operator_ >
struct	cutlass::gemm::threadblock::DefaultMmaCore< Shape_, WarpShape_, GemmShape< 1, 1, 1 >, ElementA_, layout::RowMajor, ElementB_, layout::RowMajor, ElementC_, LayoutC_, arch::OpClassSimt, 2, Operator_ >
struct	cutlass::gemm::threadblock::DefaultMmaCore< Shape_, WarpShape_, GemmShape< 1, 1, 1 >, ElementA_, layout::ColumnMajor, ElementB_, layout::ColumnMajor, ElementC_, LayoutC_, arch::OpClassSimt, 2, Operator_ >
struct	cutlass::gemm::threadblock::DefaultMmaCore< Shape_, WarpShape_, GemmShape< 1, 1, 4 >, int8_t, layout::ColumnMajor, int8_t, layout::RowMajor, ElementC_, LayoutC_, arch::OpClassSimt, 2, Operator_ >
struct	cutlass::gemm::threadblock::DefaultMmaCore< Shape_, WarpShape_, GemmShape< 1, 1, 4 >, int8_t, layout::RowMajor, int8_t, layout::ColumnMajor, ElementC_, LayoutC_, arch::OpClassSimt, 2, Operator_ >
	Partial specialization: More...
struct	cutlass::gemm::threadblock::DefaultMmaCore< Shape_, WarpShape_, GemmShape< 1, 1, 4 >, int8_t, layout::RowMajor, int8_t, layout::RowMajor, ElementC_, LayoutC_, arch::OpClassSimt, 2, Operator_ >
	Partial specialization: More...
struct	cutlass::gemm::threadblock::DefaultMmaCore< Shape_, WarpShape_, GemmShape< 1, 1, 4 >, int8_t, layout::ColumnMajor, int8_t, layout::ColumnMajor, ElementC_, LayoutC_, arch::OpClassSimt, 2, Operator_ >
	Partial specialization: More...

Namespaces
	cutlass
	cutlass::gemm
	cutlass::gemm::threadblock
	cutlass::gemm::threadblock::detail

Functions
template<typename WarpShape >
constexpr int	cutlass::gemm::threadblock::detail::simt_get_warp_threads_m ()
constexpr int	cutlass::gemm::threadblock::detail::simt_transpose_padding (int threads, int crosswise, int size_in_bits)
	Computes padding in shared memory to perform efficient transpose without bank conflicts. More...

Detailed Description

Partial specializations for threadblock::Mma operations targeting simt instructions.