CUTLASS: default_mma_core_simt.h File Reference

| | CUTLASS

CUDA Templates for Linear Algebra Subroutines and Solvers |

• Main Page

• Modules

• Namespaces

• Classes

• Files

• File List

• File Members

• include

• cutlass

• gemm

• threadblock

Classes | Namespaces | Functions

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.](default mma core__simt_8h_source.html)

|

Classes

| | struct | [cutlass::gemm::threadblock::DefaultMmaCore< Shape_, WarpShape_, GemmShape< 1, 1, 1 >, ElementA_, layout::ColumnMajor, ElementB_, layout::RowMajor, ElementC_, LayoutC_, arch::OpClassSimt, 2, Operator_ >](structcutlass_1_1gemm_1_1threadblock_1_1DefaultMmaCore_3_01Shape 00_01WarpShape 00_01GemmShafafd5c61db86cbfe90863578ddd11092.html) | | | | struct | [cutlass::gemm::threadblock::DefaultMmaCore< Shape_, WarpShape_, GemmShape< 1, 1, 1 >, ElementA_, layout::RowMajor, ElementB_, layout::ColumnMajor, ElementC_, LayoutC_, arch::OpClassSimt, 2, Operator_ >](structcutlass_1_1gemm_1_1threadblock_1_1DefaultMmaCore_3_01Shape 00_01WarpShape 00_01GemmSha8da7a0cfbbe859b701fdd9f2b8566aa7.html) | | | | struct | [cutlass::gemm::threadblock::DefaultMmaCore< Shape_, WarpShape_, GemmShape< 1, 1, 1 >, ElementA_, layout::RowMajor, ElementB_, layout::RowMajor, ElementC_, LayoutC_, arch::OpClassSimt, 2, Operator_ >](structcutlass_1_1gemm_1_1threadblock_1_1DefaultMmaCore_3_01Shape 00_01WarpShape 00_01GemmSha84e9f8afb6a4ca9f5dcd219b182d16e7.html) | | | | struct | [cutlass::gemm::threadblock::DefaultMmaCore< Shape_, WarpShape_, GemmShape< 1, 1, 1 >, ElementA_, layout::ColumnMajor, ElementB_, layout::ColumnMajor, ElementC_, LayoutC_, arch::OpClassSimt, 2, Operator_ >](structcutlass_1_1gemm_1_1threadblock_1_1DefaultMmaCore_3_01Shape 00_01WarpShape 00_01GemmShab94a11a77dd0565102710907089acee0.html) | | | | 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_ >](structcutlass_1_1gemm_1_1threadblock_1_1DefaultMmaCore_3_01Shape 00_01WarpShape 00_01GemmSha34a52cc7b2942e8c290f0032b6779b52.html) | | | | 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_ >](structcutlass_1_1gemm_1_1threadblock_1_1DefaultMmaCore_3_01Shape 00_01WarpShape 00_01GemmShaaf312aafe9da92ea9d417bcc12a8e7dc.html) | | | Partial specialization: [More...](structcutlass_1_1gemm_1_1threadblock_1_1DefaultMmaCore_3_01Shape 00_01WarpShape 00_01GemmShaaf312aafe9da92ea9d417bcc12a8e7dc.html#details)
| | | | 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_ >](structcutlass_1_1gemm_1_1threadblock_1_1DefaultMmaCore_3_01Shape 00_01WarpShape 00_01GemmSha863d4139ccaa713bc4bde32c425f4067.html) | | | Partial specialization: [More...](structcutlass_1_1gemm_1_1threadblock_1_1DefaultMmaCore_3_01Shape 00_01WarpShape 00_01GemmSha863d4139ccaa713bc4bde32c425f4067.html#details)
| | | | 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_ >](structcutlass_1_1gemm_1_1threadblock_1_1DefaultMmaCore_3_01Shape 00_01WarpShape 00_01GemmSha2c0d0b7cdb5c4bcb11e83c058eb65345.html) | | | Partial specialization: [More...](structcutlass_1_1gemm_1_1threadblock_1_1DefaultMmaCore_3_01Shape 00_01WarpShape 00_01GemmSha2c0d0b7cdb5c4bcb11e83c058eb65345.html#details)
| | |

|

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.

---

Generated by 1.8.11