- k -

• kAccessesPerInterleavedTile : cutlass::epilogue::warp::VoltaTensorOpPolicy< WarpShape_, gemm::GemmShape< 32, 32, 4 >, float, layout::RowMajor > , cutlass::epilogue::warp::VoltaTensorOpPolicy< WarpShape_, gemm::GemmShape< 32, 32, 4 >, half_t, layout::RowMajor >
• kAccessesPerIteration : cutlass::epilogue::warp::SimtPolicy< WarpShape_, Operator_, layout::RowMajor, MmaSimtPolicy_ >
• kAccessesPerQuad : cutlass::epilogue::warp::TileIteratorVoltaTensorOp< WarpShape_, gemm::GemmShape< 32, 32, 4 >, float, layout::RowMajor >::Detail , cutlass::epilogue::warp::TileIteratorVoltaTensorOp< WarpShape_, gemm::GemmShape< 32, 32, 4 >, half_t, layout::RowMajor >::Detail
• kAccessesPerVector : cutlass::transform::threadblock::PredicatedTileAccessIterator< Shape_, Element_, layout::ColumnMajor, AdvanceRank, ThreadMap_, AccessType_ > , cutlass::transform::threadblock::PredicatedTileAccessIterator< Shape_, Element_, layout::ColumnMajorInterleaved< InterleavedK >, AdvanceRank, ThreadMap_, AccessType_ > , cutlass::transform::threadblock::PredicatedTileAccessIterator< Shape_, Element_, layout::PitchLinear, AdvanceRank, ThreadMap_, AccessType_ > , cutlass::transform::threadblock::PredicatedTileAccessIterator< Shape_, Element_, layout::RowMajor, AdvanceRank, ThreadMap_, AccessType_ > , cutlass::transform::threadblock::PredicatedTileAccessIterator< Shape_, Element_, layout::RowMajorInterleaved< InterleavedK >, AdvanceRank, ThreadMap_, AccessType_ > , cutlass::transform::threadblock::PredicatedTileIterator< Shape_, Element_, layout::PitchLinear, AdvanceRank, ThreadMap_, AccessSize >
• kAccessQuadDelta : cutlass::epilogue::warp::TileIteratorVoltaTensorOp< WarpShape_, gemm::GemmShape< 32, 32, 4 >, float, layout::RowMajor >::Detail , cutlass::epilogue::warp::TileIteratorVoltaTensorOp< WarpShape_, gemm::GemmShape< 32, 32, 4 >, half_t, layout::RowMajor >::Detail
• kAccessRows : cutlass::epilogue::threadblock::detail::RowArrangement< Shape, WarpsRemaining, ElementsPerAccess, ElementSize, false > , cutlass::epilogue::threadblock::detail::RowArrangement< Shape, WarpsRemaining, ElementsPerAccess, ElementSize, true > , cutlass::epilogue::threadblock::OutputTileOptimalThreadMap< Shape_, Count_, Threads, ElementsPerAccess, ElementSize >::Detail
• kAccessSize : cutlass::layout::ColumnMajorTensorOpMultiplicandCongruous< ElementSize, Crosswise > , cutlass::layout::ColumnMajorTensorOpMultiplicandCrosswise< ElementSize, Crosswise > , cutlass::layout::ColumnMajorVoltaTensorOpMultiplicandBCongruous< ElementSize > , cutlass::layout::ColumnMajorVoltaTensorOpMultiplicandCongruous< ElementSize > , cutlass::layout::ColumnMajorVoltaTensorOpMultiplicandCrosswise< ElementSize, KBlock > , cutlass::layout::RowMajorTensorOpMultiplicandCongruous< ElementSize, Crosswise > , cutlass::layout::RowMajorTensorOpMultiplicandCrosswise< ElementSize, Crosswise > , cutlass::layout::RowMajorVoltaTensorOpMultiplicandBCongruous< ElementSize > , cutlass::layout::RowMajorVoltaTensorOpMultiplicandCongruous< ElementSize > , cutlass::layout::RowMajorVoltaTensorOpMultiplicandCrosswise< ElementSize, KBlock > , cutlass::layout::TensorOpMultiplicand< ElementSize, Crosswise > , cutlass::layout::TensorOpMultiplicandColumnMajorInterleaved< ElementSize, InterleavedK > , cutlass::layout::TensorOpMultiplicandCongruous< ElementSize, Crosswise > , cutlass::layout::TensorOpMultiplicandCongruous< 32, Crosswise > , cutlass::layout::TensorOpMultiplicandCrosswise< ElementSize, Crosswise > , cutlass::layout::TensorOpMultiplicandRowMajorInterleaved< ElementSize, InterleavedK > , cutlass::layout::VoltaTensorOpMultiplicandBCongruous< ElementSize > , cutlass::layout::VoltaTensorOpMultiplicandCongruous< ElementSize > , cutlass::layout::VoltaTensorOpMultiplicandCrosswise< ElementSize, KBlock >
• kAccessSizeInBits : cutlass::epilogue::EpilogueWorkspace< Shape_, WarpCount, FragmentC_ > , cutlass::gemm::threadblock::DefaultMmaCore< Shape_, WarpShape_, GemmShape< 8, 8, 4 >, ElementA_, layout::ColumnMajor, ElementB_, layout::ColumnMajor, ElementC_, LayoutC_, arch::OpClassTensorOp, 2, Operator_ > , cutlass::gemm::threadblock::DefaultMmaCore< Shape_, WarpShape_, GemmShape< 8, 8, 4 >, ElementA_, layout::ColumnMajor, ElementB_, layout::RowMajor, ElementC_, LayoutC_, arch::OpClassTensorOp, 2, Operator_ > , cutlass::gemm::threadblock::DefaultMmaCore< Shape_, WarpShape_, GemmShape< 8, 8, 4 >, ElementA_, layout::RowMajor, ElementB_, layout::ColumnMajor, ElementC_, LayoutC_, arch::OpClassTensorOp, 2, Operator_ > , cutlass::gemm::threadblock::DefaultMmaCore< Shape_, WarpShape_, GemmShape< 8, 8, 4 >, ElementA_, layout::RowMajor, ElementB_, layout::RowMajor, ElementC_, LayoutC_, arch::OpClassTensorOp, 2, Operator_ > , cutlass::gemm::threadblock::DefaultMmaCore< Shape_, WarpShape_, InstructionShape_, ElementA_, layout::ColumnMajor, ElementB_, layout::ColumnMajor, ElementC_, LayoutC_, arch::OpClassTensorOp, 2, Operator_ > , cutlass::gemm::threadblock::DefaultMmaCore< Shape_, WarpShape_, InstructionShape_, ElementA_, layout::ColumnMajor, ElementB_, layout::RowMajor, ElementC_, LayoutC_, arch::OpClassTensorOp, 2, Operator_ > , cutlass::gemm::threadblock::DefaultMmaCore< Shape_, WarpShape_, InstructionShape_, ElementA_, layout::ColumnMajorInterleaved< InterleavedK >, ElementB_, layout::RowMajorInterleaved< InterleavedK >, ElementC_, LayoutC_, arch::OpClassTensorOp, 2, Operator_, AccumulatorsInRowMajor > , cutlass::gemm::threadblock::DefaultMmaCore< Shape_, WarpShape_, InstructionShape_, ElementA_, layout::RowMajor, ElementB_, layout::ColumnMajor, ElementC_, LayoutC_, arch::OpClassTensorOp, 2, Operator_ > , cutlass::gemm::threadblock::DefaultMmaCore< Shape_, WarpShape_, InstructionShape_, ElementA_, layout::RowMajor, ElementB_, layout::RowMajor, ElementC_, LayoutC_, arch::OpClassTensorOp, 2, Operator_ > , cutlass::transform::threadblock::RegularTileAccessIterator< Shape_, Element_, layout::TensorOpMultiplicandCongruous< sizeof_bits< Element_ >::value, int(128/sizeof(Element_))>, AdvanceRank, ThreadMap_, Alignment >::Detail , cutlass::transform::threadblock::RegularTileAccessIterator< Shape_, Element_, layout::TensorOpMultiplicandCrosswise< sizeof_bits< Element_ >::value, Crosswise >, AdvanceRank, ThreadMap_, Alignment >::Detail , cutlass::transform::threadblock::RegularTileIterator< Shape_, Element_, layout::TensorOpMultiplicandCongruous< sizeof_bits< Element_ >::value, int(128/sizeof(Element_))>, AdvanceRank, ThreadMap_, Alignment >::Detail , cutlass::transform::threadblock::RegularTileIterator< Shape_, Element_, layout::TensorOpMultiplicandCrosswise< sizeof_bits< Element_ >::value, Crosswise >, AdvanceRank, ThreadMap_, Alignment >::Detail , cutlass::transform::threadblock::RegularTileIterator< Shape_, Element_, layout::VoltaTensorOpMultiplicandBCongruous< sizeof_bits< Element_ >::value >, AdvanceRank, ThreadMap_, Alignment >::Detail , cutlass::transform::threadblock::RegularTileIterator< Shape_, Element_, layout::VoltaTensorOpMultiplicandCongruous< sizeof_bits< Element_ >::value >, AdvanceRank, ThreadMap_, Alignment >::Detail
• kAccessWidth : cutlass::epilogue::threadblock::detail::RowArrangement< Shape, WarpsRemaining, ElementsPerAccess, ElementSize, false > , cutlass::epilogue::threadblock::detail::RowArrangement< Shape, WarpsRemaining, ElementsPerAccess, ElementSize, true > , cutlass::epilogue::threadblock::OutputTileOptimalThreadMap< Shape_, Count_, Threads, ElementsPerAccess, ElementSize >::Detail
• kAccumulatorColumnStride : cutlass::epilogue::warp::TensorOpPolicy< WarpShape, OperatorShape, layout::RowMajor >
• kAccumulatorElementCount : cutlass::epilogue::warp::SimtPolicy< WarpShape_, Operator_, layout::RowMajor, MmaSimtPolicy_ >
• kAccumulatorRowStride : cutlass::epilogue::warp::TensorOpPolicy< WarpShape, OperatorShape, layout::RowMajor >
• kActiveRank : cutlass::reference::host::detail::TensorForEachHelper< Func, Rank, RankRemaining > , cutlass::reference::host::detail::TensorForEachHelper< Func, Rank, 0 >
• kAdvanceRank : cutlass::transform::threadblock::PredicatedTileAccessIterator2dThreadTile< Shape_, Element_, layout::ColumnMajor, AdvanceRank, ThreadMap_, AccessType_ > , cutlass::transform::threadblock::PredicatedTileAccessIterator2dThreadTile< Shape_, Element_, layout::PitchLinear, AdvanceRank, ThreadMap_, AccessType_ > , cutlass::transform::threadblock::PredicatedTileAccessIterator2dThreadTile< Shape_, Element_, layout::RowMajor, AdvanceRank, ThreadMap_, AccessType_ > , cutlass::transform::threadblock::PredicatedTileAccessIterator< Shape_, Element_, layout::ColumnMajor, AdvanceRank, ThreadMap_, AccessType_ > , cutlass::transform::threadblock::PredicatedTileAccessIterator< Shape_, Element_, layout::ColumnMajorInterleaved< InterleavedK >, AdvanceRank, ThreadMap_, AccessType_ > , cutlass::transform::threadblock::PredicatedTileAccessIterator< Shape_, Element_, layout::PitchLinear, AdvanceRank, ThreadMap_, AccessType_ > , cutlass::transform::threadblock::PredicatedTileAccessIterator< Shape_, Element_, layout::RowMajor, AdvanceRank, ThreadMap_, AccessType_ > , cutlass::transform::threadblock::PredicatedTileAccessIterator< Shape_, Element_, layout::RowMajorInterleaved< InterleavedK >, AdvanceRank, ThreadMap_, AccessType_ > , cutlass::transform::threadblock::PredicatedTileIterator2dThreadTile< Shape_, Element_, layout::ColumnMajor, AdvanceRank, ThreadMap_, Transpose_ > , cutlass::transform::threadblock::PredicatedTileIterator2dThreadTile< Shape_, Element_, layout::PitchLinear, AdvanceRank, ThreadMap_, Transpose_ > , cutlass::transform::threadblock::PredicatedTileIterator2dThreadTile< Shape_, Element_, layout::RowMajor, AdvanceRank, ThreadMap_, Transpose_ > , cutlass::transform::threadblock::PredicatedTileIterator< Shape_, Element_, layout::ColumnMajor, AdvanceRank, ThreadMap_, AccessSize > , cutlass::transform::threadblock::PredicatedTileIterator< Shape_, Element_, layout::ColumnMajorInterleaved< InterleavedK >, AdvanceRank, ThreadMap_, AccessSize > , cutlass::transform::threadblock::PredicatedTileIterator< Shape_, Element_, layout::PitchLinear, AdvanceRank, ThreadMap_, AccessSize > , cutlass::transform::threadblock::PredicatedTileIterator< Shape_, Element_, layout::RowMajor, AdvanceRank, ThreadMap_, AccessSize > , cutlass::transform::threadblock::PredicatedTileIterator< Shape_, Element_, layout::RowMajorInterleaved< InterleavedK >, AdvanceRank, ThreadMap_, AccessSize > , cutlass::transform::threadblock::RegularTileAccessIterator< Shape_, Element_, layout::ColumnMajor, AdvanceRank, ThreadMap_, Alignment > , cutlass::transform::threadblock::RegularTileAccessIterator< Shape_, Element_, layout::ColumnMajorTensorOpMultiplicandCongruous< sizeof_bits< Element_ >::value, int(128/sizeof(Element_))>, AdvanceRank, ThreadMap_, Alignment > , cutlass::transform::threadblock::RegularTileAccessIterator< Shape_, Element_, layout::ColumnMajorTensorOpMultiplicandCrosswise< sizeof_bits< Element_ >::value, Crosswise >, AdvanceRank, ThreadMap_, Alignment > , cutlass::transform::threadblock::RegularTileAccessIterator< Shape_, Element_, layout::PitchLinear, AdvanceRank, ThreadMap_, Alignment > , cutlass::transform::threadblock::RegularTileAccessIterator< Shape_, Element_, layout::RowMajor, AdvanceRank, ThreadMap_, Alignment > , cutlass::transform::threadblock::RegularTileAccessIterator< Shape_, Element_, layout::RowMajorTensorOpMultiplicandCongruous< sizeof_bits< Element_ >::value, int(128/sizeof(Element_))>, AdvanceRank, ThreadMap_, Alignment > , cutlass::transform::threadblock::RegularTileAccessIterator< Shape_, Element_, layout::RowMajorTensorOpMultiplicandCrosswise< sizeof_bits< Element_ >::value, Crosswise >, AdvanceRank, ThreadMap_, Alignment > , cutlass::transform::threadblock::RegularTileAccessIterator< Shape_, Element_, layout::TensorOpMultiplicandCongruous< sizeof_bits< Element_ >::value, int(128/sizeof(Element_))>, AdvanceRank, ThreadMap_, Alignment > , cutlass::transform::threadblock::RegularTileAccessIterator< Shape_, Element_, layout::TensorOpMultiplicandCrosswise< sizeof_bits< Element_ >::value, Crosswise >, AdvanceRank, ThreadMap_, Alignment > , cutlass::transform::threadblock::RegularTileIterator2dThreadTile< Shape_, Element_, layout::ColumnMajorInterleaved< 4 >, AdvanceRank, ThreadMap_, Alignment > , cutlass::transform::threadblock::RegularTileIterator2dThreadTile< Shape_, Element_, layout::PitchLinear, AdvanceRank, ThreadMap_, Alignment > , cutlass::transform::threadblock::RegularTileIterator2dThreadTile< Shape_, Element_, layout::RowMajorInterleaved< 4 >, AdvanceRank, ThreadMap_, Alignment > , cutlass::transform::threadblock::RegularTileIterator< Shape_, Element_, layout::ColumnMajor, AdvanceRank, ThreadMap_, Alignment > , cutlass::transform::threadblock::RegularTileIterator< Shape_, Element_, layout::ColumnMajorTensorOpMultiplicandCongruous< sizeof_bits< Element_ >::value, int(128/sizeof(Element_))>, AdvanceRank, ThreadMap_, Alignment > , cutlass::transform::threadblock::RegularTileIterator< Shape_, Element_, layout::ColumnMajorTensorOpMultiplicandCrosswise< sizeof_bits< Element_ >::value, Crosswise >, AdvanceRank, ThreadMap_, Alignment > , cutlass::transform::threadblock::RegularTileIterator< Shape_, Element_, layout::ColumnMajorVoltaTensorOpMultiplicandBCongruous< sizeof_bits< Element_ >::value >, AdvanceRank, ThreadMap_, Alignment > , cutlass::transform::threadblock::RegularTileIterator< Shape_, Element_, layout::ColumnMajorVoltaTensorOpMultiplicandCongruous< sizeof_bits< Element_ >::value >, AdvanceRank, ThreadMap_, Alignment > , cutlass::transform::threadblock::RegularTileIterator< Shape_, Element_, layout::ColumnMajorVoltaTensorOpMultiplicandCrosswise< sizeof_bits< Element_ >::value, Shape_::kRow >, AdvanceRank, ThreadMap_, Alignment > , cutlass::transform::threadblock::RegularTileIterator< Shape_, Element_, layout::PitchLinear, AdvanceRank, ThreadMap_, Alignment > , cutlass::transform::threadblock::RegularTileIterator< Shape_, Element_, layout::RowMajor, AdvanceRank, ThreadMap_, Alignment > , cutlass::transform::threadblock::RegularTileIterator< Shape_, Element_, layout::RowMajorTensorOpMultiplicandCongruous< sizeof_bits< Element_ >::value, int(128/sizeof(Element_))>, AdvanceRank, ThreadMap_, Alignment > , cutlass::transform::threadblock::RegularTileIterator< Shape_, Element_, layout::RowMajorTensorOpMultiplicandCrosswise< sizeof_bits< Element_ >::value, Crosswise >, AdvanceRank, ThreadMap_, Alignment > , cutlass::transform::threadblock::RegularTileIterator< Shape_, Element_, layout::RowMajorVoltaTensorOpMultiplicandBCongruous< sizeof_bits< Element_ >::value >, AdvanceRank, ThreadMap_, Alignment > , cutlass::transform::threadblock::RegularTileIterator< Shape_, Element_, layout::RowMajorVoltaTensorOpMultiplicandCongruous< sizeof_bits< Element_ >::value >, AdvanceRank, ThreadMap_, Alignment > , cutlass::transform::threadblock::RegularTileIterator< Shape_, Element_, layout::RowMajorVoltaTensorOpMultiplicandCrosswise< sizeof_bits< Element_ >::value, Shape_::kColumn >, AdvanceRank, ThreadMap_, Alignment > , cutlass::transform::threadblock::RegularTileIterator< Shape_, Element_, layout::TensorOpMultiplicandCongruous< sizeof_bits< Element_ >::value, int(128/sizeof(Element_))>, AdvanceRank, ThreadMap_, Alignment > , cutlass::transform::threadblock::RegularTileIterator< Shape_, Element_, layout::TensorOpMultiplicandCrosswise< sizeof_bits< Element_ >::value, Crosswise >, AdvanceRank, ThreadMap_, Alignment > , cutlass::transform::threadblock::RegularTileIterator< Shape_, Element_, layout::VoltaTensorOpMultiplicandBCongruous< sizeof_bits< Element_ >::value >, AdvanceRank, ThreadMap_, Alignment > , cutlass::transform::threadblock::RegularTileIterator< Shape_, Element_, layout::VoltaTensorOpMultiplicandCongruous< sizeof_bits< Element_ >::value >, AdvanceRank, ThreadMap_, Alignment > , cutlass::transform::threadblock::RegularTileIterator< Shape_, Element_, layout::VoltaTensorOpMultiplicandCrosswise< sizeof_bits< Element_ >::value, Shape_::kContiguous >, AdvanceRank, ThreadMap_, Alignment >
• kAlign : cutlass::AlignedBuffer< T, N, Align >
• kAlignment : cutlass::epilogue::threadblock::SharedLoadIterator< ThreadMap_, Element_, MaxAlignment > , cutlass::transform::threadblock::RegularTileAccessIterator< Shape_, Element_, layout::ColumnMajor, AdvanceRank, ThreadMap_, Alignment > , cutlass::transform::threadblock::RegularTileAccessIterator< Shape_, Element_, layout::ColumnMajorTensorOpMultiplicandCongruous< sizeof_bits< Element_ >::value, int(128/sizeof(Element_))>, AdvanceRank, ThreadMap_, Alignment > , cutlass::transform::threadblock::RegularTileAccessIterator< Shape_, Element_, layout::ColumnMajorTensorOpMultiplicandCrosswise< sizeof_bits< Element_ >::value, Crosswise >, AdvanceRank, ThreadMap_, Alignment > , cutlass::transform::threadblock::RegularTileAccessIterator< Shape_, Element_, layout::PitchLinear, AdvanceRank, ThreadMap_, Alignment > , cutlass::transform::threadblock::RegularTileAccessIterator< Shape_, Element_, layout::RowMajor, AdvanceRank, ThreadMap_, Alignment > , cutlass::transform::threadblock::RegularTileAccessIterator< Shape_, Element_, layout::RowMajorTensorOpMultiplicandCongruous< sizeof_bits< Element_ >::value, int(128/sizeof(Element_))>, AdvanceRank, ThreadMap_, Alignment > , cutlass::transform::threadblock::RegularTileAccessIterator< Shape_, Element_, layout::RowMajorTensorOpMultiplicandCrosswise< sizeof_bits< Element_ >::value, Crosswise >, AdvanceRank, ThreadMap_, Alignment > , cutlass::transform::threadblock::RegularTileAccessIterator< Shape_, Element_, layout::TensorOpMultiplicandCongruous< sizeof_bits< Element_ >::value, int(128/sizeof(Element_))>, AdvanceRank, ThreadMap_, Alignment > , cutlass::transform::threadblock::RegularTileAccessIterator< Shape_, Element_, layout::TensorOpMultiplicandCrosswise< sizeof_bits< Element_ >::value, Crosswise >, AdvanceRank, ThreadMap_, Alignment > , cutlass::transform::threadblock::RegularTileIterator2dThreadTile< Shape_, Element_, layout::ColumnMajorInterleaved< 4 >, AdvanceRank, ThreadMap_, Alignment > , cutlass::transform::threadblock::RegularTileIterator2dThreadTile< Shape_, Element_, layout::PitchLinear, AdvanceRank, ThreadMap_, Alignment > , cutlass::transform::threadblock::RegularTileIterator2dThreadTile< Shape_, Element_, layout::RowMajorInterleaved< 4 >, AdvanceRank, ThreadMap_, Alignment > , cutlass::transform::threadblock::RegularTileIterator< Shape_, Element_, layout::ColumnMajor, AdvanceRank, ThreadMap_, Alignment > , cutlass::transform::threadblock::RegularTileIterator< Shape_, Element_, layout::ColumnMajorTensorOpMultiplicandCongruous< sizeof_bits< Element_ >::value, int(128/sizeof(Element_))>, AdvanceRank, ThreadMap_, Alignment > , cutlass::transform::threadblock::RegularTileIterator< Shape_, Element_, layout::ColumnMajorTensorOpMultiplicandCrosswise< sizeof_bits< Element_ >::value, Crosswise >, AdvanceRank, ThreadMap_, Alignment > , cutlass::transform::threadblock::RegularTileIterator< Shape_, Element_, layout::PitchLinear, AdvanceRank, ThreadMap_, Alignment > , cutlass::transform::threadblock::RegularTileIterator< Shape_, Element_, layout::RowMajor, AdvanceRank, ThreadMap_, Alignment > , cutlass::transform::threadblock::RegularTileIterator< Shape_, Element_, layout::RowMajorTensorOpMultiplicandCongruous< sizeof_bits< Element_ >::value, int(128/sizeof(Element_))>, AdvanceRank, ThreadMap_, Alignment > , cutlass::transform::threadblock::RegularTileIterator< Shape_, Element_, layout::RowMajorTensorOpMultiplicandCrosswise< sizeof_bits< Element_ >::value, Crosswise >, AdvanceRank, ThreadMap_, Alignment > , cutlass::transform::threadblock::RegularTileIterator< Shape_, Element_, layout::RowMajorVoltaTensorOpMultiplicandCrosswise< sizeof_bits< Element_ >::value, Shape_::kColumn >, AdvanceRank, ThreadMap_, Alignment > , cutlass::transform::threadblock::RegularTileIterator< Shape_, Element_, layout::TensorOpMultiplicandCongruous< sizeof_bits< Element_ >::value, int(128/sizeof(Element_))>, AdvanceRank, ThreadMap_, Alignment > , cutlass::transform::threadblock::RegularTileIterator< Shape_, Element_, layout::TensorOpMultiplicandCrosswise< sizeof_bits< Element_ >::value, Crosswise >, AdvanceRank, ThreadMap_, Alignment >
• kAlignmentA : cutlass::gemm::device::DefaultGemmConfiguration< arch::OpClassSimt, ArchTag, ElementA, ElementB, ElementC, ElementAccumulator > , cutlass::gemm::device::DefaultGemmConfiguration< arch::OpClassSimt, ArchTag, int8_t, int8_t, ElementC, int32_t > , cutlass::gemm::device::DefaultGemmConfiguration< arch::OpClassTensorOp, arch::Sm70, ElementA, ElementB, ElementC, ElementAccumulator > , cutlass::gemm::device::DefaultGemmConfiguration< arch::OpClassTensorOp, arch::Sm75, ElementA, ElementB, ElementC, ElementAccumulator > , cutlass::gemm::device::DefaultGemmConfiguration< arch::OpClassTensorOp, arch::Sm75, int4b_t, int4b_t, ElementC, int32_t > , cutlass::gemm::device::DefaultGemmConfiguration< arch::OpClassTensorOp, arch::Sm75, int4b_t, uint4b_t, ElementC, int32_t > , cutlass::gemm::device::DefaultGemmConfiguration< arch::OpClassTensorOp, arch::Sm75, int8_t, int8_t, ElementC, int32_t > , cutlass::gemm::device::DefaultGemmConfiguration< arch::OpClassTensorOp, arch::Sm75, int8_t, uint8_t, ElementC, int32_t > , cutlass::gemm::device::DefaultGemmConfiguration< arch::OpClassTensorOp, arch::Sm75, uint4b_t, int4b_t, ElementC, int32_t > , cutlass::gemm::device::DefaultGemmConfiguration< arch::OpClassTensorOp, arch::Sm75, uint4b_t, uint4b_t, ElementC, int32_t > , cutlass::gemm::device::DefaultGemmConfiguration< arch::OpClassTensorOp, arch::Sm75, uint8_t, int8_t, ElementC, int32_t > , cutlass::gemm::device::DefaultGemmConfiguration< arch::OpClassTensorOp, arch::Sm75, uint8_t, uint8_t, ElementC, int32_t > , cutlass::gemm::device::DefaultGemmConfiguration< arch::OpClassWmmaTensorOp, ArchTag, ElementA, ElementB, ElementC, ElementAccumulator > , cutlass::gemm::device::Gemm< ElementA_, LayoutA_, ElementB_, LayoutB_, ElementC_, LayoutC_, ElementAccumulator_, OperatorClass_, ArchTag_, ThreadblockShape_, WarpShape_, InstructionShape_, EpilogueOutputOp_, ThreadblockSwizzle_, Stages, AlignmentA, AlignmentB, SplitKSerial, Operator_, IsBetaZero > , cutlass::gemm::device::Gemm< ElementA_, LayoutA_, ElementB_, LayoutB_, ElementC_, layout::ColumnMajor, ElementAccumulator_, OperatorClass_, ArchTag_, ThreadblockShape_, WarpShape_, InstructionShape_, EpilogueOutputOp_, ThreadblockSwizzle_, Stages, AlignmentA, AlignmentB, SplitKSerial, Operator_, IsBetaZero > , cutlass::gemm::device::GemmBatched< ElementA_, LayoutA_, ElementB_, LayoutB_, ElementC_, LayoutC_, ElementAccumulator_, OperatorClass_, ArchTag_, ThreadblockShape_, WarpShape_, InstructionShape_, EpilogueOutputOp_, ThreadblockSwizzle_, Stages, AlignmentA, AlignmentB, Operator_ > , cutlass::gemm::device::GemmBatched< ElementA_, LayoutA_, ElementB_, LayoutB_, ElementC_, layout::ColumnMajor, ElementAccumulator_, OperatorClass_, ArchTag_, ThreadblockShape_, WarpShape_, InstructionShape_, EpilogueOutputOp_, ThreadblockSwizzle_, Stages, AlignmentA, AlignmentB, Operator_ >
• kAlignmentB : cutlass::gemm::device::DefaultGemmConfiguration< arch::OpClassSimt, ArchTag, ElementA, ElementB, ElementC, ElementAccumulator > , cutlass::gemm::device::DefaultGemmConfiguration< arch::OpClassSimt, ArchTag, int8_t, int8_t, ElementC, int32_t > , cutlass::gemm::device::DefaultGemmConfiguration< arch::OpClassTensorOp, arch::Sm70, ElementA, ElementB, ElementC, ElementAccumulator > , cutlass::gemm::device::DefaultGemmConfiguration< arch::OpClassTensorOp, arch::Sm75, ElementA, ElementB, ElementC, ElementAccumulator > , cutlass::gemm::device::DefaultGemmConfiguration< arch::OpClassTensorOp, arch::Sm75, int4b_t, int4b_t, ElementC, int32_t > , cutlass::gemm::device::DefaultGemmConfiguration< arch::OpClassTensorOp, arch::Sm75, int4b_t, uint4b_t, ElementC, int32_t > , cutlass::gemm::device::DefaultGemmConfiguration< arch::OpClassTensorOp, arch::Sm75, int8_t, int8_t, ElementC, int32_t > , cutlass::gemm::device::DefaultGemmConfiguration< arch::OpClassTensorOp, arch::Sm75, int8_t, uint8_t, ElementC, int32_t > , cutlass::gemm::device::DefaultGemmConfiguration< arch::OpClassTensorOp, arch::Sm75, uint4b_t, int4b_t, ElementC, int32_t > , cutlass::gemm::device::DefaultGemmConfiguration< arch::OpClassTensorOp, arch::Sm75, uint4b_t, uint4b_t, ElementC, int32_t > , cutlass::gemm::device::DefaultGemmConfiguration< arch::OpClassTensorOp, arch::Sm75, uint8_t, int8_t, ElementC, int32_t > , cutlass::gemm::device::DefaultGemmConfiguration< arch::OpClassTensorOp, arch::Sm75, uint8_t, uint8_t, ElementC, int32_t > , cutlass::gemm::device::DefaultGemmConfiguration< arch::OpClassWmmaTensorOp, ArchTag, ElementA, ElementB, ElementC, ElementAccumulator > , cutlass::gemm::device::Gemm< ElementA_, LayoutA_, ElementB_, LayoutB_, ElementC_, LayoutC_, ElementAccumulator_, OperatorClass_, ArchTag_, ThreadblockShape_, WarpShape_, InstructionShape_, EpilogueOutputOp_, ThreadblockSwizzle_, Stages, AlignmentA, AlignmentB, SplitKSerial, Operator_, IsBetaZero > , cutlass::gemm::device::Gemm< ElementA_, LayoutA_, ElementB_, LayoutB_, ElementC_, layout::ColumnMajor, ElementAccumulator_, OperatorClass_, ArchTag_, ThreadblockShape_, WarpShape_, InstructionShape_, EpilogueOutputOp_, ThreadblockSwizzle_, Stages, AlignmentA, AlignmentB, SplitKSerial, Operator_, IsBetaZero > , cutlass::gemm::device::GemmBatched< ElementA_, LayoutA_, ElementB_, LayoutB_, ElementC_, LayoutC_, ElementAccumulator_, OperatorClass_, ArchTag_, ThreadblockShape_, WarpShape_, InstructionShape_, EpilogueOutputOp_, ThreadblockSwizzle_, Stages, AlignmentA, AlignmentB, Operator_ > , cutlass::gemm::device::GemmBatched< ElementA_, LayoutA_, ElementB_, LayoutB_, ElementC_, layout::ColumnMajor, ElementAccumulator_, OperatorClass_, ArchTag_, ThreadblockShape_, WarpShape_, InstructionShape_, EpilogueOutputOp_, ThreadblockSwizzle_, Stages, AlignmentA, AlignmentB, Operator_ >
• kAlignmentC : cutlass::gemm::device::Gemm< ElementA_, LayoutA_, ElementB_, LayoutB_, ElementC_, LayoutC_, ElementAccumulator_, OperatorClass_, ArchTag_, ThreadblockShape_, WarpShape_, InstructionShape_, EpilogueOutputOp_, ThreadblockSwizzle_, Stages, AlignmentA, AlignmentB, SplitKSerial, Operator_, IsBetaZero > , cutlass::gemm::device::Gemm< ElementA_, LayoutA_, ElementB_, LayoutB_, ElementC_, layout::ColumnMajor, ElementAccumulator_, OperatorClass_, ArchTag_, ThreadblockShape_, WarpShape_, InstructionShape_, EpilogueOutputOp_, ThreadblockSwizzle_, Stages, AlignmentA, AlignmentB, SplitKSerial, Operator_, IsBetaZero > , cutlass::gemm::device::GemmBatched< ElementA_, LayoutA_, ElementB_, LayoutB_, ElementC_, LayoutC_, ElementAccumulator_, OperatorClass_, ArchTag_, ThreadblockShape_, WarpShape_, InstructionShape_, EpilogueOutputOp_, ThreadblockSwizzle_, Stages, AlignmentA, AlignmentB, Operator_ > , cutlass::gemm::device::GemmBatched< ElementA_, LayoutA_, ElementB_, LayoutB_, ElementC_, layout::ColumnMajor, ElementAccumulator_, OperatorClass_, ArchTag_, ThreadblockShape_, WarpShape_, InstructionShape_, EpilogueOutputOp_, ThreadblockSwizzle_, Stages, AlignmentA, AlignmentB, Operator_ >
• kAlignmentK : cutlass::gemm::kernel::GemmSplitKParallel< Mma_, Epilogue_, ThreadblockSwizzle_ >
• kBatch : cutlass::gemm::BatchedGemmCoord
• kBits : cutlass::integer_subbyte< Bits, Signed >
