Back to Cutlass

CUTLASS: cutlass::gemm::threadblock Namespace Reference

docs/namespacecutlass_1_1gemm_1_1threadblock.html

4.4.215.5 KB
Original Source

| | CUTLASS

CUDA Templates for Linear Algebra Subroutines and Solvers |

Namespaces | Classes | Functions

cutlass::gemm::threadblock Namespace Reference

|

Namespaces

| | | detail | | |

|

Classes

| | struct | DefaultGemvCore | | | | struct | DefaultMma | | | | struct | DefaultMma< ElementA, LayoutA, kAlignmentA, ElementB, LayoutB, kAlignmentB, ElementAccumulator, layout::ColumnMajorInterleaved< InterleavedK >, OperatorClass, ArchTag, ThreadblockShape, WarpShape, InstructionShape, 2, Operator, true > | | | Specialization for column-major-interleaved output. More...
| | | | struct | DefaultMma< ElementA, LayoutA, kAlignmentA, ElementB, LayoutB, kAlignmentB, ElementAccumulator, layout::RowMajor, arch::OpClassSimt, ArchTag, ThreadblockShape, WarpShape, InstructionShape, 2, Operator, false > | | | Specialization for row-major output (OperatorClass Simt) More...
| | | | struct | DefaultMma< ElementA, LayoutA, kAlignmentA, ElementB, LayoutB, kAlignmentB, ElementAccumulator, layout::RowMajor, arch::OpClassTensorOp, ArchTag, ThreadblockShape, WarpShape, InstructionShape, 2, Operator, false > | | | Specialization for row-major output (OperatorClass Simt) More...
| | | | struct | DefaultMma< int8_t, LayoutA, kAlignmentA, int8_t, LayoutB, kAlignmentB, ElementAccumulator, layout::RowMajor, arch::OpClassSimt, ArchTag, ThreadblockShape, WarpShape, GemmShape< 1, 1, 4 >, 2, Operator, false > | | | | struct | DefaultMmaCore | | | | struct | [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 | [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 | [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_01GemmSha46446d1e3871e31d2e728f710d78c8c1.html) | | | | struct | [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 | [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 | [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)
| | | | struct | [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 | [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 | [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 | [DefaultMmaCore< Shape_, WarpShape_, GemmShape< 8, 8, 4 >, ElementA_, layout::ColumnMajor, ElementB_, layout::ColumnMajor, ElementC_, LayoutC_, arch::OpClassTensorOp, 2, Operator_ >](structcutlass_1_1gemm_1_1threadblock_1_1DefaultMmaCore_3_01Shape 00_01WarpShape 00_01GemmShaf03a122202ad10acdc96f280106d678b.html) | | | | struct | [DefaultMmaCore< Shape_, WarpShape_, GemmShape< 8, 8, 4 >, ElementA_, layout::ColumnMajor, ElementB_, layout::RowMajor, ElementC_, LayoutC_, arch::OpClassTensorOp, 2, Operator_ >](structcutlass_1_1gemm_1_1threadblock_1_1DefaultMmaCore_3_01Shape 00_01WarpShape 00_01GemmSha69bef08ea63dd930f99d9788105873dd.html) | | | | struct | [DefaultMmaCore< Shape_, WarpShape_, GemmShape< 8, 8, 4 >, ElementA_, layout::RowMajor, ElementB_, layout::ColumnMajor, ElementC_, LayoutC_, arch::OpClassTensorOp, 2, Operator_ >](structcutlass_1_1gemm_1_1threadblock_1_1DefaultMmaCore_3_01Shape 00_01WarpShape 00_01GemmSha3adf608332a8c9ee7014fced0da8a9ca.html) | | | | struct | [DefaultMmaCore< Shape_, WarpShape_, GemmShape< 8, 8, 4 >, ElementA_, layout::RowMajor, ElementB_, layout::RowMajor, ElementC_, LayoutC_, arch::OpClassTensorOp, 2, Operator_ >](structcutlass_1_1gemm_1_1threadblock_1_1DefaultMmaCore_3_01Shape 00_01WarpShape 00_01GemmShab7edfba3cdf43a07e3c4d719d87565a4.html) | | | | struct | [DefaultMmaCore< Shape_, WarpShape_, InstructionShape_, ElementA_, layout::ColumnMajor, ElementB_, layout::ColumnMajor, ElementC_, LayoutC_, arch::OpClassTensorOp, 2, Operator_ >](structcutlass_1_1gemm_1_1threadblock_1_1DefaultMmaCore_3_01Shape 00_01WarpShape 00_01Instruc803d38bc1e4618c07c47f54c87ae2678.html) | | | | struct | [DefaultMmaCore< Shape_, WarpShape_, InstructionShape_, ElementA_, layout::ColumnMajor, ElementB_, layout::RowMajor, ElementC_, LayoutC_, arch::OpClassTensorOp, 2, Operator_ >](structcutlass_1_1gemm_1_1threadblock_1_1DefaultMmaCore_3_01Shape 00_01WarpShape 00_01Instrucf60fe02fcdd80d28b7fd419133465dcc.html) | | | | struct | [DefaultMmaCore< Shape_, WarpShape_, InstructionShape_, ElementA_, layout::ColumnMajorInterleaved< InterleavedK >, ElementB_, layout::RowMajorInterleaved< InterleavedK >, ElementC_, LayoutC_, arch::OpClassTensorOp, 2, Operator_, AccumulatorsInRowMajor >](structcutlass_1_1gemm_1_1threadblock_1_1DefaultMmaCore_3_01Shape 00_01WarpShape 00_01Instruc2bf00737f4ad0a9da9a8be6d3e66c152.html) | | | | struct | [DefaultMmaCore< Shape_, WarpShape_, InstructionShape_, ElementA_, layout::RowMajor, ElementB_, layout::ColumnMajor, ElementC_, LayoutC_, arch::OpClassTensorOp, 2, Operator_ >](structcutlass_1_1gemm_1_1threadblock_1_1DefaultMmaCore_3_01Shape 00_01WarpShape 00_01Instruc24092ddc01fc83dabb7db4c14880fe60.html) | | | | struct | [DefaultMmaCore< Shape_, WarpShape_, InstructionShape_, ElementA_, layout::RowMajor, ElementB_, layout::RowMajor, ElementC_, LayoutC_, arch::OpClassTensorOp, 2, Operator_ >](structcutlass_1_1gemm_1_1threadblock_1_1DefaultMmaCore_3_01Shape 00_01WarpShape 00_01Instruc4fee9f2965b8468bfb42b94a74527d22.html) | | | | struct | GemmBatchedIdentityThreadblockSwizzle | | | Threadblock swizzling function for batched GEMMs. More...
| | | | struct | GemmHorizontalThreadblockSwizzle | | | Threadblock swizzling function for GEMMs. More...
| | | | struct | GemmIdentityThreadblockSwizzle | | | Threadblock swizzling function for GEMMs. More...
| | | | struct | GemmSplitKHorizontalThreadblockSwizzle | | | Threadblock swizzling function for split-K GEMMs. More...
| | | | struct | GemmSplitKIdentityThreadblockSwizzle | | | Threadblock swizzling function for split-K GEMMs. More...
| | | | class | Gemv | | | Structure to compute the matrix-vector product using SIMT math instructions. More...
| | | | struct | GemvBatchedStridedThreadblockDefaultSwizzle | | | Threadblock swizzling function for batched GEMVs. More...
| | | | class | MmaBase | | | | class | MmaPipelined | | | Structure to compute the matrix product targeting CUDA cores and SIMT math instructions. More...
| | | | struct | MmaPolicy | | | Policy object describing MmaTensorOp. More...
| | | | class | MmaSingleStage | | | Structure to compute the matrix product targeting CUDA cores and SIMT math instructions. More...
| | |

