Partial specialization for Turing Architecture.
#include <default_gemm.h>
|
using | Mma = typename cutlass::gemm::threadblock::DefaultMma< ElementA, LayoutA, kAlignmentA, ElementB, LayoutB, kAlignmentB, ElementAccumulator, layout::RowMajor, arch::OpClassTensorOp, arch::Sm75, ThreadblockShape, WarpShape, InstructionShape, 2, Operator >::ThreadblockMma |
| Define the threadblock-scoped matrix multiply-accumulate. More...
|
|
using | Epilogue = typename cutlass::epilogue::threadblock::DefaultEpilogueTensorOp< ThreadblockShape, typename Mma::Operator, kPartitionsK, EpilogueOutputOp, EpilogueOutputOp::kCount >::Epilogue |
| Define the epilogue. More...
|
|
using | GemmKernel = kernel::Gemm< Mma, Epilogue, ThreadblockSwizzle, SplitKSerial > |
| Define the kernel-level GEMM operator. More...
|
|
|
static const int | kPartitionsK = ThreadblockShape::kK / WarpShape::kK |
|
template<typename ElementA , typename LayoutA , int kAlignmentA, typename ElementB , typename LayoutB , int kAlignmentB, typename ElementC , typename ElementAccumulator , typename ThreadblockShape , typename WarpShape , typename InstructionShape , typename EpilogueOutputOp , typename ThreadblockSwizzle , bool SplitKSerial, typename Operator >
using 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 >::Epilogue = typename cutlass::epilogue::threadblock::DefaultEpilogueTensorOp< ThreadblockShape, typename Mma::Operator, kPartitionsK, EpilogueOutputOp, EpilogueOutputOp::kCount >::Epilogue |
template<typename ElementA , typename LayoutA , int kAlignmentA, typename ElementB , typename LayoutB , int kAlignmentB, typename ElementC , typename ElementAccumulator , typename ThreadblockShape , typename WarpShape , typename InstructionShape , typename EpilogueOutputOp , typename ThreadblockSwizzle , bool SplitKSerial, typename Operator >
using 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 >::GemmKernel = kernel::Gemm<Mma, Epilogue, ThreadblockSwizzle, SplitKSerial> |
template<typename ElementA , typename LayoutA , int kAlignmentA, typename ElementB , typename LayoutB , int kAlignmentB, typename ElementC , typename ElementAccumulator , typename ThreadblockShape , typename WarpShape , typename InstructionShape , typename EpilogueOutputOp , typename ThreadblockSwizzle , bool SplitKSerial, typename Operator >
using 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 >::Mma = typename cutlass::gemm::threadblock::DefaultMma< ElementA, LayoutA, kAlignmentA, ElementB, LayoutB, kAlignmentB, ElementAccumulator, layout::RowMajor, arch::OpClassTensorOp, arch::Sm75, ThreadblockShape, WarpShape, InstructionShape, 2, Operator >::ThreadblockMma |
template<typename ElementA , typename LayoutA , int kAlignmentA, typename ElementB , typename LayoutB , int kAlignmentB, typename ElementC , typename ElementAccumulator , typename ThreadblockShape , typename WarpShape , typename InstructionShape , typename EpilogueOutputOp , typename ThreadblockSwizzle , bool SplitKSerial, typename Operator >
const int 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 >::kPartitionsK = ThreadblockShape::kK / WarpShape::kK |
|
static |
The documentation for this struct was generated from the following file: