CUTLASS: cutlass::layout::RowMajorVoltaTensorOpMultiplicandCrosswise< ElementSize, KBlock > Struct Template Reference

CUTLASS

CUDA Templates for Linear Algebra Subroutines and Solvers

#include <tensor_op_multiplicand_sm70.h>

Public Types
using	Index = int32_t
	Index type used for coordinates. More...

using	LongIndex = int64_t
	Long index type used for offsets. More...

using	TensorCoord = MatrixCoord
	Logical coordinate. More...

using	Stride = Coord< kStrideRank, Index, LongIndex >
	Stride vector. More...

using	Base = VoltaTensorOpMultiplicandCrosswise< ElementSize, KBlock >

Public Member Functions
CUTLASS_HOST_DEVICE	RowMajorVoltaTensorOpMultiplicandCrosswise (Index ldm=0)
	Ctor. More...

CUTLASS_HOST_DEVICE	RowMajorVoltaTensorOpMultiplicandCrosswise (Stride stride)
	Ctor. More...

CUTLASS_HOST_DEVICE LongIndex	operator() (TensorCoord const &coord) const

CUTLASS_HOST_DEVICE TensorCoord	inverse (LongIndex offset) const
	Inverse of layout function, mapping linear offset to logical coordinate. More...

CUTLASS_HOST_DEVICE Stride	stride () const
	Returns the stride of the layout. More...

CUTLASS_HOST_DEVICE Stride &	stride ()
	Returns the stride of the layout. More...

CUTLASS_HOST_DEVICE LongIndex	capacity (TensorCoord const &extent) const

Static Public Member Functions
static CUTLASS_HOST_DEVICE RowMajorVoltaTensorOpMultiplicandCrosswise	packed (TensorCoord const &extent)
	Helper returns a layout to a tightly packed tensor. More...

Static Public Attributes
static int const	kRank = 2
	Logical rank of tensor. More...

static int const	kStrideRank = 1
	Rank of stride vector. More...

static int const	kAccessSize = Base::kAccessSize
	This layout is optimized for 64b accesses. More...

static int const	kElementSize = Base::kElementSize

static int const	kElementsPerAccess = Base::kElementsPerAccess

Detailed Description

template<int ElementSize, int KBlock>
struct cutlass::layout::RowMajorVoltaTensorOpMultiplicandCrosswise< ElementSize, KBlock >

Template mapping a row-major view of pitch-linear memory to TensorOpMultiplicandCrosswise

Member Typedef Documentation

template<int ElementSize, int KBlock>

using cutlass::layout::RowMajorVoltaTensorOpMultiplicandCrosswise< ElementSize, KBlock >::Base = VoltaTensorOpMultiplicandCrosswise<ElementSize, KBlock>

template<int ElementSize, int KBlock>

using cutlass::layout::RowMajorVoltaTensorOpMultiplicandCrosswise< ElementSize, KBlock >::Index = int32_t

template<int ElementSize, int KBlock>

using cutlass::layout::RowMajorVoltaTensorOpMultiplicandCrosswise< ElementSize, KBlock >::LongIndex = int64_t

template<int ElementSize, int KBlock>

using cutlass::layout::RowMajorVoltaTensorOpMultiplicandCrosswise< ElementSize, KBlock >::Stride = Coord<kStrideRank, Index, LongIndex>

template<int ElementSize, int KBlock>

using cutlass::layout::RowMajorVoltaTensorOpMultiplicandCrosswise< ElementSize, KBlock >::TensorCoord = MatrixCoord

Constructor & Destructor Documentation

template<int ElementSize, int KBlock>

CUTLASS_HOST_DEVICE cutlass::layout::RowMajorVoltaTensorOpMultiplicandCrosswise< ElementSize, KBlock >::RowMajorVoltaTensorOpMultiplicandCrosswise ( Index ldm = 0 )

inline

template<int ElementSize, int KBlock>

CUTLASS_HOST_DEVICE cutlass::layout::RowMajorVoltaTensorOpMultiplicandCrosswise< ElementSize, KBlock >::RowMajorVoltaTensorOpMultiplicandCrosswise ( Stride stride )

inline

Member Function Documentation

template<int ElementSize, int KBlock>

CUTLASS_HOST_DEVICE LongIndex cutlass::layout::RowMajorVoltaTensorOpMultiplicandCrosswise< ElementSize, KBlock >::capacity ( TensorCoord const & extent ) const

inline

Compute the number of contiguous elements needed to store a tensor with the given size

template<int ElementSize, int KBlock>

CUTLASS_HOST_DEVICE TensorCoord cutlass::layout::RowMajorVoltaTensorOpMultiplicandCrosswise< ElementSize, KBlock >::inverse ( LongIndex offset ) const

inline

template<int ElementSize, int KBlock>

CUTLASS_HOST_DEVICE LongIndex cutlass::layout::RowMajorVoltaTensorOpMultiplicandCrosswise< ElementSize, KBlock >::operator() ( TensorCoord const & coord ) const

inline

Returns the offset of a coordinate in linear memory. Assumes coordinate has convention (contiguous, strided)

template<int ElementSize, int KBlock>

static CUTLASS_HOST_DEVICE RowMajorVoltaTensorOpMultiplicandCrosswise cutlass::layout::RowMajorVoltaTensorOpMultiplicandCrosswise< ElementSize, KBlock >::packed ( TensorCoord const & extent )

inlinestatic

template<int ElementSize, int KBlock>

CUTLASS_HOST_DEVICE Stride cutlass::layout::RowMajorVoltaTensorOpMultiplicandCrosswise< ElementSize, KBlock >::stride ( ) const

inline

template<int ElementSize, int KBlock>

CUTLASS_HOST_DEVICE Stride& cutlass::layout::RowMajorVoltaTensorOpMultiplicandCrosswise< ElementSize, KBlock >::stride ( )

inline

Member Data Documentation

template<int ElementSize, int KBlock>

int const cutlass::layout::RowMajorVoltaTensorOpMultiplicandCrosswise< ElementSize, KBlock >::kAccessSize = Base::kAccessSize

static

template<int ElementSize, int KBlock>

int const cutlass::layout::RowMajorVoltaTensorOpMultiplicandCrosswise< ElementSize, KBlock >::kElementSize = Base::kElementSize

static

template<int ElementSize, int KBlock>

int const cutlass::layout::RowMajorVoltaTensorOpMultiplicandCrosswise< ElementSize, KBlock >::kElementsPerAccess = Base::kElementsPerAccess

static

template<int ElementSize, int KBlock>

int const cutlass::layout::RowMajorVoltaTensorOpMultiplicandCrosswise< ElementSize, KBlock >::kRank = 2

static

template<int ElementSize, int KBlock>

int const cutlass::layout::RowMajorVoltaTensorOpMultiplicandCrosswise< ElementSize, KBlock >::kStrideRank = 1

static

The documentation for this struct was generated from the following file:

tensor_op_multiplicand_sm70.h