|

Functions

| | CUTLASS_DEVICE int | RematerializeThreadIdxX () | | | Helper to rematerialize block Idx. Reduces register liveness. More...
| | | | CUTLASS_DEVICE int | RematerializeThreadIdxY () | | | Helper to rematerialize block Idx. Reduces register liveness. More...
| | | | CUTLASS_DEVICE int | RematerializeThreadIdxZ () | | | Helper to rematerialize block Idx. Reduces register liveness. More...
| | | | CUTLASS_DEVICE int | RematerializeBlockIdxX () | | | Helper to rematerialize block Idx. Reduces register liveness. More...
| | | | CUTLASS_DEVICE int | RematerializeBlockIdxY () | | | Helper to rematerialize block Idx. Reduces register liveness. More...
| | | | CUTLASS_DEVICE int | RematerializeBlockIdxZ () | | | Helper to rematerialize block Idx. Reduces register liveness. More...
| | | | CUTLASS_DEVICE int | RematerializeBlockDimX () | | | Helper to rematerialize block Dim. Reduces register liveness. More...
| | | | CUTLASS_DEVICE int | RematerializeBlockDimY () | | | Helper to rematerialize block Dim. Reduces register liveness. More...
| | | | CUTLASS_DEVICE int | RematerializeBlockDimZ () | | | Helper to rematerialize block Dim. Reduces register liveness. More...
| | |

Function Documentation

| CUTLASS_DEVICE int cutlass::gemm::threadblock::RematerializeBlockDimX | ( | | ) | |

| CUTLASS_DEVICE int cutlass::gemm::threadblock::RematerializeBlockDimY | ( | | ) | |

| CUTLASS_DEVICE int cutlass::gemm::threadblock::RematerializeBlockDimZ | ( | | ) | |

| CUTLASS_DEVICE int cutlass::gemm::threadblock::RematerializeBlockIdxX | ( | | ) | |

| CUTLASS_DEVICE int cutlass::gemm::threadblock::RematerializeBlockIdxY | ( | | ) | |

| CUTLASS_DEVICE int cutlass::gemm::threadblock::RematerializeBlockIdxZ | ( | | ) | |

| CUTLASS_DEVICE int cutlass::gemm::threadblock::RematerializeThreadIdxX | ( | | ) | |

| CUTLASS_DEVICE int cutlass::gemm::threadblock::RematerializeThreadIdxY | ( | | ) | |

| CUTLASS_DEVICE int cutlass::gemm::threadblock::RematerializeThreadIdxZ | ( | | ) | |


Generated by 1.8.11