default_mma_core_wmma.h

Go to the documentation of this file.

/***************************************************************************************************
 * Copyright (c) 2017-2019, NVIDIA CORPORATION.  All rights reserved.
 *
 * Redistribution and use in source and binary forms, with or without modification, are permitted
 * provided that the following conditions are met:
 *     * Redistributions of source code must retain the above copyright notice, this list of
 *       conditions and the following disclaimer.
 *     * Redistributions in binary form must reproduce the above copyright notice, this list of
 *       conditions and the following disclaimer in the documentation and/or other materials
 *       provided with the distribution.
 *     * Neither the name of the NVIDIA CORPORATION nor the names of its contributors may be used
 *       to endorse or promote products derived from this software without specific prior written
 *       permission.
 *
 * THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS "AS IS" AND ANY EXPRESS OR
 * IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND
 * FITNESS FOR A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL NVIDIA CORPORATION BE LIABLE
 * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING,
 * BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS;
 * OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT,
 * STRICT LIABILITY, OR TOR (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE
 * OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
 *
 **************************************************************************************************/
#pragma once

#include "cutlass/cutlass.h"
#include "cutlass/array.h"
#include "cutlass/fast_math.h"
#include "cutlass/arch/wmma.h"

#if defined(CUTLASS_ARCH_WMMA_ENABLED)

#include "cutlass/numeric_types.h"
#include "cutlass/matrix_shape.h"

#include "cutlass/transform/threadblock/regular_tile_iterator_pitch_linear.h"

#include "cutlass/gemm/warp/mma_tensor_op_wmma.h"

#include "cutlass/gemm/warp/mma_tensor_op_policy.h"
#include "cutlass/gemm/threadblock/default_mma_core.h"


namespace cutlass {
namespace gemm {
namespace threadblock {


template <
    typename Shape_,
    typename WarpShape_,
    typename InstructionShape_,
    typename ElementA_,
    typename ElementB_,
    typename ElementC_,
    typename LayoutC_,
    typename Operator_,
    int Stages>
struct DefaultMmaCore<Shape_, WarpShape_, InstructionShape_, ElementA_,
                      layout::ColumnMajor, ElementB_, layout::RowMajor,
                      ElementC_, LayoutC_, arch::OpClassWmmaTensorOp, Stages,
                      Operator_> {
  using Shape = Shape_;
  using WarpShape = WarpShape_;
  using InstructionShape = InstructionShape_;
  using ElementA = ElementA_;
  using LayoutA = layout::ColumnMajor;
  using ElementB = ElementB_;
  using LayoutB = layout::RowMajor;
  using ElementC = ElementC_;
  using LayoutC = LayoutC_;
  using OperatorClass = arch::OpClassWmmaTensorOp;

  using WarpCount = GemmShape<
    Shape::kM / WarpShape::kM,
    Shape::kN / WarpShape::kN,
    Shape::kK / WarpShape::kK
  >;

  // Divisility requirements
  static_assert(
    !(Shape::kM % WarpShape::kM) &&
    !(Shape::kN % WarpShape::kN),
    "Threadblock-scoped GEMM should be divisible by warp-scoped GEMM size."
  );

  static int const kWarpSize = warp::WarpSize<arch::OpClassWmmaTensorOp>::value;

  static int const kThreads = WarpCount::kCount * kWarpSize;

  static int const kAccessSizeInBits = 128;

  using Operator = Operator_;

  //
  // Shared memory layouts
  //
  // NOTE: shared memory layout for wmma is same as the operands' layout in the global memory
  using SmemLayoutA = LayoutA;
  using SmemLayoutB = LayoutB;

  // Pad shared memory to avoid bank conflicts
  static int const kPaddingA = 128 / sizeof_bits<ElementA>::value;
  static int const kPaddingB = 128 / sizeof_bits<ElementB>::value;

  //
  // Iterators to write to shared memory
  //
  
  using IteratorThreadMapA = transform::PitchLinearStripminedThreadMap<
    layout::PitchLinearShape<Shape::kM, Shape::kK>,
    kThreads,
    kAccessSizeInBits / sizeof_bits<ElementB>::value
  >;

  using SmemIteratorA = transform::threadblock::RegularTileIterator<
    MatrixShape<Shape::kM, Shape::kK>, 
    ElementA, 
    SmemLayoutA,
    1,
    IteratorThreadMapA
  >;

  using IteratorThreadMapB = transform::PitchLinearStripminedThreadMap<
    layout::PitchLinearShape<Shape::kN, Shape::kK>,
    kThreads,
    kAccessSizeInBits / sizeof_bits<ElementB>::value
  >;

  using SmemIteratorB = transform::threadblock::RegularTileIterator<
    MatrixShape<Shape::kK, Shape::kN>, 
    ElementB, 
    SmemLayoutB,
    0,
    IteratorThreadMapB
  >;