• kBlockColumns : cutlass::layout::ColumnMajorBlockLinear< BlockRows, BlockColumns > , cutlass::layout::RowMajorBlockLinear< BlockRows, BlockColumns >
• kBlockRows : cutlass::layout::ColumnMajorBlockLinear< BlockRows, BlockColumns > , cutlass::layout::RowMajorBlockLinear< BlockRows, BlockColumns >
• kBytes : cutlass::AlignedBuffer< T, N, Align > , cutlass::PredicateVector< kPredicates_, kPredicatesPerByte_, kPredicateStart_ >
• kC : cutlass::Tensor4DCoord
• kCluster : cutlass::epilogue::threadblock::OutputTileShape< Column, Row, Group, Cluster, Tile >
• kColumn : cutlass::epilogue::threadblock::OutputTileShape< Column, Row, Group, Cluster, Tile > , cutlass::MatrixShape< Row_, Column_ >
• kColumns : cutlass::thread::Matrix< Element, Rows, Columns, Layout >
• kColumnsPerQuad : cutlass::epilogue::warp::TileIteratorVoltaTensorOp< WarpShape_, gemm::GemmShape< 32, 32, 4 >, float, layout::RowMajor >::Detail , cutlass::epilogue::warp::TileIteratorVoltaTensorOp< WarpShape_, gemm::GemmShape< 32, 32, 4 >, half_t, layout::RowMajor >::Detail
• kCompactedDeltaCluster : cutlass::epilogue::threadblock::OutputTileOptimalThreadMap< Shape_, Count_, Threads, ElementsPerAccess, ElementSize >::Detail
• kCompactedDeltaGroup : cutlass::epilogue::threadblock::OutputTileOptimalThreadMap< Shape_, Count_, Threads, ElementsPerAccess, ElementSize >::Detail
• kContiguous : cutlass::layout::PitchLinearShape< Contiguous, Strided >
• kContiguousElementsPerLine : cutlass::gemm::warp::MmaVoltaTensorOpMultiplicandTileIterator< Shape_, Operand_, Element_, cutlass::layout::VoltaTensorOpMultiplicandCrosswise< sizeof_bits< Element_ >::value, KBlock >, InstructionShape_, OpDelta_, 32 >::Policy , cutlass::transform::threadblock::RegularTileIterator< Shape_, Element_, layout::VoltaTensorOpMultiplicandCrosswise< sizeof_bits< Element_ >::value, Shape_::kContiguous >, AdvanceRank, ThreadMap_, Alignment >::Detail
• kCount : cutlass::AlignedBuffer< T, N, Align > , cutlass::epilogue::thread::Convert< ElementOutput_, Count, ElementAccumulator_, Round > , cutlass::epilogue::thread::LinearCombination< ElementOutput_, Count, ElementAccumulator_, ElementCompute_, Round > , cutlass::epilogue::thread::LinearCombinationClamp< ElementOutput_, Count, ElementAccumulator_, ElementCompute_, Round > , cutlass::epilogue::thread::LinearCombinationRelu< ElementOutput_, Count, ElementAccumulator_, ElementCompute_, Round > , cutlass::epilogue::thread::LinearCombinationRelu< ElementOutput_, Count, int, float, Round > , cutlass::epilogue::thread::ReductionOpPlus< Element_, Count > , cutlass::epilogue::threadblock::InterleavedPredicatedTileIterator< ThreadMap_, Element_, InterleavedK >::Mask , cutlass::epilogue::threadblock::OutputTileShape< Column, Row, Group, Cluster, Tile > , cutlass::epilogue::threadblock::PredicatedTileIterator< ThreadMap_, Element_ >::Mask , cutlass::gemm::GemmShape< M, N, K > , cutlass::layout::PitchLinearShape< Contiguous, Strided > , cutlass::MatrixShape< Row_, Column_ > , cutlass::reduction::thread::ReduceAdd< ElementAccumulator_, Element_, Count >
• kCrosswise : cutlass::gemm::warp::MmaTensorOpMultiplicandTileIterator< Shape_, Operand_, Element_, cutlass::layout::ColumnMajorTensorOpMultiplicandCrosswise< sizeof_bits< Element_ >::value, Crosswise >, InstructionShape_, OpDelta_, 32, PartitionsK_ > , cutlass::gemm::warp::MmaTensorOpMultiplicandTileIterator< Shape_, Operand_, Element_, cutlass::layout::RowMajorTensorOpMultiplicandCrosswise< sizeof_bits< Element_ >::value, Crosswise >, InstructionShape_, OpDelta_, 32, PartitionsK_ > , cutlass::gemm::warp::MmaTensorOpMultiplicandTileIterator< Shape_, Operand_, Element_, cutlass::layout::TensorOpMultiplicandCrosswise< sizeof_bits< Element_ >::value, Crosswise >, InstructionShape_, OpDelta_, 32, PartitionsK_ > , cutlass::layout::TensorOpMultiplicand< ElementSize, Crosswise > , cutlass::layout::TensorOpMultiplicandCrosswise< ElementSize, Crosswise > , cutlass::transform::threadblock::RegularTileAccessIterator< Shape_, Element_, layout::TensorOpMultiplicandCrosswise< sizeof_bits< Element_ >::value, Crosswise >, AdvanceRank, ThreadMap_, Alignment >
• kDeltaCluster : cutlass::epilogue::threadblock::OutputTileOptimalThreadMap< Shape_, Count_, Threads, ElementsPerAccess, ElementSize >::Detail
• kDeltaColumn : cutlass::epilogue::threadblock::detail::RowArrangement< Shape, WarpsRemaining, ElementsPerAccess, ElementSize, false > , cutlass::epilogue::threadblock::detail::RowArrangement< Shape, WarpsRemaining, ElementsPerAccess, ElementSize, true >
• kDeltaGroup : cutlass::epilogue::threadblock::OutputTileOptimalThreadMap< Shape_, Count_, Threads, ElementsPerAccess, ElementSize >::Detail
• kDeltaRow : cutlass::epilogue::threadblock::detail::RowArrangement< Shape, WarpsRemaining, ElementsPerAccess, ElementSize, false > , cutlass::epilogue::threadblock::detail::RowArrangement< Shape, WarpsRemaining, ElementsPerAccess, ElementSize, true >
• kElementCount : cutlass::transform::thread::Transpose< ElementCount_, layout::PitchLinearShape< 4, 4 >, int8_t >
• kElements : cutlass::Array< T, N, false > , cutlass::Array< T, N, true > , cutlass::transform::threadblock::PredicatedTileIterator2dThreadTile< Shape_, Element_, layout::PitchLinear, AdvanceRank, ThreadMap_, Transpose_ >::AccessType
• kElementSize : cutlass::epilogue::threadblock::detail::RowArrangement< Shape, WarpsRemaining, ElementsPerAccess, ElementSize, false > , cutlass::epilogue::threadblock::detail::RowArrangement< Shape, WarpsRemaining, ElementsPerAccess, ElementSize, true > , cutlass::epilogue::threadblock::InterleavedOutputTileThreadMap< WarpCount_, MmaCount_, Threads, ElementsPerAccess, ElementSize > , cutlass::epilogue::threadblock::OutputTileOptimalThreadMap< Shape_, Count_, Threads, ElementsPerAccess, ElementSize > , cutlass::layout::ColumnMajorTensorOpMultiplicandCongruous< ElementSize, Crosswise > , cutlass::layout::ColumnMajorTensorOpMultiplicandCrosswise< ElementSize, Crosswise > , cutlass::layout::ColumnMajorVoltaTensorOpMultiplicandBCongruous< ElementSize > , cutlass::layout::ColumnMajorVoltaTensorOpMultiplicandCongruous< ElementSize > , cutlass::layout::ColumnMajorVoltaTensorOpMultiplicandCrosswise< ElementSize, KBlock > , cutlass::layout::RowMajorTensorOpMultiplicandCongruous< ElementSize, Crosswise > , cutlass::layout::RowMajorTensorOpMultiplicandCrosswise< ElementSize, Crosswise > , cutlass::layout::RowMajorVoltaTensorOpMultiplicandBCongruous< ElementSize > , cutlass::layout::RowMajorVoltaTensorOpMultiplicandCongruous< ElementSize > , cutlass::layout::RowMajorVoltaTensorOpMultiplicandCrosswise< ElementSize, KBlock > , cutlass::layout::TensorOpMultiplicand< ElementSize, Crosswise > , cutlass::layout::TensorOpMultiplicandColumnMajorInterleaved< ElementSize, InterleavedK > , cutlass::layout::TensorOpMultiplicandCongruous< ElementSize, Crosswise > , cutlass::layout::TensorOpMultiplicandCongruous< 32, Crosswise > , cutlass::layout::TensorOpMultiplicandCrosswise< ElementSize, Crosswise > , cutlass::layout::TensorOpMultiplicandRowMajorInterleaved< ElementSize, InterleavedK > , cutlass::layout::VoltaTensorOpMultiplicandBCongruous< ElementSize > , cutlass::layout::VoltaTensorOpMultiplicandCongruous< ElementSize > , cutlass::layout::VoltaTensorOpMultiplicandCrosswise< ElementSize, KBlock >
• kElementsPerAccess : cutlass::epilogue::EpilogueWorkspace< Shape_, WarpCount, FragmentC_ > , cutlass::epilogue::threadblock::DefaultEpilogueComplexTensorOp< Shape_, WarpMmaTensorOp_, PartitionsK, OutputOp_, ElementsPerAccess > , cutlass::epilogue::threadblock::DefaultEpilogueSimt< Shape_, WarpMmaSimt_, OutputOp_, ElementsPerAccess > , cutlass::epilogue::threadblock::DefaultEpilogueTensorOp< Shape_, WarpMmaTensorOp_, PartitionsK, OutputOp_, ElementsPerAccess > , cutlass::epilogue::threadblock::DefaultEpilogueVoltaTensorOp< Shape_, WarpMmaTensorOp_, PartitionsK, OutputOp_, ElementsPerAccess > , cutlass::epilogue::threadblock::DefaultEpilogueWmmaTensorOp< Shape_, WarpMmaTensorOp_, PartitionsK, OutputOp_, ElementsPerAccess > , cutlass::epilogue::threadblock::DefaultInterleavedEpilogueTensorOp< Shape_, WarpMmaTensorOp_, PartitionsK, OutputOp_, ElementsPerAccess, InterleavedK, IsBetaZero, isSplitK > , cutlass::epilogue::threadblock::DefaultInterleavedThreadMapTensorOp< ThreadblockShape_, WarpShape_, PartitionsK, Element_, ElementsPerAccess, InterleavedK > , cutlass::epilogue::threadblock::DefaultThreadMapSimt< ThreadblockShape_, WarpShape_, MmaSimtPolicy_, PartitionsK, Element_, ElementsPerAccess > , cutlass::epilogue::threadblock::DefaultThreadMapTensorOp< ThreadblockShape_, WarpShape_, PartitionsK, Element_, ElementsPerAccess > , cutlass::epilogue::threadblock::DefaultThreadMapVoltaTensorOp< ThreadblockShape_, WarpShape_, PartitionsK, ElementOutput_, ElementsPerAccess, float > , cutlass::epilogue::threadblock::DefaultThreadMapVoltaTensorOp< ThreadblockShape_, WarpShape_, PartitionsK, ElementOutput_, ElementsPerAccess, half_t > , cutlass::epilogue::threadblock::DefaultThreadMapWmmaTensorOp< ThreadblockShape_, WarpShape_, InstructionShape_, PartitionsK, Element_, ElementsPerAccess > , cutlass::epilogue::threadblock::detail::RowArrangement< Shape, WarpsRemaining, ElementsPerAccess, ElementSize, false > , cutlass::epilogue::threadblock::detail::RowArrangement< Shape, WarpsRemaining, ElementsPerAccess, ElementSize, true > , cutlass::epilogue::threadblock::Epilogue< Shape_, WarpMmaOperator_, PartitionsK, OutputTileIterator_, AccumulatorFragmentIterator_, WarpTileIterator_, SharedLoadIterator_, OutputOp_, Padding_ > , cutlass::epilogue::threadblock::InterleavedEpilogue< Shape_, WarpMmaOperator_, PartitionsK, OutputTileIterator_, AccumulatorFragmentIterator_, OutputOp_, InterleavedK, IsBetaZero > , cutlass::epilogue::threadblock::InterleavedOutputTileThreadMap< WarpCount_, MmaCount_, Threads, ElementsPerAccess, ElementSize > , cutlass::epilogue::threadblock::InterleavedPredicatedTileIterator< ThreadMap_, Element_, InterleavedK > , cutlass::epilogue::threadblock::OutputTileOptimalThreadMap< Shape_, Count_, Threads, ElementsPerAccess, ElementSize >::CompactedThreadMap , cutlass::epilogue::threadblock::OutputTileOptimalThreadMap< Shape_, Count_, Threads, ElementsPerAccess, ElementSize > , cutlass::epilogue::threadblock::OutputTileThreadMap< ThreadMap_, Shape_, Iterations_, Delta_, Count_ > , cutlass::epilogue::threadblock::PredicatedTileIterator< ThreadMap_, Element_ > , cutlass::epilogue::threadblock::SharedLoadIterator< ThreadMap_, Element_, MaxAlignment > , cutlass::epilogue::warp::SimtPolicy< WarpShape_, Operator_, layout::RowMajor, MmaSimtPolicy_ > , cutlass::epilogue::warp::TensorOpPolicy< WarpShape, OperatorShape, layout::ColumnMajorInterleaved< InterleavedK > > , cutlass::epilogue::warp::TensorOpPolicy< WarpShape, OperatorShape, layout::RowMajor > , cutlass::epilogue::warp::TileIteratorVoltaTensorOp< WarpShape_, gemm::GemmShape< 32, 32, 4 >, float, layout::RowMajor > , cutlass::epilogue::warp::TileIteratorVoltaTensorOp< WarpShape_, gemm::GemmShape< 32, 32, 4 >, half_t, layout::RowMajor > , cutlass::epilogue::warp::VoltaTensorOpPolicy< WarpShape_, gemm::GemmShape< 32, 32, 4 >, float, layout::RowMajor > , cutlass::epilogue::warp::VoltaTensorOpPolicy< WarpShape_, gemm::GemmShape< 32, 32, 4 >, half_t, layout::RowMajor > , cutlass::gemm::threadblock::DefaultMmaCore< Shape_, WarpShape_, GemmShape< 1, 1, 1 >, ElementA_, layout::ColumnMajor, ElementB_, layout::ColumnMajor, ElementC_, LayoutC_, arch::OpClassSimt, 2, Operator_ > , cutlass::gemm::threadblock::DefaultMmaCore< Shape_, WarpShape_, GemmShape< 1, 1, 1 >, ElementA_, layout::ColumnMajor, ElementB_, layout::RowMajor, ElementC_, LayoutC_, arch::OpClassSimt, 2, Operator_ > , cutlass::gemm::threadblock::DefaultMmaCore< Shape_, WarpShape_, GemmShape< 1, 1, 1 >, ElementA_, layout::RowMajor, ElementB_, layout::ColumnMajor, ElementC_, LayoutC_, arch::OpClassSimt, 2, Operator_ > , cutlass::gemm::threadblock::DefaultMmaCore< Shape_, WarpShape_, GemmShape< 1, 1, 1 >, ElementA_, layout::RowMajor, ElementB_, layout::RowMajor, ElementC_, LayoutC_, arch::OpClassSimt, 2, Operator_ > , cutlass::gemm::threadblock::DefaultMmaCore< Shape_, WarpShape_, InstructionShape_, ElementA_, layout::ColumnMajorInterleaved< InterleavedK >, ElementB_, layout::RowMajorInterleaved< InterleavedK >, ElementC_, LayoutC_, arch::OpClassTensorOp, 2, Operator_, AccumulatorsInRowMajor > , cutlass::gemm::warp::MmaVoltaTensorOpMultiplicandTileIterator< Shape_, Operand_, Element_, cutlass::layout::VoltaTensorOpMultiplicandCrosswise< sizeof_bits< Element_ >::value, KBlock >, InstructionShape_, OpDelta_, 32 >::Policy , cutlass::layout::ColumnMajorTensorOpMultiplicandCongruous< ElementSize, Crosswise > , cutlass::layout::ColumnMajorTensorOpMultiplicandCrosswise< ElementSize, Crosswise > , cutlass::layout::ColumnMajorVoltaTensorOpMultiplicandBCongruous< ElementSize > , cutlass::layout::ColumnMajorVoltaTensorOpMultiplicandCongruous< ElementSize > , cutlass::layout::ColumnMajorVoltaTensorOpMultiplicandCrosswise< ElementSize, KBlock > , cutlass::layout::RowMajorTensorOpMultiplicandCongruous< ElementSize, Crosswise > , cutlass::layout::RowMajorTensorOpMultiplicandCrosswise< ElementSize, Crosswise > , cutlass::layout::RowMajorVoltaTensorOpMultiplicandBCongruous< ElementSize > , cutlass::layout::RowMajorVoltaTensorOpMultiplicandCongruous< ElementSize > , cutlass::layout::RowMajorVoltaTensorOpMultiplicandCrosswise< ElementSize, KBlock > , cutlass::layout::TensorOpMultiplicand< ElementSize, Crosswise > , cutlass::layout::TensorOpMultiplicandColumnMajorInterleaved< ElementSize, InterleavedK > , cutlass::layout::TensorOpMultiplicandCongruous< ElementSize, Crosswise > , cutlass::layout::TensorOpMultiplicandCongruous< 32, Crosswise > , cutlass::layout::TensorOpMultiplicandCrosswise< ElementSize, Crosswise > , cutlass::layout::TensorOpMultiplicandRowMajorInterleaved< ElementSize, InterleavedK > , cutlass::layout::VoltaTensorOpMultiplicandBCongruous< ElementSize > , cutlass::layout::VoltaTensorOpMultiplicandCongruous< ElementSize > , cutlass::layout::VoltaTensorOpMultiplicandCrosswise< ElementSize, KBlock > , cutlass::reduction::kernel::ReduceSplitK< Shape_, OutputOp_, ReductionOp_, PartitionsPerStage > , cutlass::transform::PitchLinear2DThreadTileStripminedThreadMap< Shape_, Threads, cutlass::layout::PitchLinearShape< 4, 4 > > , cutlass::transform::PitchLinearStripminedThreadMap< Shape_, Threads, ElementsPerAccess > , cutlass::transform::PitchLinearTilePolicyStripminedThreadContiguous< Shape, Threads, ElementsPerAccess > , cutlass::transform::PitchLinearTilePolicyStripminedThreadStrided< Shape, Threads, ElementsPerAccess > , cutlass::transform::PitchLinearWarpRakedThreadMap< Shape_, Threads, WarpThreadArrangement_, ElementsPerAccess > , cutlass::transform::PitchLinearWarpStripedThreadMap< Shape_, Threads, WarpThreadArrangement_, ElementsPerAccess > , cutlass::transform::TransposePitchLinearThreadMap2DThreadTile< ThreadMap_ > , cutlass::transform::TransposePitchLinearThreadMap< ThreadMap_, WarpThreadArrangement_ > , cutlass::transform::TransposePitchLinearThreadMapSimt< ThreadMap_ >
