CUTLASS
CUDA Templates for Linear Algebra Subroutines and Solvers
|
Mapping function for 4-D NHWC tensors.
#include <tensor.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 = Tensor4DCoord |
Logical coordinate (n, h, w, c) More... | |
using | Stride = Coord< kStrideRank > |
Stride vector. More... | |
Public Member Functions | |
CUTLASS_HOST_DEVICE | TensorNHWC (Stride const &stride=Stride(0)) |
Constructor. More... | |
CUTLASS_HOST_DEVICE | TensorNHWC (typename Stride::Index c, typename Stride::Index wc, typename Stride::Index hwc) |
Constructor. More... | |
CUTLASS_HOST_DEVICE LongIndex | operator() (TensorCoord const &coord) const |
Returns the offset of a coordinate (n, h, w, c) in linear memory. More... | |
CUTLASS_HOST_DEVICE | operator RowMajor () |
Returns a RowMajor equivalent for a TensorNHWC layout. More... | |
CUTLASS_HOST_DEVICE TensorCoord | inverse (LongIndex index) const |
Returns the logical coordinate (n, h, w, c) from a given offset in linear memory. 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 |
Compute the number of contiguous elements needed to store a tensor with the given size. More... | |
Static Public Member Functions | |
static CUTLASS_HOST_DEVICE TensorNHWC | packed (TensorCoord const &extent) |
Helper returns a layout to a tightly packed NHWC tensor. More... | |
Static Public Attributes | |
static int const | kRank = 4 |
Logical rank of tensor. More... | |
static int const | kStrideRank = 3 |
Rank of stride vector. More... | |
using cutlass::layout::TensorNHWC::Index = int32_t |
using cutlass::layout::TensorNHWC::LongIndex = int64_t |
|
inline |
|
inline |
|
inline |
|
inline |
|
inlineexplicit |
|
inline |
|
inlinestatic |
|
inline |
|
inline |
|
static |
|
static |