  //
  // Warp-level matrix multiply operator
  //

  // Define the warp-level tensor op
  using Policy = cutlass::gemm::warp::MmaTensorOpPolicy<
    cutlass::arch::Wmma<
      InstructionShape,
      ElementA,
      LayoutA,
      ElementB,
      LayoutB,
      ElementC,
      LayoutC,
      Operator
    >,
    cutlass::MatrixShape<1, 1>
  >;

  using MmaTensorOp = cutlass::gemm::warp::MmaTensorOpWmma<
    WarpShape,
    ElementA,
    SmemLayoutA,
    ElementB,
    SmemLayoutB,
    ElementC,
    LayoutC,
    Policy
  >;

  using MmaPolicy = MmaPolicy<
    MmaTensorOp,
    MatrixShape<kPaddingA, 0>,
    MatrixShape<0, kPaddingB>,
    WarpCount::kK
  >;
};


template <
    typename Shape_,
    typename WarpShape_,
    typename InstructionShape_,
    typename ElementA_,
    typename ElementB_,
    typename ElementC_,
    typename LayoutC_,
    typename Operator_,
    int Stages>
struct DefaultMmaCore<Shape_, WarpShape_, InstructionShape_, ElementA_,
                      layout::RowMajor, ElementB_, layout::ColumnMajor,
                      ElementC_, LayoutC_, arch::OpClassWmmaTensorOp, Stages,
                      Operator_> {
  using Shape = Shape_;
  using WarpShape = WarpShape_;
  using InstructionShape = InstructionShape_;
  using ElementA = ElementA_;
  using LayoutA = layout::RowMajor;
  using ElementB = ElementB_;
  using LayoutB = layout::ColumnMajor;
  using ElementC = ElementC_;
  using LayoutC = LayoutC_;
  using OperatorClass = arch::OpClassWmmaTensorOp;

  using WarpCount = GemmShape<
    Shape::kM / WarpShape::kM,
    Shape::kN / WarpShape::kN,
    Shape::kK / WarpShape::kK
  >;

  // Divisility requirements
  static_assert(
    !(Shape::kM % WarpShape::kM) &&
    !(Shape::kN % WarpShape::kN),
    "Threadblock-scoped GEMM should be divisible by warp-scoped GEMM size."
  );

  static int const kWarpSize = warp::WarpSize<arch::OpClassWmmaTensorOp>::value;

  static int const kThreads = WarpCount::kCount * kWarpSize;


  static int const kAccessSizeInBits = 128;

  using Operator = Operator_;

  // Warp thread arrangement 
  static int const kWarpThreadArrangementContiguousA =
      Shape::kK / (kAccessSizeInBits / sizeof_bits<ElementA>::value);

  static int const kWarpThreadArrangementStridedA =
      kWarpSize / kWarpThreadArrangementContiguousA;

  static int const kWarpThreadArrangementContiguousB =
      Shape::kK / (kAccessSizeInBits / sizeof_bits<ElementA>::value);

  static int const kWarpThreadArrangementStridedB =
      kWarpSize / kWarpThreadArrangementContiguousB;

  //
  // Shared memory layouts
  //

  // shared memory layout for wmma is same as the operands' layout in global memory
  using SmemLayoutA = LayoutA;
  using SmemLayoutB = LayoutB;
  
  // Pad shared memory to avoid bank conflicts
  static int const kPaddingA = 128 / sizeof_bits<ElementA>::value;
  static int const kPaddingB = 128 / sizeof_bits<ElementB>::value;

  //
  // Iterators to write to shared memory 
  //
  using IteratorThreadMapA = transform::PitchLinearStripminedThreadMap<
    layout::PitchLinearShape<Shape::kK, Shape::kM>,
    kThreads,
    kAccessSizeInBits / sizeof_bits<ElementA>::value
  >;

  using SmemIteratorA = transform::threadblock::RegularTileIterator<
    MatrixShape<Shape::kM, Shape::kK>, 
    ElementA, 
    SmemLayoutA,
    1,
    IteratorThreadMapA 
  >;

  using IteratorThreadMapB = transform::PitchLinearStripminedThreadMap<
    layout::PitchLinearShape<Shape::kK, Shape::kN>,
    kThreads,
    kAccessSizeInBits / sizeof_bits<ElementB>::value
  >;  

  using SmemIteratorB = transform::threadblock::RegularTileIterator<
    MatrixShape<Shape::kK, Shape::kN>, 
    ElementB, 
    SmemLayoutB,
    0,
    IteratorThreadMapB // SmemThreadMapB 
  >;

