CUTLASS
CUDA Templates for Linear Algebra Subroutines and Solvers
Public Types | Public Member Functions | List of all members
cutlass::arch::Mma< gemm::GemmShape< 8, 8, 4 >, 8, half_t, layout::ColumnMajor, half_t, layout::RowMajor, half_t, layout::RowMajor, OpMultiplyAdd > Struct Template Reference

Matrix multiply-add operation: F16 = F16 * F16 + F16.

#include <mma_sm70.h>

Public Types

using Shape = gemm::GemmShape< 8, 8, 4 >
 
using ElementA = half_t
 
using LayoutA = layout::ColumnMajor
 
using FragmentA = Array< half_t, 4 >
 
using ElementB = half_t
 
using LayoutB = layout::RowMajor
 
using FragmentB = Array< half_t, 4 >
 
using ElementC = half_t
 
using LayoutC = layout::RowMajor
 
using FragmentC = Array< half_t, 8 >
 
using Operator = OpMultiplyAdd
 

Public Member Functions

CUTLASS_HOST_DEVICE void operator() (FragmentC &d, FragmentA const &a, FragmentB const &b, FragmentC const &c)
 

Member Typedef Documentation

Member Function Documentation

CUTLASS_HOST_DEVICE void cutlass::arch::Mma< gemm::GemmShape< 8, 8, 4 >, 8, half_t, layout::ColumnMajor, half_t, layout::RowMajor, half_t, layout::RowMajor, OpMultiplyAdd >::operator() ( FragmentC d,
FragmentA const &  a,
FragmentB const &  b,
FragmentC const &  c 
)
inline

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