CUTLASS: cutlass::layout::TensorOpMultiplicandRowMajorInterleaved< ElementSize, InterleavedK > Struct Template Reference

CUTLASS

CUDA Templates for Linear Algebra Subroutines and Solvers

Template based on element size (in bits) - defined in terms of pitch-linear memory.

#include <tensor_op_multiplicand_sm75.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 = PitchLinearCoord
	Logical coordinate. More...

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

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

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

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

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

Static Public Member Functions
static CUTLASS_HOST_DEVICE TensorOpMultiplicandRowMajorInterleaved	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 = 128
	This layout is optimized for 128b accesses. More...

static int const	kElementSize = ElementSize

static int const	kElementsPerAccess = kAccessSize / kElementSize

static int const	kInterleavedK = InterleavedK

Member Typedef Documentation

template<int ElementSize, int InterleavedK>

using cutlass::layout::TensorOpMultiplicandRowMajorInterleaved< ElementSize, InterleavedK >::Index = int32_t

template<int ElementSize, int InterleavedK>

using cutlass::layout::TensorOpMultiplicandRowMajorInterleaved< ElementSize, InterleavedK >::LongIndex = int64_t

template<int ElementSize, int InterleavedK>

using cutlass::layout::TensorOpMultiplicandRowMajorInterleaved< ElementSize, InterleavedK >::Stride = Coord<kStrideRank, Index, LongIndex>

template<int ElementSize, int InterleavedK>

using cutlass::layout::TensorOpMultiplicandRowMajorInterleaved< ElementSize, InterleavedK >::TensorCoord = PitchLinearCoord

Constructor & Destructor Documentation

template<int ElementSize, int InterleavedK>

CUTLASS_HOST_DEVICE cutlass::layout::TensorOpMultiplicandRowMajorInterleaved< ElementSize, InterleavedK >::TensorOpMultiplicandRowMajorInterleaved ( Index ldm = 0 )

inline

template<int ElementSize, int InterleavedK>

CUTLASS_HOST_DEVICE cutlass::layout::TensorOpMultiplicandRowMajorInterleaved< ElementSize, InterleavedK >::TensorOpMultiplicandRowMajorInterleaved ( Stride stride )

inline

Member Function Documentation

template<int ElementSize, int InterleavedK>

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

inline

template<int ElementSize, int InterleavedK>

CUTLASS_HOST_DEVICE LongIndex cutlass::layout::TensorOpMultiplicandRowMajorInterleaved< ElementSize, InterleavedK >::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 InterleavedK>

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

inlinestatic

template<int ElementSize, int InterleavedK>

CUTLASS_HOST_DEVICE Stride cutlass::layout::TensorOpMultiplicandRowMajorInterleaved< ElementSize, InterleavedK >::stride ( ) const

inline

template<int ElementSize, int InterleavedK>

CUTLASS_HOST_DEVICE Stride& cutlass::layout::TensorOpMultiplicandRowMajorInterleaved< ElementSize, InterleavedK >::stride ( )

inline

Member Data Documentation

template<int ElementSize, int InterleavedK>

int const cutlass::layout::TensorOpMultiplicandRowMajorInterleaved< ElementSize, InterleavedK >::kAccessSize = 128

static

template<int ElementSize, int InterleavedK>

int const cutlass::layout::TensorOpMultiplicandRowMajorInterleaved< ElementSize, InterleavedK >::kElementSize = ElementSize

static

template<int ElementSize, int InterleavedK>

int const cutlass::layout::TensorOpMultiplicandRowMajorInterleaved< ElementSize, InterleavedK >::kElementsPerAccess = kAccessSize / kElementSize

static

template<int ElementSize, int InterleavedK>

int const cutlass::layout::TensorOpMultiplicandRowMajorInterleaved< ElementSize, InterleavedK >::kInterleavedK = InterleavedK

static

template<int ElementSize, int InterleavedK>

int const cutlass::layout::TensorOpMultiplicandRowMajorInterleaved< ElementSize, InterleavedK >::kRank = 2

static

template<int ElementSize, int InterleavedK>

int const cutlass::layout::TensorOpMultiplicandRowMajorInterleaved< ElementSize, InterleavedK >::kStrideRank = 1

static

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

tensor_op_multiplicand_sm75.h