• kElementsPerIteration : cutlass::epilogue::warp::SimtPolicy< WarpShape_, Operator_, layout::RowMajor, MmaSimtPolicy_ >
• kElementsPerMma : cutlass::epilogue::warp::VoltaTensorOpPolicy< WarpShape_, gemm::GemmShape< 32, 32, 4 >, float, layout::RowMajor > , cutlass::epilogue::warp::VoltaTensorOpPolicy< WarpShape_, gemm::GemmShape< 32, 32, 4 >, half_t, layout::RowMajor >
• kElementsPerStoredItem : cutlass::Array< T, N, false > , cutlass::HostTensor< Element_, Layout_ >
• kEpilogueElementsPerAccess : cutlass::gemm::kernel::DefaultGemm< ElementA, LayoutA, kAlignmentA, ElementB, LayoutB, kAlignmentB, ElementC, layout::RowMajor, ElementAccumulator, arch::OpClassSimt, ArchTag, ThreadblockShape, WarpShape, GemmShape< 1, 1, 1 >, EpilogueOutputOp, ThreadblockSwizzle, 2, SplitKSerial, Operator > , cutlass::gemm::kernel::DefaultGemm< int8_t, LayoutA, kAlignmentA, int8_t, LayoutB, kAlignmentB, ElementC, LayoutC, ElementAccumulator, arch::OpClassSimt, ArchTag, ThreadblockShape, WarpShape, GemmShape< 1, 1, 4 >, EpilogueOutputOp, ThreadblockSwizzle, 2, SplitKSerial, Operator, false >
• keys : cutlass::CommandLine
• kFactor : cutlass::layout::TensorOpMultiplicand< ElementSize, Crosswise > , cutlass::layout::TensorOpMultiplicandCrosswise< ElementSize, Crosswise >
• kGroup : cutlass::epilogue::threadblock::OutputTileShape< Column, Row, Group, Cluster, Tile >
• kGroupCount : cutlass::epilogue::threadblock::DefaultThreadMapSimt< ThreadblockShape_, WarpShape_, MmaSimtPolicy_, PartitionsK, Element_, ElementsPerAccess >::Detail
• kGroupPerTile : cutlass::gemm::warp::MmaSimtTileIterator< Shape_, Operand::kA, Element_, layout::ColumnMajorInterleaved< 4 >, Policy_, PartitionsK, PartitionGroupSize > , cutlass::gemm::warp::MmaSimtTileIterator< Shape_, Operand::kB, Element_, layout::RowMajorInterleaved< 4 >, Policy_, PartitionsK, PartitionGroupSize >
• kGroupsPerTile : cutlass::gemm::warp::MmaTensorOpMultiplicandTileIterator< Shape_, Operand_, Element_, cutlass::layout::TensorOpMultiplicandCongruous< sizeof_bits< Element_ >::value, 64 >, InstructionShape_, OpDelta_, 32, PartitionsK_ >::Policy , cutlass::gemm::warp::MmaTensorOpMultiplicandTileIterator< Shape_, Operand_, Element_, cutlass::layout::TensorOpMultiplicandCrosswise< sizeof_bits< Element_ >::value, Crosswise >, InstructionShape_, OpDelta_, 32, PartitionsK_ >::Policy
• kH : cutlass::Tensor4DCoord
• kImaginaryIndex : cutlass::epilogue::warp::FragmentIteratorComplexTensorOp< WarpShape_, OperatorShape_, OperatorElementC_, OperatorFragmentC_, layout::RowMajor >
• kind : cutlass::Distribution , cutlass::library::OperationDescription
• kInterleave : cutlass::gemm::warp::MmaSimtTileIterator< Shape_, Operand::kA, Element_, layout::ColumnMajorInterleaved< 4 >, Policy_, PartitionsK, PartitionGroupSize > , cutlass::gemm::warp::MmaSimtTileIterator< Shape_, Operand::kB, Element_, layout::RowMajorInterleaved< 4 >, Policy_, PartitionsK, PartitionGroupSize > , cutlass::layout::ColumnMajorInterleaved< Interleave > , cutlass::layout::RowMajorInterleaved< Interleave > , cutlass::layout::TensorCxRSKx< Interleave > , cutlass::layout::TensorNCxHWx< Interleave >
• kInterleavedK : cutlass::epilogue::threadblock::DefaultInterleavedThreadMapTensorOp< ThreadblockShape_, WarpShape_, PartitionsK, Element_, ElementsPerAccess, InterleavedK > , cutlass::epilogue::warp::FragmentIteratorTensorOp< WarpShape_, OperatorShape_, OperatorElementC_, OperatorFragmentC_, layout::ColumnMajorInterleaved< InterleavedK > > , cutlass::gemm::threadblock::DefaultMmaCore< Shape_, WarpShape_, InstructionShape_, ElementA_, layout::ColumnMajorInterleaved< InterleavedK >, ElementB_, layout::RowMajorInterleaved< InterleavedK >, ElementC_, LayoutC_, arch::OpClassTensorOp, 2, Operator_, AccumulatorsInRowMajor > , cutlass::layout::TensorOpMultiplicandColumnMajorInterleaved< ElementSize, InterleavedK > , cutlass::layout::TensorOpMultiplicandRowMajorInterleaved< ElementSize, InterleavedK > , cutlass::transform::threadblock::PredicatedTileAccessIterator< Shape_, Element_, layout::ColumnMajorInterleaved< InterleavedK >, AdvanceRank, ThreadMap_, AccessType_ > , cutlass::transform::threadblock::PredicatedTileAccessIterator< Shape_, Element_, layout::RowMajorInterleaved< InterleavedK >, AdvanceRank, ThreadMap_, AccessType_ > , cutlass::transform::threadblock::PredicatedTileIterator< Shape_, Element_, layout::ColumnMajorInterleaved< InterleavedK >, AdvanceRank, ThreadMap_, AccessSize > , cutlass::transform::threadblock::PredicatedTileIterator< Shape_, Element_, layout::RowMajorInterleaved< InterleavedK >, AdvanceRank, ThreadMap_, AccessSize >
• kInterleavedTilesM : cutlass::epilogue::threadblock::DefaultThreadMapVoltaTensorOp< ThreadblockShape_, WarpShape_, PartitionsK, ElementOutput_, ElementsPerAccess, float >::Detail , cutlass::epilogue::threadblock::DefaultThreadMapVoltaTensorOp< ThreadblockShape_, WarpShape_, PartitionsK, ElementOutput_, ElementsPerAccess, half_t >::Detail
• kIsBetaZero : cutlass::gemm::device::Gemm< ElementA_, LayoutA_, ElementB_, LayoutB_, ElementC_, LayoutC_, ElementAccumulator_, OperatorClass_, ArchTag_, ThreadblockShape_, WarpShape_, InstructionShape_, EpilogueOutputOp_, ThreadblockSwizzle_, Stages, AlignmentA, AlignmentB, SplitKSerial, Operator_, IsBetaZero > , cutlass::gemm::device::Gemm< ElementA_, LayoutA_, ElementB_, LayoutB_, ElementC_, layout::ColumnMajor, ElementAccumulator_, OperatorClass_, ArchTag_, ThreadblockShape_, WarpShape_, InstructionShape_, EpilogueOutputOp_, ThreadblockSwizzle_, Stages, AlignmentA, AlignmentB, SplitKSerial, Operator_, IsBetaZero >
• kIsConventionalLayout : cutlass::gemm::thread::detail::EnableMma_Crow_SM60< LayoutA, LayoutB >
• kIterarionsPerAccess : cutlass::transform::threadblock::RegularTileIterator< Shape_, Element_, layout::VoltaTensorOpMultiplicandCrosswise< sizeof_bits< Element_ >::value, Shape_::kContiguous >, AdvanceRank, ThreadMap_, Alignment >::Detail
• kIterations : cutlass::epilogue::EpilogueWorkspace< Shape_, WarpCount, FragmentC_ > , cutlass::epilogue::threadblock::DefaultThreadMapSimt< ThreadblockShape_, WarpShape_, MmaSimtPolicy_, PartitionsK, Element_, ElementsPerAccess >::Detail , cutlass::epilogue::threadblock::InterleavedPredicatedTileIterator< ThreadMap_, Element_, InterleavedK > , cutlass::epilogue::threadblock::PredicatedTileIterator< ThreadMap_, Element_ > , cutlass::epilogue::warp::FragmentIteratorComplexTensorOp< WarpShape_, OperatorShape_, OperatorElementC_, OperatorFragmentC_, layout::RowMajor > , cutlass::epilogue::warp::FragmentIteratorSimt< WarpShape_, Operator_, layout::RowMajor, MmaSimtPolicy_ > , cutlass::epilogue::warp::FragmentIteratorTensorOp< WarpShape_, OperatorShape_, OperatorElementC_, OperatorFragmentC_, layout::ColumnMajorInterleaved< InterleavedK > > , cutlass::epilogue::warp::FragmentIteratorTensorOp< WarpShape_, OperatorShape_, OperatorElementC_, OperatorFragmentC_, layout::RowMajor > , cutlass::epilogue::warp::FragmentIteratorVoltaTensorOp< WarpShape_, gemm::GemmShape< 32, 32, 4 >, float, layout::RowMajor > , cutlass::epilogue::warp::FragmentIteratorVoltaTensorOp< WarpShape_, gemm::GemmShape< 32, 32, 4 >, half_t, layout::RowMajor > , cutlass::epilogue::warp::SimtPolicy< WarpShape_, Operator_, layout::RowMajor, MmaSimtPolicy_ > , cutlass::epilogue::warp::TensorOpPolicy< WarpShape, OperatorShape, layout::ColumnMajorInterleaved< InterleavedK > > , cutlass::epilogue::warp::TensorOpPolicy< WarpShape, OperatorShape, layout::RowMajor > , cutlass::epilogue::warp::TileIteratorSimt< WarpShape_, Operator_, Element_, layout::RowMajor, MmaSimtPolicy_ > , cutlass::epilogue::warp::TileIteratorTensorOp< WarpShape_, OperatorShape_, Element_, layout::RowMajor > , cutlass::epilogue::warp::TileIteratorVoltaTensorOp< WarpShape_, gemm::GemmShape< 32, 32, 4 >, float, layout::RowMajor > , cutlass::epilogue::warp::TileIteratorVoltaTensorOp< WarpShape_, gemm::GemmShape< 32, 32, 4 >, half_t, layout::RowMajor > , cutlass::epilogue::warp::VoltaTensorOpPolicy< WarpShape_, gemm::GemmShape< 32, 32, 4 >, float, layout::RowMajor > , cutlass::epilogue::warp::VoltaTensorOpPolicy< WarpShape_, gemm::GemmShape< 32, 32, 4 >, half_t, layout::RowMajor >
• kIterationsCluster : cutlass::epilogue::threadblock::OutputTileOptimalThreadMap< Shape_, Count_, Threads, ElementsPerAccess, ElementSize >::Detail
• kIterationsColumn : cutlass::epilogue::threadblock::detail::RowArrangement< Shape, WarpsRemaining, ElementsPerAccess, ElementSize, false > , cutlass::epilogue::threadblock::detail::RowArrangement< Shape, WarpsRemaining, ElementsPerAccess, ElementSize, true >
• kIterationsGroup : cutlass::epilogue::threadblock::OutputTileOptimalThreadMap< Shape_, Count_, Threads, ElementsPerAccess, ElementSize >::Detail
• kIterationsPerInstruction : cutlass::epilogue::warp::TensorOpPolicy< WarpShape, OperatorShape, layout::ColumnMajorInterleaved< InterleavedK > > , cutlass::epilogue::warp::TensorOpPolicy< WarpShape, OperatorShape, layout::RowMajor >
• kIterationsRow : cutlass::epilogue::threadblock::detail::RowArrangement< Shape, WarpsRemaining, ElementsPerAccess, ElementSize, false > , cutlass::epilogue::threadblock::detail::RowArrangement< Shape, WarpsRemaining, ElementsPerAccess, ElementSize, true >
• kK : cutlass::gemm::BatchedGemmCoord , cutlass::gemm::GemmCoord , cutlass::gemm::GemmShape< M, N, K >
• kKBlock : cutlass::gemm::warp::MmaVoltaTensorOpMultiplicandTileIterator< Shape_, Operand_, Element_, cutlass::layout::ColumnMajorVoltaTensorOpMultiplicandCrosswise< sizeof_bits< Element_ >::value, KBlock >, InstructionShape_, OpDelta_, 32 > , cutlass::gemm::warp::MmaVoltaTensorOpMultiplicandTileIterator< Shape_, Operand_, Element_, cutlass::layout::RowMajorVoltaTensorOpMultiplicandCrosswise< sizeof_bits< Element_ >::value, KBlock >, InstructionShape_, OpDelta_, 32 > , cutlass::gemm::warp::MmaVoltaTensorOpMultiplicandTileIterator< Shape_, Operand_, Element_, cutlass::layout::VoltaTensorOpMultiplicandCrosswise< sizeof_bits< Element_ >::value, KBlock >, InstructionShape_, OpDelta_, 32 > , cutlass::layout::VoltaTensorOpMultiplicandCrosswise< ElementSize, KBlock >
• kKN : cutlass::gemm::GemmShape< M, N, K >
• kLanesInQuad : cutlass::epilogue::warp::TileIteratorTensorOp< WarpShape_, OperatorShape_, Element_, layout::RowMajor >::Detail , cutlass::epilogue::warp::TileIteratorVoltaTensorOp< WarpShape_, gemm::GemmShape< 32, 32, 4 >, float, layout::RowMajor >::Detail , cutlass::epilogue::warp::TileIteratorVoltaTensorOp< WarpShape_, gemm::GemmShape< 32, 32, 4 >, half_t, layout::RowMajor >::Detail
• kLdsmOpInner : cutlass::gemm::warp::MmaTensorOpMultiplicandTileIterator< Shape_, Operand_, Element_, cutlass::layout::TensorOpMultiplicandCongruous< sizeof_bits< Element_ >::value, 64 >, InstructionShape_, OpDelta_, 32, PartitionsK_ >::Policy , cutlass::gemm::warp::MmaTensorOpMultiplicandTileIterator< Shape_, Operand_, Element_, cutlass::layout::TensorOpMultiplicandCrosswise< sizeof_bits< Element_ >::value, Crosswise >, InstructionShape_, OpDelta_, 32, PartitionsK_ >::Policy
• kLdsmOpOuter : cutlass::gemm::warp::MmaTensorOpMultiplicandTileIterator< Shape_, Operand_, Element_, cutlass::layout::TensorOpMultiplicandCongruous< sizeof_bits< Element_ >::value, 64 >, InstructionShape_, OpDelta_, 32, PartitionsK_ >::Policy , cutlass::gemm::warp::MmaTensorOpMultiplicandTileIterator< Shape_, Operand_, Element_, cutlass::layout::TensorOpMultiplicandCrosswise< sizeof_bits< Element_ >::value, Crosswise >, InstructionShape_, OpDelta_, 32, PartitionsK_ >::Policy
• kM : cutlass::gemm::BatchedGemmCoord , cutlass::gemm::GemmCoord , cutlass::gemm::GemmShape< M, N, K >
• kMask : cutlass::Array< T, N, false > , cutlass::integer_subbyte< Bits, Signed >
• kMemoryAccessSize : cutlass::epilogue::threadblock::detail::RowArrangement< Shape, WarpsRemaining, ElementsPerAccess, ElementSize, true >
• kMinAlignment : cutlass::epilogue::threadblock::SharedLoadIterator< ThreadMap_, Element_, MaxAlignment >
• kMinComputeCapability : cutlass::arch::Sm50 , cutlass::arch::Sm60 , cutlass::arch::Sm61 , cutlass::arch::Sm70 , cutlass::arch::Sm72 , cutlass::arch::Sm75
• kMK : cutlass::gemm::GemmShape< M, N, K >
• kMN : cutlass::gemm::GemmShape< M, N, K >
