CUTLASS
CUDA Templates for Linear Algebra Subroutines and Solvers
Classes | Public Types | Static Public Attributes | List of all members
cutlass::epilogue::threadblock::DefaultThreadMapSimt< ThreadblockShape_, WarpShape_, MmaSimtPolicy_, PartitionsK, Element_, ElementsPerAccess > Struct Template Reference

Defines the optimal thread map for SIMT accumulator layouts.

#include <default_thread_map_simt.h>

Classes

struct  Detail
 

Public Types

using ThreadblockShape = ThreadblockShape_
 
using WarpShape = WarpShape_
 
using MmaSimtPolicy = MmaSimtPolicy_
 
using Element = Element_
 
using Type = OutputTileOptimalThreadMap< OutputTileShape< ThreadblockShape::kN, 1, MmaSimtPolicy::WarpShape::kRow, Detail::WarpCount::kM, 1 >, OutputTileShape< 1, MmaSimtPolicy::LaneMmaShape::kM, Detail::kGroupCount, 1, Detail::kIterations >, Detail::kThreads, kElementsPerAccess, sizeof_bits< Element >::value >
 ThreadMap to be used by epilogue::PredicatedTileIterator satisfying concept OutputTileThreadMap. More...
 

Static Public Attributes

static int const kPartitionsK = PartitionsK
 
static int const kElementsPerAccess = ElementsPerAccess
 

Member Typedef Documentation

template<typename ThreadblockShape_ , typename WarpShape_ , typename MmaSimtPolicy_ , int PartitionsK, typename Element_ , int ElementsPerAccess>
using cutlass::epilogue::threadblock::DefaultThreadMapSimt< ThreadblockShape_, WarpShape_, MmaSimtPolicy_, PartitionsK, Element_, ElementsPerAccess >::Element = Element_
template<typename ThreadblockShape_ , typename WarpShape_ , typename MmaSimtPolicy_ , int PartitionsK, typename Element_ , int ElementsPerAccess>
using cutlass::epilogue::threadblock::DefaultThreadMapSimt< ThreadblockShape_, WarpShape_, MmaSimtPolicy_, PartitionsK, Element_, ElementsPerAccess >::MmaSimtPolicy = MmaSimtPolicy_
template<typename ThreadblockShape_ , typename WarpShape_ , typename MmaSimtPolicy_ , int PartitionsK, typename Element_ , int ElementsPerAccess>
using cutlass::epilogue::threadblock::DefaultThreadMapSimt< ThreadblockShape_, WarpShape_, MmaSimtPolicy_, PartitionsK, Element_, ElementsPerAccess >::ThreadblockShape = ThreadblockShape_
template<typename ThreadblockShape_ , typename WarpShape_ , typename MmaSimtPolicy_ , int PartitionsK, typename Element_ , int ElementsPerAccess>
using cutlass::epilogue::threadblock::DefaultThreadMapSimt< ThreadblockShape_, WarpShape_, MmaSimtPolicy_, PartitionsK, Element_, ElementsPerAccess >::Type = OutputTileOptimalThreadMap< OutputTileShape< ThreadblockShape::kN, 1, MmaSimtPolicy::WarpShape::kRow, Detail::WarpCount::kM, 1>, OutputTileShape< 1, MmaSimtPolicy::LaneMmaShape::kM, Detail::kGroupCount, 1, Detail::kIterations>, Detail::kThreads, kElementsPerAccess, sizeof_bits<Element>::value >
template<typename ThreadblockShape_ , typename WarpShape_ , typename MmaSimtPolicy_ , int PartitionsK, typename Element_ , int ElementsPerAccess>
using cutlass::epilogue::threadblock::DefaultThreadMapSimt< ThreadblockShape_, WarpShape_, MmaSimtPolicy_, PartitionsK, Element_, ElementsPerAccess >::WarpShape = WarpShape_

Member Data Documentation

template<typename ThreadblockShape_ , typename WarpShape_ , typename MmaSimtPolicy_ , int PartitionsK, typename Element_ , int ElementsPerAccess>
int const cutlass::epilogue::threadblock::DefaultThreadMapSimt< ThreadblockShape_, WarpShape_, MmaSimtPolicy_, PartitionsK, Element_, ElementsPerAccess >::kElementsPerAccess = ElementsPerAccess
static
template<typename ThreadblockShape_ , typename WarpShape_ , typename MmaSimtPolicy_ , int PartitionsK, typename Element_ , int ElementsPerAccess>
int const cutlass::epilogue::threadblock::DefaultThreadMapSimt< ThreadblockShape_, WarpShape_, MmaSimtPolicy_, PartitionsK, Element_, ElementsPerAccess >::kPartitionsK = PartitionsK
static

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