• kMNK : cutlass::gemm::GemmShape< M, N, K >
• kN : cutlass::gemm::BatchedGemmCoord , cutlass::gemm::GemmCoord , cutlass::gemm::GemmShape< M, N, K > , cutlass::Tensor4DCoord
• kOpDelta : cutlass::gemm::warp::MmaTensorOpMultiplicandTileIterator< Shape_, Operand_, Element_, cutlass::layout::ColumnMajorTensorOpMultiplicandCongruous< sizeof_bits< Element_ >::value, int(128/sizeof(Element_))>, InstructionShape_, OpDelta_, 32, PartitionsK_ > , cutlass::gemm::warp::MmaTensorOpMultiplicandTileIterator< Shape_, Operand_, Element_, cutlass::layout::ColumnMajorTensorOpMultiplicandCrosswise< sizeof_bits< Element_ >::value, Crosswise >, InstructionShape_, OpDelta_, 32, PartitionsK_ > , cutlass::gemm::warp::MmaTensorOpMultiplicandTileIterator< Shape_, Operand_, Element_, cutlass::layout::RowMajorTensorOpMultiplicandCongruous< sizeof_bits< Element_ >::value, int(128/sizeof(Element_))>, InstructionShape_, OpDelta_, 32, PartitionsK_ > , cutlass::gemm::warp::MmaTensorOpMultiplicandTileIterator< Shape_, Operand_, Element_, cutlass::layout::RowMajorTensorOpMultiplicandCrosswise< sizeof_bits< Element_ >::value, Crosswise >, InstructionShape_, OpDelta_, 32, PartitionsK_ > , cutlass::gemm::warp::MmaTensorOpMultiplicandTileIterator< Shape_, Operand_, Element_, cutlass::layout::TensorOpMultiplicandCongruous< sizeof_bits< Element_ >::value, 64 >, InstructionShape_, OpDelta_, 32, PartitionsK_ > , cutlass::gemm::warp::MmaTensorOpMultiplicandTileIterator< Shape_, Operand_, Element_, cutlass::layout::TensorOpMultiplicandCrosswise< sizeof_bits< Element_ >::value, Crosswise >, InstructionShape_, OpDelta_, 32, PartitionsK_ > , cutlass::gemm::warp::MmaVoltaTensorOpMultiplicandTileIterator< Shape_, Operand::kA, Element_, cutlass::layout::ColumnMajorVoltaTensorOpMultiplicandCongruous< sizeof_bits< Element_ >::value >, InstructionShape_, OpDelta_, 32 > , cutlass::gemm::warp::MmaVoltaTensorOpMultiplicandTileIterator< Shape_, Operand::kA, Element_, cutlass::layout::VoltaTensorOpMultiplicandCongruous< sizeof_bits< Element_ >::value >, InstructionShape_, OpDelta_, 32 > , cutlass::gemm::warp::MmaVoltaTensorOpMultiplicandTileIterator< Shape_, Operand::kB, Element_, cutlass::layout::RowMajorVoltaTensorOpMultiplicandBCongruous< sizeof_bits< Element_ >::value >, InstructionShape_, OpDelta_, 32 > , cutlass::gemm::warp::MmaVoltaTensorOpMultiplicandTileIterator< Shape_, Operand::kB, Element_, cutlass::layout::VoltaTensorOpMultiplicandBCongruous< sizeof_bits< Element_ >::value >, InstructionShape_, OpDelta_, 32 > , cutlass::gemm::warp::MmaVoltaTensorOpMultiplicandTileIterator< Shape_, Operand_, Element_, cutlass::layout::ColumnMajorVoltaTensorOpMultiplicandCrosswise< sizeof_bits< Element_ >::value, KBlock >, InstructionShape_, OpDelta_, 32 > , cutlass::gemm::warp::MmaVoltaTensorOpMultiplicandTileIterator< Shape_, Operand_, Element_, cutlass::layout::RowMajorVoltaTensorOpMultiplicandCrosswise< sizeof_bits< Element_ >::value, KBlock >, InstructionShape_, OpDelta_, 32 > , cutlass::gemm::warp::MmaVoltaTensorOpMultiplicandTileIterator< Shape_, Operand_, Element_, cutlass::layout::VoltaTensorOpMultiplicandCrosswise< sizeof_bits< Element_ >::value, KBlock >, InstructionShape_, OpDelta_, 32 >
• kOperand : cutlass::gemm::warp::MmaSimtTileIterator< Shape_, Operand::kA, Element_, layout::ColumnMajor, Policy_, PartitionsK, PartitionGroupSize > , cutlass::gemm::warp::MmaSimtTileIterator< Shape_, Operand::kA, Element_, layout::ColumnMajorInterleaved< 4 >, Policy_, PartitionsK, PartitionGroupSize > , cutlass::gemm::warp::MmaSimtTileIterator< Shape_, Operand::kB, Element_, layout::RowMajor, Policy_, PartitionsK, PartitionGroupSize > , cutlass::gemm::warp::MmaSimtTileIterator< Shape_, Operand::kB, Element_, layout::RowMajorInterleaved< 4 >, Policy_, PartitionsK, PartitionGroupSize > , cutlass::gemm::warp::MmaSimtTileIterator< Shape_, Operand::kC, Element_, layout::ColumnMajor, Policy_ > , cutlass::gemm::warp::MmaSimtTileIterator< Shape_, Operand::kC, Element_, layout::RowMajor, Policy_ > , cutlass::gemm::warp::MmaTensorOpAccumulatorTileIterator< Shape_, Element_, cutlass::layout::ColumnMajor, InstructionShape_, OpDelta_ > , cutlass::gemm::warp::MmaTensorOpAccumulatorTileIterator< Shape_, Element_, cutlass::layout::ColumnMajorInterleaved< InterleavedN >, InstructionShape_, OpDelta_ > , cutlass::gemm::warp::MmaTensorOpAccumulatorTileIterator< Shape_, Element_, cutlass::layout::RowMajor, InstructionShape_, OpDelta_ > , cutlass::gemm::warp::MmaTensorOpMultiplicandTileIterator< Shape_, Operand_, Element_, cutlass::layout::ColumnMajorTensorOpMultiplicandCongruous< sizeof_bits< Element_ >::value, int(128/sizeof(Element_))>, InstructionShape_, OpDelta_, 32, PartitionsK_ > , cutlass::gemm::warp::MmaTensorOpMultiplicandTileIterator< Shape_, Operand_, Element_, cutlass::layout::ColumnMajorTensorOpMultiplicandCrosswise< sizeof_bits< Element_ >::value, Crosswise >, InstructionShape_, OpDelta_, 32, PartitionsK_ > , cutlass::gemm::warp::MmaTensorOpMultiplicandTileIterator< Shape_, Operand_, Element_, cutlass::layout::RowMajorTensorOpMultiplicandCongruous< sizeof_bits< Element_ >::value, int(128/sizeof(Element_))>, InstructionShape_, OpDelta_, 32, PartitionsK_ > , cutlass::gemm::warp::MmaTensorOpMultiplicandTileIterator< Shape_, Operand_, Element_, cutlass::layout::RowMajorTensorOpMultiplicandCrosswise< sizeof_bits< Element_ >::value, Crosswise >, InstructionShape_, OpDelta_, 32, PartitionsK_ > , cutlass::gemm::warp::MmaTensorOpMultiplicandTileIterator< Shape_, Operand_, Element_, cutlass::layout::TensorOpMultiplicandCongruous< sizeof_bits< Element_ >::value, 64 >, InstructionShape_, OpDelta_, 32, PartitionsK_ > , cutlass::gemm::warp::MmaTensorOpMultiplicandTileIterator< Shape_, Operand_, Element_, cutlass::layout::TensorOpMultiplicandCrosswise< sizeof_bits< Element_ >::value, Crosswise >, InstructionShape_, OpDelta_, 32, PartitionsK_ > , cutlass::gemm::warp::MmaVoltaTensorOpAccumulatorTileIterator< Shape_, Element_, Layout_, InstructionShape_, OpDelta_ > , cutlass::gemm::warp::MmaVoltaTensorOpMultiplicandTileIterator< Shape_, Operand::kA, Element_, cutlass::layout::ColumnMajorVoltaTensorOpMultiplicandCongruous< sizeof_bits< Element_ >::value >, InstructionShape_, OpDelta_, 32 > , cutlass::gemm::warp::MmaVoltaTensorOpMultiplicandTileIterator< Shape_, Operand::kA, Element_, cutlass::layout::VoltaTensorOpMultiplicandCongruous< sizeof_bits< Element_ >::value >, InstructionShape_, OpDelta_, 32 > , cutlass::gemm::warp::MmaVoltaTensorOpMultiplicandTileIterator< Shape_, Operand::kB, Element_, cutlass::layout::RowMajorVoltaTensorOpMultiplicandBCongruous< sizeof_bits< Element_ >::value >, InstructionShape_, OpDelta_, 32 > , cutlass::gemm::warp::MmaVoltaTensorOpMultiplicandTileIterator< Shape_, Operand::kB, Element_, cutlass::layout::VoltaTensorOpMultiplicandBCongruous< sizeof_bits< Element_ >::value >, InstructionShape_, OpDelta_, 32 > , cutlass::gemm::warp::MmaVoltaTensorOpMultiplicandTileIterator< Shape_, Operand_, Element_, cutlass::layout::ColumnMajorVoltaTensorOpMultiplicandCrosswise< sizeof_bits< Element_ >::value, KBlock >, InstructionShape_, OpDelta_, 32 > , cutlass::gemm::warp::MmaVoltaTensorOpMultiplicandTileIterator< Shape_, Operand_, Element_, cutlass::layout::RowMajorVoltaTensorOpMultiplicandCrosswise< sizeof_bits< Element_ >::value, KBlock >, InstructionShape_, OpDelta_, 32 > , cutlass::gemm::warp::MmaVoltaTensorOpMultiplicandTileIterator< Shape_, Operand_, Element_, cutlass::layout::VoltaTensorOpMultiplicandCrosswise< sizeof_bits< Element_ >::value, KBlock >, InstructionShape_, OpDelta_, 32 >
• kPaddingM : cutlass::gemm::threadblock::DefaultMmaCore< Shape_, WarpShape_, GemmShape< 1, 1, 1 >, ElementA_, layout::RowMajor, ElementB_, layout::ColumnMajor, ElementC_, LayoutC_, arch::OpClassSimt, 2, Operator_ > , cutlass::gemm::threadblock::DefaultMmaCore< Shape_, WarpShape_, GemmShape< 1, 1, 1 >, ElementA_, layout::RowMajor, ElementB_, layout::RowMajor, ElementC_, LayoutC_, arch::OpClassSimt, 2, Operator_ > , cutlass::gemm::threadblock::DefaultMmaCore< Shape_, WarpShape_, GemmShape< 1, 1, 4 >, int8_t, layout::ColumnMajor, int8_t, layout::ColumnMajor, ElementC_, LayoutC_, arch::OpClassSimt, 2, Operator_ > , cutlass::gemm::threadblock::DefaultMmaCore< Shape_, WarpShape_, GemmShape< 1, 1, 4 >, int8_t, layout::RowMajor, int8_t, layout::ColumnMajor, ElementC_, LayoutC_, arch::OpClassSimt, 2, Operator_ > , cutlass::gemm::threadblock::DefaultMmaCore< Shape_, WarpShape_, GemmShape< 1, 1, 4 >, int8_t, layout::RowMajor, int8_t, layout::RowMajor, ElementC_, LayoutC_, arch::OpClassSimt, 2, Operator_ >
• kPaddingN : cutlass::gemm::threadblock::DefaultMmaCore< Shape_, WarpShape_, GemmShape< 1, 1, 1 >, ElementA_, layout::ColumnMajor, ElementB_, layout::ColumnMajor, ElementC_, LayoutC_, arch::OpClassSimt, 2, Operator_ > , cutlass::gemm::threadblock::DefaultMmaCore< Shape_, WarpShape_, GemmShape< 1, 1, 1 >, ElementA_, layout::RowMajor, ElementB_, layout::ColumnMajor, ElementC_, LayoutC_, arch::OpClassSimt, 2, Operator_ > , cutlass::gemm::threadblock::DefaultMmaCore< Shape_, WarpShape_, GemmShape< 1, 1, 4 >, int8_t, layout::ColumnMajor, int8_t, layout::ColumnMajor, ElementC_, LayoutC_, arch::OpClassSimt, 2, Operator_ > , cutlass::gemm::threadblock::DefaultMmaCore< Shape_, WarpShape_, GemmShape< 1, 1, 4 >, int8_t, layout::RowMajor, int8_t, layout::ColumnMajor, ElementC_, LayoutC_, arch::OpClassSimt, 2, Operator_ > , cutlass::gemm::threadblock::DefaultMmaCore< Shape_, WarpShape_, GemmShape< 1, 1, 4 >, int8_t, layout::RowMajor, int8_t, layout::RowMajor, ElementC_, LayoutC_, arch::OpClassSimt, 2, Operator_ >
• kPartitionsK : cutlass::epilogue::threadblock::DefaultEpilogueComplexTensorOp< Shape_, WarpMmaTensorOp_, PartitionsK, OutputOp_, ElementsPerAccess > , cutlass::epilogue::threadblock::DefaultEpilogueSimt< Shape_, WarpMmaSimt_, OutputOp_, ElementsPerAccess > , cutlass::epilogue::threadblock::DefaultEpilogueTensorOp< Shape_, WarpMmaTensorOp_, PartitionsK, OutputOp_, ElementsPerAccess > , cutlass::epilogue::threadblock::DefaultEpilogueVoltaTensorOp< Shape_, WarpMmaTensorOp_, PartitionsK, OutputOp_, ElementsPerAccess > , cutlass::epilogue::threadblock::DefaultEpilogueWmmaTensorOp< Shape_, WarpMmaTensorOp_, PartitionsK, OutputOp_, ElementsPerAccess > , cutlass::epilogue::threadblock::DefaultInterleavedEpilogueTensorOp< Shape_, WarpMmaTensorOp_, PartitionsK, OutputOp_, ElementsPerAccess, InterleavedK, IsBetaZero, isSplitK > , cutlass::epilogue::threadblock::DefaultInterleavedThreadMapTensorOp< ThreadblockShape_, WarpShape_, PartitionsK, Element_, ElementsPerAccess, InterleavedK > , cutlass::epilogue::threadblock::DefaultThreadMapSimt< ThreadblockShape_, WarpShape_, MmaSimtPolicy_, PartitionsK, Element_, ElementsPerAccess > , cutlass::epilogue::threadblock::DefaultThreadMapTensorOp< ThreadblockShape_, WarpShape_, PartitionsK, Element_, ElementsPerAccess > , cutlass::epilogue::threadblock::DefaultThreadMapVoltaTensorOp< ThreadblockShape_, WarpShape_, PartitionsK, ElementOutput_, ElementsPerAccess, float > , cutlass::epilogue::threadblock::DefaultThreadMapVoltaTensorOp< ThreadblockShape_, WarpShape_, PartitionsK, ElementOutput_, ElementsPerAccess, half_t > , cutlass::epilogue::threadblock::DefaultThreadMapWmmaTensorOp< ThreadblockShape_, WarpShape_, InstructionShape_, PartitionsK, Element_, ElementsPerAccess > , cutlass::epilogue::threadblock::Epilogue< Shape_, WarpMmaOperator_, PartitionsK, OutputTileIterator_, AccumulatorFragmentIterator_, WarpTileIterator_, SharedLoadIterator_, OutputOp_, Padding_ > , cutlass::epilogue::threadblock::EpilogueBase< Shape_, WarpMmaOperator_, PartitionsK, AccumulatorFragmentIterator_, WarpTileIterator_, Padding_ > , cutlass::epilogue::threadblock::InterleavedEpilogue< Shape_, WarpMmaOperator_, PartitionsK, OutputTileIterator_, AccumulatorFragmentIterator_, OutputOp_, InterleavedK, IsBetaZero > , cutlass::gemm::kernel::DefaultGemm< ElementA, layout::ColumnMajorInterleaved< InterleavedK >, kAlignmentA, ElementB, layout::RowMajorInterleaved< InterleavedK >, kAlignmentB, ElementC, layout::ColumnMajorInterleaved< InterleavedK >, int32_t, arch::OpClassTensorOp, arch::Sm75, ThreadblockShape, WarpShape, InstructionShape, EpilogueOutputOp, ThreadblockSwizzle, 2, SplitKSerial, Operator, IsBetaZero > , cutlass::gemm::kernel::DefaultGemm< ElementA, LayoutA, kAlignmentA, ElementB, LayoutB, kAlignmentB, ElementC, layout::RowMajor, ElementAccumulator, arch::OpClassTensorOp, arch::Sm70, ThreadblockShape, WarpShape, GemmShape< 8, 8, 4 >, EpilogueOutputOp, ThreadblockSwizzle, 2, SplitKSerial, Operator > , cutlass::gemm::kernel::DefaultGemm< ElementA, LayoutA, kAlignmentA, ElementB, LayoutB, kAlignmentB, ElementC, layout::RowMajor, ElementAccumulator, arch::OpClassTensorOp, arch::Sm75, ThreadblockShape, WarpShape, InstructionShape, EpilogueOutputOp, ThreadblockSwizzle, 2, SplitKSerial, Operator > , cutlass::gemm::threadblock::MmaPolicy< Operator_, SmemPaddingA_, SmemPaddingB_, PartitionsK > , cutlass::gemm::warp::MmaSimtTileIterator< Shape_, Operand::kA, Element_, layout::ColumnMajorInterleaved< 4 >, Policy_, PartitionsK, PartitionGroupSize > , cutlass::gemm::warp::MmaSimtTileIterator< Shape_, Operand::kB, Element_, layout::RowMajorInterleaved< 4 >, Policy_, PartitionsK, PartitionGroupSize > , cutlass::gemm::warp::MmaTensorOp< Shape_, ElementA_, LayoutA_, ElementB_, LayoutB_, ElementC_, LayoutC_, Policy_, PartitionsK_, AccumulatorsInRowMajor, PartitionsN_, Enable > , cutlass::gemm::warp::MmaTensorOpMultiplicandTileIterator< Shape_, Operand_, Element_, cutlass::layout::TensorOpMultiplicandCongruous< sizeof_bits< Element_ >::value, 64 >, InstructionShape_, OpDelta_, 32, PartitionsK_ > , cutlass::gemm::warp::MmaTensorOpMultiplicandTileIterator< Shape_, Operand_, Element_, cutlass::layout::TensorOpMultiplicandCrosswise< sizeof_bits< Element_ >::value, Crosswise >, InstructionShape_, OpDelta_, 32, PartitionsK_ >
• kPartitionsN : cutlass::gemm::warp::MmaTensorOp< Shape_, ElementA_, LayoutA_, ElementB_, LayoutB_, ElementC_, LayoutC_, Policy_, PartitionsK_, AccumulatorsInRowMajor, PartitionsN_, Enable >
• kPartitionsPerStage : cutlass::reduction::kernel::ReduceSplitK< Shape_, OutputOp_, ReductionOp_, PartitionsPerStage >
• kPointerCount : cutlass::transform::threadblock::RegularTileAccessIterator< Shape_, Element_, layout::TensorOpMultiplicandCongruous< sizeof_bits< Element_ >::value, int(128/sizeof(Element_))>, AdvanceRank, ThreadMap_, Alignment >::Detail , cutlass::transform::threadblock::RegularTileAccessIterator< Shape_, Element_, layout::TensorOpMultiplicandCrosswise< sizeof_bits< Element_ >::value, Crosswise >, AdvanceRank, ThreadMap_, Alignment >::Detail , cutlass::transform::threadblock::RegularTileIterator< Shape_, Element_, layout::VoltaTensorOpMultiplicandBCongruous< sizeof_bits< Element_ >::value >, AdvanceRank, ThreadMap_, Alignment >::Detail , cutlass::transform::threadblock::RegularTileIterator< Shape_, Element_, layout::VoltaTensorOpMultiplicandCongruous< sizeof_bits< Element_ >::value >, AdvanceRank, ThreadMap_, Alignment >::Detail , cutlass::transform::threadblock::RegularTileIterator< Shape_, Element_, layout::VoltaTensorOpMultiplicandCrosswise< sizeof_bits< Element_ >::value, Shape_::kContiguous >, AdvanceRank, ThreadMap_, Alignment >::Detail
• kPredicateByteCount : cutlass::transform::threadblock::PredicatedTileAccessIterator2dThreadTile< Shape_, Element_, layout::PitchLinear, AdvanceRank, ThreadMap_, AccessType_ > , cutlass::transform::threadblock::PredicatedTileAccessIterator< Shape_, Element_, layout::PitchLinear, AdvanceRank, ThreadMap_, AccessType_ >
• kPredicateCount : cutlass::transform::threadblock::PredicatedTileAccessIterator< Shape_, Element_, layout::PitchLinear, AdvanceRank, ThreadMap_, AccessType_ >
• kPredicateMask : cutlass::transform::threadblock::PredicatedTileAccessIterator2dThreadTile< Shape_, Element_, layout::PitchLinear, AdvanceRank, ThreadMap_, AccessType_ > , cutlass::transform::threadblock::PredicatedTileAccessIterator< Shape_, Element_, layout::PitchLinear, AdvanceRank, ThreadMap_, AccessType_ >
• kPredicates : cutlass::PredicateVector< kPredicates_, kPredicatesPerByte_, kPredicateStart_ >
• kPredicatesPerByte : cutlass::PredicateVector< kPredicates_, kPredicatesPerByte_, kPredicateStart_ > , cutlass::transform::threadblock::PredicatedTileAccessIterator2dThreadTile< Shape_, Element_, layout::PitchLinear, AdvanceRank, ThreadMap_, AccessType_ > , cutlass::transform::threadblock::PredicatedTileAccessIterator< Shape_, Element_, layout::PitchLinear, AdvanceRank, ThreadMap_, AccessType_ >
• kPredicatesPerWord : cutlass::transform::threadblock::PredicatedTileAccessIterator2dThreadTile< Shape_, Element_, layout::PitchLinear, AdvanceRank, ThreadMap_, AccessType_ > , cutlass::transform::threadblock::PredicatedTileAccessIterator< Shape_, Element_, layout::PitchLinear, AdvanceRank, ThreadMap_, AccessType_ >
• kPredicateStart : cutlass::PredicateVector< kPredicates_, kPredicatesPerByte_, kPredicateStart_ >
• kPredicateWordCount : cutlass::transform::threadblock::PredicatedTileAccessIterator2dThreadTile< Shape_, Element_, layout::PitchLinear, AdvanceRank, ThreadMap_, AccessType_ > , cutlass::transform::threadblock::PredicatedTileAccessIterator< Shape_, Element_, layout::PitchLinear, AdvanceRank, ThreadMap_, AccessType_ >
• kRank : cutlass::Coord< Rank_, Index_, LongIndex_ > , cutlass::HostTensor< Element_, Layout_ > , cutlass::IdentityTensorLayout< Rank > , cutlass::layout::ColumnMajor , cutlass::layout::ColumnMajorBlockLinear< BlockRows, BlockColumns > , cutlass::layout::ColumnMajorInterleaved< Interleave > , cutlass::layout::ColumnMajorTensorOpMultiplicandCongruous< ElementSize, Crosswise > , cutlass::layout::ColumnMajorTensorOpMultiplicandCrosswise< ElementSize, Crosswise > , cutlass::layout::ColumnMajorVoltaTensorOpMultiplicandBCongruous< ElementSize > , cutlass::layout::ColumnMajorVoltaTensorOpMultiplicandCongruous< ElementSize > , cutlass::layout::ColumnMajorVoltaTensorOpMultiplicandCrosswise< ElementSize, KBlock > , cutlass::layout::ContiguousMatrix , cutlass::layout::GeneralMatrix , cutlass::layout::PackedVectorLayout , cutlass::layout::PitchLinear , cutlass::layout::RowMajor , cutlass::layout::RowMajorBlockLinear< BlockRows, BlockColumns > , cutlass::layout::RowMajorInterleaved< Interleave > , cutlass::layout::RowMajorTensorOpMultiplicandCongruous< ElementSize, Crosswise > , cutlass::layout::RowMajorTensorOpMultiplicandCrosswise< ElementSize, Crosswise > , cutlass::layout::RowMajorVoltaTensorOpMultiplicandBCongruous< ElementSize > , cutlass::layout::RowMajorVoltaTensorOpMultiplicandCongruous< ElementSize > , cutlass::layout::RowMajorVoltaTensorOpMultiplicandCrosswise< ElementSize, KBlock > , cutlass::layout::TensorCxRSKx< Interleave > , cutlass::layout::TensorNCHW , cutlass::layout::TensorNCxHWx< Interleave > , cutlass::layout::TensorNHWC , cutlass::layout::TensorOpMultiplicand< ElementSize, Crosswise > , cutlass::layout::TensorOpMultiplicandColumnMajorInterleaved< ElementSize, InterleavedK > , cutlass::layout::TensorOpMultiplicandCongruous< ElementSize, Crosswise > , cutlass::layout::TensorOpMultiplicandCongruous< 32, Crosswise > , cutlass::layout::TensorOpMultiplicandCrosswise< ElementSize, Crosswise > , cutlass::layout::TensorOpMultiplicandRowMajorInterleaved< ElementSize, InterleavedK > , cutlass::layout::VoltaTensorOpMultiplicandBCongruous< ElementSize > , cutlass::layout::VoltaTensorOpMultiplicandCongruous< ElementSize > , cutlass::layout::VoltaTensorOpMultiplicandCrosswise< ElementSize, KBlock > , cutlass::TensorRef< Element_, Layout_ > , cutlass::TensorView< Element_, Layout_ > , cutlass::thread::Matrix< Element, Rows, Columns, Layout >
• kRealIndex : cutlass::epilogue::warp::FragmentIteratorComplexTensorOp< WarpShape_, OperatorShape_, OperatorElementC_, OperatorFragmentC_, layout::RowMajor >
• kRound : cutlass::epilogue::thread::Convert< ElementOutput_, Count, ElementAccumulator_, Round > , cutlass::epilogue::thread::LinearCombination< ElementOutput_, Count, ElementAccumulator_, ElementCompute_, Round > , cutlass::epilogue::thread::LinearCombinationClamp< ElementOutput_, Count, ElementAccumulator_, ElementCompute_, Round > , cutlass::epilogue::thread::LinearCombinationRelu< ElementOutput_, Count, ElementAccumulator_, ElementCompute_, Round > , cutlass::epilogue::thread::LinearCombinationRelu< ElementOutput_, Count, int, float, Round >
• kRow : cutlass::epilogue::threadblock::OutputTileShape< Column, Row, Group, Cluster, Tile > , cutlass::MatrixShape< Row_, Column_ >
• kRows : cutlass::thread::Matrix< Element, Rows, Columns, Layout >
• kRowsPerIteration : cutlass::epilogue::warp::SimtPolicy< WarpShape_, Operator_, layout::RowMajor, MmaSimtPolicy_ > , cutlass::epilogue::warp::TensorOpPolicy< WarpShape, OperatorShape, layout::ColumnMajorInterleaved< InterleavedK > > , cutlass::epilogue::warp::TensorOpPolicy< WarpShape, OperatorShape, layout::RowMajor > , cutlass::epilogue::warp::VoltaTensorOpPolicy< WarpShape_, gemm::GemmShape< 32, 32, 4 >, float, layout::RowMajor > , cutlass::epilogue::warp::VoltaTensorOpPolicy< WarpShape_, gemm::GemmShape< 32, 32, 4 >, half_t, layout::RowMajor >
• kRowsPerMmaTile : cutlass::epilogue::warp::VoltaTensorOpPolicy< WarpShape_, gemm::GemmShape< 32, 32, 4 >, float, layout::RowMajor >
• kRowsPerQuad : cutlass::epilogue::warp::TileIteratorVoltaTensorOp< WarpShape_, gemm::GemmShape< 32, 32, 4 >, float, layout::RowMajor >::Detail , cutlass::epilogue::warp::TileIteratorVoltaTensorOp< WarpShape_, gemm::GemmShape< 32, 32, 4 >, half_t, layout::RowMajor >::Detail
• kShapeRow : cutlass::epilogue::threadblock::detail::RowArrangement< Shape, WarpsRemaining, ElementsPerAccess, ElementSize, true >::Detail
• kShapeWidth : cutlass::epilogue::threadblock::detail::RowArrangement< Shape, WarpsRemaining, ElementsPerAccess, ElementSize, true >::Detail
• kSharedMemAlignment : cutlass::epilogue::threadblock::DefaultEpilogueVoltaTensorOp< Shape_, WarpMmaTensorOp_, PartitionsK, OutputOp_, ElementsPerAccess >
• kSigned : cutlass::integer_subbyte< Bits, Signed >
• kSizeBits : cutlass::Array< T, N, false >
• kSplitKSerial : cutlass::gemm::device::Gemm< ElementA_, LayoutA_, ElementB_, LayoutB_, ElementC_, LayoutC_, ElementAccumulator_, OperatorClass_, ArchTag_, ThreadblockShape_, WarpShape_, InstructionShape_, EpilogueOutputOp_, ThreadblockSwizzle_, Stages, AlignmentA, AlignmentB, SplitKSerial, Operator_, IsBetaZero > , cutlass::gemm::device::Gemm< ElementA_, LayoutA_, ElementB_, LayoutB_, ElementC_, layout::ColumnMajor, ElementAccumulator_, OperatorClass_, ArchTag_, ThreadblockShape_, WarpShape_, InstructionShape_, EpilogueOutputOp_, ThreadblockSwizzle_, Stages, AlignmentA, AlignmentB, SplitKSerial, Operator_, IsBetaZero > , cutlass::gemm::device::GemmBatched< ElementA_, LayoutA_, ElementB_, LayoutB_, ElementC_, layout::ColumnMajor, ElementAccumulator_, OperatorClass_, ArchTag_, ThreadblockShape_, WarpShape_, InstructionShape_, EpilogueOutputOp_, ThreadblockSwizzle_, Stages, AlignmentA, AlignmentB, Operator_ > , cutlass::gemm::device::GemmComplex< ElementA_, LayoutA_, ElementB_, LayoutB_, ElementC_, LayoutC_, ElementAccumulator_, OperatorClass_, ArchTag_, ThreadblockShape_, WarpShape_, InstructionShape_, EpilogueOutputOp_, ThreadblockSwizzle_, Stages, TransformA, TransformB, SplitKSerial > , cutlass::gemm::device::GemmComplex< ElementA_, LayoutA_, ElementB_, LayoutB_, ElementC_, layout::ColumnMajor, ElementAccumulator_, OperatorClass_, ArchTag_, ThreadblockShape_, WarpShape_, InstructionShape_, EpilogueOutputOp_, ThreadblockSwizzle_, Stages, TransformA, TransformB, SplitKSerial > , cutlass::gemm::kernel::Gemm< Mma_, Epilogue_, ThreadblockSwizzle_, SplitKSerial >
• kStages : cutlass::gemm::device::DefaultGemmConfiguration< arch::OpClassSimt, ArchTag, ElementA, ElementB, ElementC, ElementAccumulator > , cutlass::gemm::device::DefaultGemmConfiguration< arch::OpClassSimt, ArchTag, int8_t, int8_t, ElementC, int32_t > , cutlass::gemm::device::DefaultGemmConfiguration< arch::OpClassTensorOp, arch::Sm70, ElementA, ElementB, ElementC, ElementAccumulator > , cutlass::gemm::device::DefaultGemmConfiguration< arch::OpClassTensorOp, arch::Sm75, ElementA, ElementB, ElementC, ElementAccumulator > , cutlass::gemm::device::DefaultGemmConfiguration< arch::OpClassTensorOp, arch::Sm75, int4b_t, int4b_t, ElementC, int32_t > , cutlass::gemm::device::DefaultGemmConfiguration< arch::OpClassTensorOp, arch::Sm75, int4b_t, uint4b_t, ElementC, int32_t > , cutlass::gemm::device::DefaultGemmConfiguration< arch::OpClassTensorOp, arch::Sm75, int8_t, int8_t, ElementC, int32_t > , cutlass::gemm::device::DefaultGemmConfiguration< arch::OpClassTensorOp, arch::Sm75, int8_t, uint8_t, ElementC, int32_t > , cutlass::gemm::device::DefaultGemmConfiguration< arch::OpClassTensorOp, arch::Sm75, uint4b_t, int4b_t, ElementC, int32_t > , cutlass::gemm::device::DefaultGemmConfiguration< arch::OpClassTensorOp, arch::Sm75, uint4b_t, uint4b_t, ElementC, int32_t > , cutlass::gemm::device::DefaultGemmConfiguration< arch::OpClassTensorOp, arch::Sm75, uint8_t, int8_t, ElementC, int32_t > , cutlass::gemm::device::DefaultGemmConfiguration< arch::OpClassTensorOp, arch::Sm75, uint8_t, uint8_t, ElementC, int32_t > , cutlass::gemm::device::DefaultGemmConfiguration< arch::OpClassWmmaTensorOp, ArchTag, ElementA, ElementB, ElementC, ElementAccumulator > , cutlass::gemm::device::Gemm< ElementA_, LayoutA_, ElementB_, LayoutB_, ElementC_, LayoutC_, ElementAccumulator_, OperatorClass_, ArchTag_, ThreadblockShape_, WarpShape_, InstructionShape_, EpilogueOutputOp_, ThreadblockSwizzle_, Stages, AlignmentA, AlignmentB, SplitKSerial, Operator_, IsBetaZero > , cutlass::gemm::device::Gemm< ElementA_, LayoutA_, ElementB_, LayoutB_, ElementC_, layout::ColumnMajor, ElementAccumulator_, OperatorClass_, ArchTag_, ThreadblockShape_, WarpShape_, InstructionShape_, EpilogueOutputOp_, ThreadblockSwizzle_, Stages, AlignmentA, AlignmentB, SplitKSerial, Operator_, IsBetaZero > , cutlass::gemm::device::GemmBatched< ElementA_, LayoutA_, ElementB_, LayoutB_, ElementC_, LayoutC_, ElementAccumulator_, OperatorClass_, ArchTag_, ThreadblockShape_, WarpShape_, InstructionShape_, EpilogueOutputOp_, ThreadblockSwizzle_, Stages, AlignmentA, AlignmentB, Operator_ > , cutlass::gemm::device::GemmBatched< ElementA_, LayoutA_, ElementB_, LayoutB_, ElementC_, layout::ColumnMajor, ElementAccumulator_, OperatorClass_, ArchTag_, ThreadblockShape_, WarpShape_, InstructionShape_, EpilogueOutputOp_, ThreadblockSwizzle_, Stages, AlignmentA, AlignmentB, Operator_ > , cutlass::gemm::device::GemmComplex< ElementA_, LayoutA_, ElementB_, LayoutB_, ElementC_, LayoutC_, ElementAccumulator_, OperatorClass_, ArchTag_, ThreadblockShape_, WarpShape_, InstructionShape_, EpilogueOutputOp_, ThreadblockSwizzle_, Stages, TransformA, TransformB, SplitKSerial > , cutlass::gemm::device::GemmComplex< ElementA_, LayoutA_, ElementB_, LayoutB_, ElementC_, layout::ColumnMajor, ElementAccumulator_, OperatorClass_, ArchTag_, ThreadblockShape_, WarpShape_, InstructionShape_, EpilogueOutputOp_, ThreadblockSwizzle_, Stages, TransformA, TransformB, SplitKSerial > , cutlass::gemm::device::GemmSplitKParallel< ElementA_, LayoutA_, ElementB_, LayoutB_, ElementC_, LayoutC_, ElementAccumulator_, OperatorClass_, ArchTag_, ThreadblockShape_, WarpShape_, InstructionShape_, EpilogueOutputOp_, ConvertScaledOp_, ReductionOp_, ThreadblockSwizzle_, Stages, kAlignmentA, kAlignmentB, Operator_ > , cutlass::gemm::device::GemmSplitKParallel< ElementA_, LayoutA_, ElementB_, LayoutB_, ElementC_, layout::ColumnMajor, ElementAccumulator_, OperatorClass_, ArchTag_, ThreadblockShape_, WarpShape_, InstructionShape_, EpilogueOutputOp_, ConvertScaledOp_, ReductionOp_, ThreadblockSwizzle_, Stages, kAlignmentA, kAlignmentB, Operator_ > , cutlass::gemm::threadblock::MmaBase< Shape_, Policy_, Stages, Enable >
• kStorageElements : cutlass::Array< T, N, false > , cutlass::Array< T, N, true >
• kStrided : cutlass::layout::PitchLinearShape< Contiguous, Strided >
• kStrideRank : cutlass::IdentityTensorLayout< Rank > , cutlass::layout::ColumnMajor , cutlass::layout::ColumnMajorBlockLinear< BlockRows, BlockColumns > , cutlass::layout::ColumnMajorInterleaved< Interleave > , cutlass::layout::ColumnMajorTensorOpMultiplicandCongruous< ElementSize, Crosswise > , cutlass::layout::ColumnMajorTensorOpMultiplicandCrosswise< ElementSize, Crosswise > , cutlass::layout::ColumnMajorVoltaTensorOpMultiplicandBCongruous< ElementSize > , cutlass::layout::ColumnMajorVoltaTensorOpMultiplicandCongruous< ElementSize > , cutlass::layout::ColumnMajorVoltaTensorOpMultiplicandCrosswise< ElementSize, KBlock > , cutlass::layout::ContiguousMatrix , cutlass::layout::GeneralMatrix , cutlass::layout::PackedVectorLayout , cutlass::layout::PitchLinear , cutlass::layout::RowMajor , cutlass::layout::RowMajorBlockLinear< BlockRows, BlockColumns > , cutlass::layout::RowMajorInterleaved< Interleave > , cutlass::layout::RowMajorTensorOpMultiplicandCongruous< ElementSize, Crosswise > , cutlass::layout::RowMajorTensorOpMultiplicandCrosswise< ElementSize, Crosswise > , cutlass::layout::RowMajorVoltaTensorOpMultiplicandBCongruous< ElementSize > , cutlass::layout::RowMajorVoltaTensorOpMultiplicandCongruous< ElementSize > , cutlass::layout::RowMajorVoltaTensorOpMultiplicandCrosswise< ElementSize, KBlock > , cutlass::layout::TensorCxRSKx< Interleave > , cutlass::layout::TensorNCHW , cutlass::layout::TensorNCxHWx< Interleave > , cutlass::layout::TensorNHWC , cutlass::layout::TensorOpMultiplicand< ElementSize, Crosswise > , cutlass::layout::TensorOpMultiplicandColumnMajorInterleaved< ElementSize, InterleavedK > , cutlass::layout::TensorOpMultiplicandCongruous< ElementSize, Crosswise > , cutlass::layout::TensorOpMultiplicandCongruous< 32, Crosswise > , cutlass::layout::TensorOpMultiplicandCrosswise< ElementSize, Crosswise > , cutlass::layout::TensorOpMultiplicandRowMajorInterleaved< ElementSize, InterleavedK > , cutlass::layout::VoltaTensorOpMultiplicandBCongruous< ElementSize > , cutlass::layout::VoltaTensorOpMultiplicandCongruous< ElementSize > , cutlass::layout::VoltaTensorOpMultiplicandCrosswise< ElementSize, KBlock >
• kTargetAccessRows : cutlass::epilogue::threadblock::detail::RowArrangement< Shape, WarpsRemaining, ElementsPerAccess, ElementSize, true >::Detail
• kTargetMemoryAccessWidth : cutlass::epilogue::threadblock::detail::RowArrangement< Shape, WarpsRemaining, ElementsPerAccess, ElementSize, true >::Detail
• kTensorOpRows : cutlass::epilogue::threadblock::DefaultInterleavedThreadMapTensorOp< ThreadblockShape_, WarpShape_, PartitionsK, Element_, ElementsPerAccess, InterleavedK >::Detail , cutlass::epilogue::threadblock::DefaultThreadMapTensorOp< ThreadblockShape_, WarpShape_, PartitionsK, Element_, ElementsPerAccess >::Detail , cutlass::epilogue::threadblock::DefaultThreadMapVoltaTensorOp< ThreadblockShape_, WarpShape_, PartitionsK, ElementOutput_, ElementsPerAccess, float >::Detail , cutlass::epilogue::threadblock::DefaultThreadMapVoltaTensorOp< ThreadblockShape_, WarpShape_, PartitionsK, ElementOutput_, ElementsPerAccess, half_t >::Detail , cutlass::epilogue::threadblock::DefaultThreadMapWmmaTensorOp< ThreadblockShape_, WarpShape_, InstructionShape_, PartitionsK, Element_, ElementsPerAccess >::Detail
• kThreadblockAccesses : cutlass::epilogue::EpilogueWorkspace< Shape_, WarpCount, FragmentC_ >
• kThreadCount : cutlass::gemm::kernel::Gemm< Mma_, Epilogue_, ThreadblockSwizzle_, SplitKSerial > , cutlass::gemm::kernel::GemmBatched< Mma_, Epilogue_, ThreadblockSwizzle_ > , cutlass::gemm::kernel::GemmSplitKParallel< Mma_, Epilogue_, ThreadblockSwizzle_ > , cutlass::gemm::warp::MmaComplexTensorOp< Shape_, complex< RealElementA >, LayoutA_, complex< RealElementB >, LayoutB_, complex< RealElementC >, LayoutC_, Policy_, TransformA, TransformB, Enable > , cutlass::gemm::warp::MmaTensorOp< Shape_, ElementA_, LayoutA_, ElementB_, LayoutB_, ElementC_, LayoutC_, Policy_, PartitionsK_, AccumulatorsInRowMajor, PartitionsN_, Enable > , cutlass::gemm::warp::MmaVoltaTensorOp< Shape_, ElementA_, LayoutA_, ElementB_, LayoutB_, ElementC_, LayoutC_, Policy_, Enable >
• kThreads : cutlass::epilogue::threadblock::DefaultInterleavedThreadMapTensorOp< ThreadblockShape_, WarpShape_, PartitionsK, Element_, ElementsPerAccess, InterleavedK >::Detail , cutlass::epilogue::threadblock::DefaultThreadMapSimt< ThreadblockShape_, WarpShape_, MmaSimtPolicy_, PartitionsK, Element_, ElementsPerAccess >::Detail , cutlass::epilogue::threadblock::DefaultThreadMapTensorOp< ThreadblockShape_, WarpShape_, PartitionsK, Element_, ElementsPerAccess >::Detail , cutlass::epilogue::threadblock::DefaultThreadMapVoltaTensorOp< ThreadblockShape_, WarpShape_, PartitionsK, ElementOutput_, ElementsPerAccess, float >::Detail , cutlass::epilogue::threadblock::DefaultThreadMapVoltaTensorOp< ThreadblockShape_, WarpShape_, PartitionsK, ElementOutput_, ElementsPerAccess, half_t >::Detail , cutlass::epilogue::threadblock::DefaultThreadMapWmmaTensorOp< ThreadblockShape_, WarpShape_, InstructionShape_, PartitionsK, Element_, ElementsPerAccess >::Detail , cutlass::epilogue::threadblock::InterleavedOutputTileThreadMap< WarpCount_, MmaCount_, Threads, ElementsPerAccess, ElementSize > , cutlass::epilogue::threadblock::InterleavedPredicatedTileIterator< ThreadMap_, Element_, InterleavedK > , cutlass::epilogue::threadblock::OutputTileOptimalThreadMap< Shape_, Count_, Threads, ElementsPerAccess, ElementSize >::CompactedThreadMap , cutlass::epilogue::threadblock::OutputTileOptimalThreadMap< Shape_, Count_, Threads, ElementsPerAccess, ElementSize > , cutlass::epilogue::threadblock::OutputTileThreadMap< ThreadMap_, Shape_, Iterations_, Delta_, Count_ > , cutlass::epilogue::threadblock::PredicatedTileIterator< ThreadMap_, Element_ > , cutlass::epilogue::threadblock::SharedLoadIterator< ThreadMap_, Element_, MaxAlignment > , cutlass::gemm::threadblock::DefaultMmaCore< Shape_, WarpShape_, GemmShape< 1, 1, 1 >, ElementA_, layout::ColumnMajor, ElementB_, layout::ColumnMajor, ElementC_, LayoutC_, arch::OpClassSimt, 2, Operator_ > , cutlass::gemm::threadblock::DefaultMmaCore< Shape_, WarpShape_, GemmShape< 1, 1, 1 >, ElementA_, layout::ColumnMajor, ElementB_, layout::RowMajor, ElementC_, LayoutC_, arch::OpClassSimt, 2, Operator_ > , cutlass::gemm::threadblock::DefaultMmaCore< Shape_, WarpShape_, GemmShape< 1, 1, 1 >, ElementA_, layout::ColumnMajor, ElementB_, layout::RowMajor, ElementC_, LayoutC_, arch::OpClassSimt, 2, Operator_, > , cutlass::gemm::threadblock::DefaultMmaCore< Shape_, WarpShape_, GemmShape< 1, 1, 1 >, ElementA_, layout::RowMajor, ElementB_, layout::ColumnMajor, ElementC_, LayoutC_, arch::OpClassSimt, 2, Operator_ > , cutlass::gemm::threadblock::DefaultMmaCore< Shape_, WarpShape_, GemmShape< 1, 1, 1 >, ElementA_, layout::RowMajor, ElementB_, layout::RowMajor, ElementC_, LayoutC_, arch::OpClassSimt, 2, Operator_ > , cutlass::gemm::threadblock::DefaultMmaCore< Shape_, WarpShape_, GemmShape< 1, 1, 4 >, int8_t, layout::ColumnMajor, int8_t, layout::ColumnMajor, ElementC_, LayoutC_, arch::OpClassSimt, 2, Operator_ > , cutlass::gemm::threadblock::DefaultMmaCore< Shape_, WarpShape_, GemmShape< 1, 1, 4 >, int8_t, layout::ColumnMajor, int8_t, layout::RowMajor, ElementC_, LayoutC_, arch::OpClassSimt, 2, Operator_ > , cutlass::gemm::threadblock::DefaultMmaCore< Shape_, WarpShape_, GemmShape< 1, 1, 4 >, int8_t, layout::RowMajor, int8_t, layout::ColumnMajor, ElementC_, LayoutC_, arch::OpClassSimt, 2, Operator_ > , cutlass::gemm::threadblock::DefaultMmaCore< Shape_, WarpShape_, GemmShape< 1, 1, 4 >, int8_t, layout::RowMajor, int8_t, layout::RowMajor, ElementC_, LayoutC_, arch::OpClassSimt, 2, Operator_ > , cutlass::gemm::threadblock::DefaultMmaCore< Shape_, WarpShape_, GemmShape< 8, 8, 4 >, ElementA_, layout::ColumnMajor, ElementB_, layout::ColumnMajor, ElementC_, LayoutC_, arch::OpClassTensorOp, 2, Operator_ > , cutlass::gemm::threadblock::DefaultMmaCore< Shape_, WarpShape_, GemmShape< 8, 8, 4 >, ElementA_, layout::ColumnMajor, ElementB_, layout::RowMajor, ElementC_, LayoutC_, arch::OpClassTensorOp, 2, Operator_ > , cutlass::gemm::threadblock::DefaultMmaCore< Shape_, WarpShape_, GemmShape< 8, 8, 4 >, ElementA_, layout::RowMajor, ElementB_, layout::ColumnMajor, ElementC_, LayoutC_, arch::OpClassTensorOp, 2, Operator_ > , cutlass::gemm::threadblock::DefaultMmaCore< Shape_, WarpShape_, GemmShape< 8, 8, 4 >, ElementA_, layout::RowMajor, ElementB_, layout::RowMajor, ElementC_, LayoutC_, arch::OpClassTensorOp, 2, Operator_ > , cutlass::gemm::threadblock::DefaultMmaCore< Shape_, WarpShape_, InstructionShape_, ElementA_, layout::ColumnMajor, ElementB_, layout::ColumnMajor, ElementC_, LayoutC_, arch::OpClassTensorOp, 2, Operator_ > , cutlass::gemm::threadblock::DefaultMmaCore< Shape_, WarpShape_, InstructionShape_, ElementA_, layout::ColumnMajor, ElementB_, layout::RowMajor, ElementC_, LayoutC_, arch::OpClassTensorOp, 2, Operator_ > , cutlass::gemm::threadblock::DefaultMmaCore< Shape_, WarpShape_, InstructionShape_, ElementA_, layout::ColumnMajorInterleaved< InterleavedK >, ElementB_, layout::RowMajorInterleaved< InterleavedK >, ElementC_, LayoutC_, arch::OpClassTensorOp, 2, Operator_, AccumulatorsInRowMajor > , cutlass::gemm::threadblock::DefaultMmaCore< Shape_, WarpShape_, InstructionShape_, ElementA_, layout::RowMajor, ElementB_, layout::ColumnMajor, ElementC_, LayoutC_, arch::OpClassTensorOp, 2, Operator_ > , cutlass::gemm::threadblock::DefaultMmaCore< Shape_, WarpShape_, InstructionShape_, ElementA_, layout::RowMajor, ElementB_, layout::RowMajor, ElementC_, LayoutC_, arch::OpClassTensorOp, 2, Operator_ > , cutlass::gemm::warp::MmaTensorOpAccumulatorTileIterator< Shape_, Element_, cutlass::layout::ColumnMajor, InstructionShape_, OpDelta_ > , cutlass::gemm::warp::MmaTensorOpAccumulatorTileIterator< Shape_, Element_, cutlass::layout::ColumnMajorInterleaved< InterleavedN >, InstructionShape_, OpDelta_ > , cutlass::gemm::warp::MmaTensorOpAccumulatorTileIterator< Shape_, Element_, cutlass::layout::RowMajor, InstructionShape_, OpDelta_ > , cutlass::gemm::warp::MmaTensorOpMultiplicandTileIterator< Shape_, Operand_, Element_, cutlass::layout::ColumnMajorTensorOpMultiplicandCongruous< sizeof_bits< Element_ >::value, int(128/sizeof(Element_))>, InstructionShape_, OpDelta_, 32, PartitionsK_ > , cutlass::gemm::warp::MmaTensorOpMultiplicandTileIterator< Shape_, Operand_, Element_, cutlass::layout::ColumnMajorTensorOpMultiplicandCrosswise< sizeof_bits< Element_ >::value, Crosswise >, InstructionShape_, OpDelta_, 32, PartitionsK_ > , cutlass::gemm::warp::MmaTensorOpMultiplicandTileIterator< Shape_, Operand_, Element_, cutlass::layout::RowMajorTensorOpMultiplicandCongruous< sizeof_bits< Element_ >::value, int(128/sizeof(Element_))>, InstructionShape_, OpDelta_, 32, PartitionsK_ > , cutlass::gemm::warp::MmaTensorOpMultiplicandTileIterator< Shape_, Operand_, Element_, cutlass::layout::RowMajorTensorOpMultiplicandCrosswise< sizeof_bits< Element_ >::value, Crosswise >, InstructionShape_, OpDelta_, 32, PartitionsK_ > , cutlass::gemm::warp::MmaTensorOpMultiplicandTileIterator< Shape_, Operand_, Element_, cutlass::layout::TensorOpMultiplicandCongruous< sizeof_bits< Element_ >::value, 64 >, InstructionShape_, OpDelta_, 32, PartitionsK_ > , cutlass::gemm::warp::MmaTensorOpMultiplicandTileIterator< Shape_, Operand_, Element_, cutlass::layout::TensorOpMultiplicandCrosswise< sizeof_bits< Element_ >::value, Crosswise >, InstructionShape_, OpDelta_, 32, PartitionsK_ > , cutlass::gemm::warp::MmaVoltaTensorOpAccumulatorTileIterator< Shape_, Element_, Layout_, InstructionShape_, OpDelta_ > , cutlass::gemm::warp::MmaVoltaTensorOpMultiplicandTileIterator< Shape_, Operand::kA, Element_, cutlass::layout::ColumnMajorVoltaTensorOpMultiplicandCongruous< sizeof_bits< Element_ >::value >, InstructionShape_, OpDelta_, 32 > , cutlass::gemm::warp::MmaVoltaTensorOpMultiplicandTileIterator< Shape_, Operand::kA, Element_, cutlass::layout::VoltaTensorOpMultiplicandCongruous< sizeof_bits< Element_ >::value >, InstructionShape_, OpDelta_, 32 > , cutlass::gemm::warp::MmaVoltaTensorOpMultiplicandTileIterator< Shape_, Operand::kB, Element_, cutlass::layout::RowMajorVoltaTensorOpMultiplicandBCongruous< sizeof_bits< Element_ >::value >, InstructionShape_, OpDelta_, 32 > , cutlass::gemm::warp::MmaVoltaTensorOpMultiplicandTileIterator< Shape_, Operand::kB, Element_, cutlass::layout::VoltaTensorOpMultiplicandBCongruous< sizeof_bits< Element_ >::value >, InstructionShape_, OpDelta_, 32 > , cutlass::gemm::warp::MmaVoltaTensorOpMultiplicandTileIterator< Shape_, Operand_, Element_, cutlass::layout::ColumnMajorVoltaTensorOpMultiplicandCrosswise< sizeof_bits< Element_ >::value, KBlock >, InstructionShape_, OpDelta_, 32 > , cutlass::gemm::warp::MmaVoltaTensorOpMultiplicandTileIterator< Shape_, Operand_, Element_, cutlass::layout::RowMajorVoltaTensorOpMultiplicandCrosswise< sizeof_bits< Element_ >::value, KBlock >, InstructionShape_, OpDelta_, 32 > , cutlass::gemm::warp::MmaVoltaTensorOpMultiplicandTileIterator< Shape_, Operand_, Element_, cutlass::layout::VoltaTensorOpMultiplicandCrosswise< sizeof_bits< Element_ >::value, KBlock >, InstructionShape_, OpDelta_, 32 > , cutlass::reduction::BatchedReductionTraits< ScalarA_, ScalarC_, ScalarD_, ScalarAlphaBeta_, ScalarAccum_, ReductionSize_, OutputTile_, SubTile_, ThreadShape_, Index_, BlockSwizzle_, maxInReg_, maxOutReg_, Functor_ > , cutlass::transform::PitchLinear2DThreadTileStripminedThreadMap< Shape_, Threads, cutlass::layout::PitchLinearShape< 4, 4 > > , cutlass::transform::PitchLinearStripminedThreadMap< Shape_, Threads, ElementsPerAccess > , cutlass::transform::PitchLinearTilePolicyStripminedThreadContiguous< Shape, Threads, ElementsPerAccess > , cutlass::transform::PitchLinearTilePolicyStripminedThreadStrided< Shape, Threads, ElementsPerAccess > , cutlass::transform::PitchLinearWarpRakedThreadMap< Shape_, Threads, WarpThreadArrangement_, ElementsPerAccess > , cutlass::transform::PitchLinearWarpStripedThreadMap< Shape_, Threads, WarpThreadArrangement_, ElementsPerAccess > , cutlass::transform::TransposePitchLinearThreadMap2DThreadTile< ThreadMap_ > , cutlass::transform::TransposePitchLinearThreadMap< ThreadMap_, WarpThreadArrangement_ > , cutlass::transform::TransposePitchLinearThreadMapSimt< ThreadMap_ >
• kThreadsPerN : cutlass::gemm::threadblock::DefaultGemvCore< Shape_, ThreadShape_, ElementA_, LayoutA_, ElementB_, LayoutB_, ElementC_, LayoutC_ >
• kTile : cutlass::epilogue::threadblock::OutputTileShape< Column, Row, Group, Cluster, Tile > , cutlass::gemm::threadblock::GemmIdentityThreadblockSwizzle
• kTileShapeContiguous : cutlass::layout::TensorOpMultiplicand< ElementSize, Crosswise >
• kTileShapeStride : cutlass::layout::TensorOpMultiplicand< ElementSize, Crosswise >
• kTransformA : cutlass::gemm::device::GemmComplex< ElementA_, LayoutA_, ElementB_, LayoutB_, ElementC_, LayoutC_, ElementAccumulator_, OperatorClass_, ArchTag_, ThreadblockShape_, WarpShape_, InstructionShape_, EpilogueOutputOp_, ThreadblockSwizzle_, Stages, TransformA, TransformB, SplitKSerial > , cutlass::gemm::warp::MmaComplexTensorOp< Shape_, complex< RealElementA >, LayoutA_, complex< RealElementB >, LayoutB_, complex< RealElementC >, LayoutC_, Policy_, TransformA, TransformB, Enable >
• kTransformB : cutlass::gemm::device::GemmComplex< ElementA_, LayoutA_, ElementB_, LayoutB_, ElementC_, LayoutC_, ElementAccumulator_, OperatorClass_, ArchTag_, ThreadblockShape_, WarpShape_, InstructionShape_, EpilogueOutputOp_, ThreadblockSwizzle_, Stages, TransformA, TransformB, SplitKSerial > , cutlass::gemm::warp::MmaComplexTensorOp< Shape_, complex< RealElementA >, LayoutA_, complex< RealElementB >, LayoutB_, complex< RealElementC >, LayoutC_, Policy_, TransformA, TransformB, Enable >
• kValue : cutlass::Max< A, B > , cutlass::Min< A, B >
• kW : cutlass::Tensor4DCoord
• kWarpAccesses : cutlass::epilogue::EpilogueWorkspace< Shape_, WarpCount, FragmentC_ >
• kWarpCount : cutlass::epilogue::EpilogueWorkspace< Shape_, WarpCount, FragmentC_ > , cutlass::epilogue::threadblock::InterleavedOutputTileThreadMap< WarpCount_, MmaCount_, Threads, ElementsPerAccess, ElementSize > , cutlass::epilogue::threadblock::OutputTileOptimalThreadMap< Shape_, Count_, Threads, ElementsPerAccess, ElementSize > , cutlass::transform::PitchLinearWarpRakedThreadMap< Shape_, Threads, WarpThreadArrangement_, ElementsPerAccess >::Detail , cutlass::transform::PitchLinearWarpStripedThreadMap< Shape_, Threads, WarpThreadArrangement_, ElementsPerAccess >::Detail , cutlass::transform::TransposePitchLinearThreadMap< ThreadMap_, WarpThreadArrangement_ >::Detail
• kWarpGemmIterations : cutlass::gemm::threadblock::MmaBase< Shape_, Policy_, Stages, Enable >
• kWarpPartitionsCluster : cutlass::epilogue::threadblock::OutputTileOptimalThreadMap< Shape_, Count_, Threads, ElementsPerAccess, ElementSize >::Detail
• kWarpPartitionsColumn : cutlass::epilogue::threadblock::detail::RowArrangement< Shape, WarpsRemaining, ElementsPerAccess, ElementSize, false > , cutlass::epilogue::threadblock::detail::RowArrangement< Shape, WarpsRemaining, ElementsPerAccess, ElementSize, true >
• kWarpPartitionsGroup : cutlass::epilogue::threadblock::OutputTileOptimalThreadMap< Shape_, Count_, Threads, ElementsPerAccess, ElementSize >::Detail
• kWarpPartitionsRow : cutlass::epilogue::threadblock::detail::RowArrangement< Shape, WarpsRemaining, ElementsPerAccess, ElementSize, false > , cutlass::epilogue::threadblock::detail::RowArrangement< Shape, WarpsRemaining, ElementsPerAccess, ElementSize, true >
• kWarpsContiguous : cutlass::transform::PitchLinearWarpRakedThreadMap< Shape_, Threads, WarpThreadArrangement_, ElementsPerAccess >::Detail , cutlass::transform::PitchLinearWarpStripedThreadMap< Shape_, Threads, WarpThreadArrangement_, ElementsPerAccess >::Detail
• kWarpSize : cutlass::epilogue::EpilogueWorkspace< Shape_, WarpCount, FragmentC_ > , cutlass::epilogue::threadblock::DefaultInterleavedThreadMapTensorOp< ThreadblockShape_, WarpShape_, PartitionsK, Element_, ElementsPerAccess, InterleavedK >::Detail , cutlass::epilogue::threadblock::DefaultThreadMapSimt< ThreadblockShape_, WarpShape_, MmaSimtPolicy_, PartitionsK, Element_, ElementsPerAccess >::Detail , cutlass::epilogue::threadblock::DefaultThreadMapTensorOp< ThreadblockShape_, WarpShape_, PartitionsK, Element_, ElementsPerAccess >::Detail , cutlass::epilogue::threadblock::DefaultThreadMapVoltaTensorOp< ThreadblockShape_, WarpShape_, PartitionsK, ElementOutput_, ElementsPerAccess, float >::Detail , cutlass::epilogue::threadblock::DefaultThreadMapVoltaTensorOp< ThreadblockShape_, WarpShape_, PartitionsK, ElementOutput_, ElementsPerAccess, half_t >::Detail , cutlass::epilogue::threadblock::DefaultThreadMapWmmaTensorOp< ThreadblockShape_, WarpShape_, InstructionShape_, PartitionsK, Element_, ElementsPerAccess >::Detail , cutlass::epilogue::threadblock::detail::RowArrangement< Shape, WarpsRemaining, ElementsPerAccess, ElementSize, false > , cutlass::epilogue::threadblock::detail::RowArrangement< Shape, WarpsRemaining, ElementsPerAccess, ElementSize, true > , cutlass::epilogue::threadblock::InterleavedOutputTileThreadMap< WarpCount_, MmaCount_, Threads, ElementsPerAccess, ElementSize > , cutlass::epilogue::threadblock::OutputTileOptimalThreadMap< Shape_, Count_, Threads, ElementsPerAccess, ElementSize > , cutlass::gemm::threadblock::DefaultMmaCore< Shape_, WarpShape_, GemmShape< 1, 1, 1 >, ElementA_, layout::ColumnMajor, ElementB_, layout::ColumnMajor, ElementC_, LayoutC_, arch::OpClassSimt, 2, Operator_ > , cutlass::gemm::threadblock::DefaultMmaCore< Shape_, WarpShape_, GemmShape< 1, 1, 1 >, ElementA_, layout::ColumnMajor, ElementB_, layout::RowMajor, ElementC_, LayoutC_, arch::OpClassSimt, 2, Operator_ > , cutlass::gemm::threadblock::DefaultMmaCore< Shape_, WarpShape_, GemmShape< 1, 1, 1 >, ElementA_, layout::ColumnMajor, ElementB_, layout::RowMajor, ElementC_, LayoutC_, arch::OpClassSimt, 2, Operator_, > , cutlass::gemm::threadblock::DefaultMmaCore< Shape_, WarpShape_, GemmShape< 1, 1, 1 >, ElementA_, layout::RowMajor, ElementB_, layout::ColumnMajor, ElementC_, LayoutC_, arch::OpClassSimt, 2, Operator_ > , cutlass::gemm::threadblock::DefaultMmaCore< Shape_, WarpShape_, GemmShape< 1, 1, 1 >, ElementA_, layout::RowMajor, ElementB_, layout::RowMajor, ElementC_, LayoutC_, arch::OpClassSimt, 2, Operator_ > , cutlass::gemm::threadblock::DefaultMmaCore< Shape_, WarpShape_, GemmShape< 1, 1, 4 >, int8_t, layout::ColumnMajor, int8_t, layout::ColumnMajor, ElementC_, LayoutC_, arch::OpClassSimt, 2, Operator_ > , cutlass::gemm::threadblock::DefaultMmaCore< Shape_, WarpShape_, GemmShape< 1, 1, 4 >, int8_t, layout::ColumnMajor, int8_t, layout::RowMajor, ElementC_, LayoutC_, arch::OpClassSimt, 2, Operator_ > , cutlass::gemm::threadblock::DefaultMmaCore< Shape_, WarpShape_, GemmShape< 1, 1, 4 >, int8_t, layout::RowMajor, int8_t, layout::ColumnMajor, ElementC_, LayoutC_, arch::OpClassSimt, 2, Operator_ > , cutlass::gemm::threadblock::DefaultMmaCore< Shape_, WarpShape_, GemmShape< 1, 1, 4 >, int8_t, layout::RowMajor, int8_t, layout::RowMajor, ElementC_, LayoutC_, arch::OpClassSimt, 2, Operator_ > , cutlass::gemm::threadblock::DefaultMmaCore< Shape_, WarpShape_, GemmShape< 8, 8, 4 >, ElementA_, layout::ColumnMajor, ElementB_, layout::ColumnMajor, ElementC_, LayoutC_, arch::OpClassTensorOp, 2, Operator_ > , cutlass::gemm::threadblock::DefaultMmaCore< Shape_, WarpShape_, GemmShape< 8, 8, 4 >, ElementA_, layout::ColumnMajor, ElementB_, layout::RowMajor, ElementC_, LayoutC_, arch::OpClassTensorOp, 2, Operator_ > , cutlass::gemm::threadblock::DefaultMmaCore< Shape_, WarpShape_, GemmShape< 8, 8, 4 >, ElementA_, layout::RowMajor, ElementB_, layout::ColumnMajor, ElementC_, LayoutC_, arch::OpClassTensorOp, 2, Operator_ > , cutlass::gemm::threadblock::DefaultMmaCore< Shape_, WarpShape_, GemmShape< 8, 8, 4 >, ElementA_, layout::RowMajor, ElementB_, layout::RowMajor, ElementC_, LayoutC_, arch::OpClassTensorOp, 2, Operator_ > , cutlass::gemm::threadblock::DefaultMmaCore< Shape_, WarpShape_, InstructionShape_, ElementA_, layout::ColumnMajor, ElementB_, layout::ColumnMajor, ElementC_, LayoutC_, arch::OpClassTensorOp, 2, Operator_ > , cutlass::gemm::threadblock::DefaultMmaCore< Shape_, WarpShape_, InstructionShape_, ElementA_, layout::ColumnMajor, ElementB_, layout::RowMajor, ElementC_, LayoutC_, arch::OpClassTensorOp, 2, Operator_ > , cutlass::gemm::threadblock::DefaultMmaCore< Shape_, WarpShape_, InstructionShape_, ElementA_, layout::ColumnMajorInterleaved< InterleavedK >, ElementB_, layout::RowMajorInterleaved< InterleavedK >, ElementC_, LayoutC_, arch::OpClassTensorOp, 2, Operator_, AccumulatorsInRowMajor > , cutlass::gemm::threadblock::DefaultMmaCore< Shape_, WarpShape_, InstructionShape_, ElementA_, layout::RowMajor, ElementB_, layout::ColumnMajor, ElementC_, LayoutC_, arch::OpClassTensorOp, 2, Operator_ > , cutlass::gemm::threadblock::DefaultMmaCore< Shape_, WarpShape_, InstructionShape_, ElementA_, layout::RowMajor, ElementB_, layout::RowMajor, ElementC_, LayoutC_, arch::OpClassTensorOp, 2, Operator_ > , cutlass::transform::PitchLinearWarpRakedThreadMap< Shape_, Threads, WarpThreadArrangement_, ElementsPerAccess >::Detail , cutlass::transform::PitchLinearWarpStripedThreadMap< Shape_, Threads, WarpThreadArrangement_, ElementsPerAccess >::Detail , cutlass::transform::TransposePitchLinearThreadMap< ThreadMap_, WarpThreadArrangement_ >::Detail
• kWarpsRemainingForGroups : cutlass::epilogue::threadblock::OutputTileOptimalThreadMap< Shape_, Count_, Threads, ElementsPerAccess, ElementSize >::Detail
• kWarpsRemainingForRows : cutlass::epilogue::threadblock::OutputTileOptimalThreadMap< Shape_, Count_, Threads, ElementsPerAccess, ElementSize >::Detail
• kWarpsStrided : cutlass::transform::PitchLinearWarpRakedThreadMap< Shape_, Threads, WarpThreadArrangement_, ElementsPerAccess >::Detail , cutlass::transform::PitchLinearWarpStripedThreadMap< Shape_, Threads, WarpThreadArrangement_, ElementsPerAccess >::Detail
• kWarpThreadArrangementContiguous : cutlass::gemm::threadblock::DefaultMmaCore< Shape_, WarpShape_, InstructionShape_, ElementA_, layout::ColumnMajorInterleaved< InterleavedK >, ElementB_, layout::RowMajorInterleaved< InterleavedK >, ElementC_, LayoutC_, arch::OpClassTensorOp, 2, Operator_, AccumulatorsInRowMajor >
• kWarpThreadArrangementContiguousA : cutlass::gemm::threadblock::DefaultMmaCore< Shape_, WarpShape_, InstructionShape_, ElementA_, layout::RowMajor, ElementB_, layout::ColumnMajor, ElementC_, LayoutC_, arch::OpClassTensorOp, 2, Operator_ > , cutlass::gemm::threadblock::DefaultMmaCore< Shape_, WarpShape_, InstructionShape_, ElementA_, layout::RowMajor, ElementB_, layout::RowMajor, ElementC_, LayoutC_, arch::OpClassTensorOp, 2, Operator_ >
• kWarpThreadArrangementContiguousB : cutlass::gemm::threadblock::DefaultMmaCore< Shape_, WarpShape_, InstructionShape_, ElementA_, layout::ColumnMajor, ElementB_, layout::ColumnMajor, ElementC_, LayoutC_, arch::OpClassTensorOp, 2, Operator_ > , cutlass::gemm::threadblock::DefaultMmaCore< Shape_, WarpShape_, InstructionShape_, ElementA_, layout::RowMajor, ElementB_, layout::ColumnMajor, ElementC_, LayoutC_, arch::OpClassTensorOp, 2, Operator_ >
• kWarpThreadArrangementStrided : cutlass::gemm::threadblock::DefaultMmaCore< Shape_, WarpShape_, InstructionShape_, ElementA_, layout::ColumnMajorInterleaved< InterleavedK >, ElementB_, layout::RowMajorInterleaved< InterleavedK >, ElementC_, LayoutC_, arch::OpClassTensorOp, 2, Operator_, AccumulatorsInRowMajor >
• kWarpThreadArrangementStridedA : cutlass::gemm::threadblock::DefaultMmaCore< Shape_, WarpShape_, InstructionShape_, ElementA_, layout::RowMajor, ElementB_, layout::ColumnMajor, ElementC_, LayoutC_, arch::OpClassTensorOp, 2, Operator_ > , cutlass::gemm::threadblock::DefaultMmaCore< Shape_, WarpShape_, InstructionShape_, ElementA_, layout::RowMajor, ElementB_, layout::RowMajor, ElementC_, LayoutC_, arch::OpClassTensorOp, 2, Operator_ >
• kWarpThreadArrangementStridedB : cutlass::gemm::threadblock::DefaultMmaCore< Shape_, WarpShape_, InstructionShape_, ElementA_, layout::ColumnMajor, ElementB_, layout::ColumnMajor, ElementC_, LayoutC_, arch::OpClassTensorOp, 2, Operator_ > , cutlass::gemm::threadblock::DefaultMmaCore< Shape_, WarpShape_, InstructionShape_, ElementA_, layout::RowMajor, ElementB_, layout::ColumnMajor, ElementC_, LayoutC_, arch::OpClassTensorOp, 2, Operator_ >
• kWordCount : cutlass::PredicateVector< kPredicates_, kPredicatesPerByte_, kPredicateStart_ >