  //
  // Warp-level matrix multiply operator
  //

  // Define the warp-level tensor op
  using Policy = cutlass::gemm::warp::MmaTensorOpPolicy<
    cutlass::arch::Wmma<
      InstructionShape,
      ElementA,
      LayoutA,
      ElementB,
      LayoutB,
      ElementC,
      LayoutC,
      Operator
    >,
    cutlass::MatrixShape<1, 1>
  >;

  using MmaTensorOp = cutlass::gemm::warp::MmaTensorOpWmma<
    WarpShape,
    ElementA,
    SmemLayoutA,
    ElementB,
    SmemLayoutB,
    ElementC,
    LayoutC,
    Policy
  >;

  using MmaPolicy = MmaPolicy<
    MmaTensorOp,
    MatrixShape<0, kPaddingA>,
    MatrixShape<kPaddingB, 0>,
    WarpCount::kK
  >;
};



template <
    typename Shape_,
    typename WarpShape_,
    typename InstructionShape_,
    typename ElementA_,
    typename ElementB_,
    typename ElementC_,
    typename LayoutC_,
    typename Operator_,
    int Stages>
struct DefaultMmaCore<Shape_, WarpShape_, InstructionShape_, ElementA_,
                      layout::RowMajor, ElementB_, layout::RowMajor, ElementC_,
                      LayoutC_, arch::OpClassWmmaTensorOp, Stages, Operator_> {
  using Shape = Shape_;
  using WarpShape = WarpShape_;
  using InstructionShape = InstructionShape_;
  using ElementA = ElementA_;
  using LayoutA = layout::RowMajor;
  using ElementB = ElementB_;
  using LayoutB = layout::RowMajor;
  using ElementC = ElementC_;
  using LayoutC = LayoutC_;
  using OperatorClass = arch::OpClassWmmaTensorOp;

  using WarpCount = GemmShape<
    Shape::kM / WarpShape::kM,
    Shape::kN / WarpShape::kN,
    Shape::kK / WarpShape::kK
  >;

  // Divisility requirements
  static_assert(
    !(Shape::kM % WarpShape::kM) &&
    !(Shape::kN % WarpShape::kN),
    "Threadblock-scoped GEMM should be divisible by warp-scoped GEMM size."
  );

  static int const kWarpSize = warp::WarpSize<arch::OpClassWmmaTensorOp>::value;

  static int const kThreads = WarpCount::kCount * kWarpSize;

  static int const kAccessSizeInBits = 128;

  using Operator = Operator_;

  // Warp thread arrangement 
  static int const kWarpThreadArrangementContiguousA =
      Shape::kK / (kAccessSizeInBits / sizeof_bits<ElementA>::value);

  static int const kWarpThreadArrangementStridedA =
      kWarpSize / kWarpThreadArrangementContiguousA;

  //
  // Shared memory layouts
  //

  // shared memory layout for wmma is same as the operands' layout in global memory
  using SmemLayoutA = LayoutA;
  using SmemLayoutB = LayoutB;

  // Pad shared memory to avoid bank conflicts
  static int const kPaddingA = 128 / sizeof_bits<ElementA>::value;
  static int const kPaddingB = 128 / sizeof_bits<ElementB>::value;
  
  //
  // Iterators to write to shared memory
  //

  using IteratorThreadMapA = transform::PitchLinearStripminedThreadMap<
    layout::PitchLinearShape<Shape::kK, Shape::kM>,
    kThreads,
    kAccessSizeInBits / sizeof_bits<ElementA>::value
  >;


  using SmemIteratorA = transform::threadblock::RegularTileIterator<
    MatrixShape<Shape::kM, Shape::kK>, 
    ElementA, 
    SmemLayoutA,
    1,
    IteratorThreadMapA
  >;

  using IteratorThreadMapB = transform::PitchLinearStripminedThreadMap<
    layout::PitchLinearShape<Shape::kN, Shape::kK>,
    kThreads,
    kAccessSizeInBits / sizeof_bits<ElementB>::value
  >;

  using SmemIteratorB = transform::threadblock::RegularTileIterator<
    MatrixShape<Shape::kK, Shape::kN>, 
    ElementB, 
    SmemLayoutB,
    0,
    IteratorThreadMapB
  >;

  //
  // Warp-level matrix multiply operator
  //

  // Define the warp-level tensor op
  using Policy = cutlass::gemm::warp::MmaTensorOpPolicy<
    cutlass::arch::Wmma<
      InstructionShape,
      ElementA,
      LayoutA,
      ElementB,
      LayoutB,
      ElementC,
      LayoutC,
      Operator
    >,
    cutlass::MatrixShape<1, 1>
  >;

  using MmaTensorOp = cutlass::gemm::warp::MmaTensorOpWmma<
    WarpShape,
    ElementA,
    SmemLayoutA,
    ElementB,
    SmemLayoutB,
    ElementC,
    LayoutC,
    Policy
  >;

  using MmaPolicy = MmaPolicy<
    MmaTensorOp,
    MatrixShape<0, kPaddingA>,
    MatrixShape<0, kPaddingB>,
    WarpCount::kK
  >;
};


template <
    typename Shape_,
    typename WarpShape_,
    typename InstructionShape_,
    typename ElementA_,
    typename ElementB_,
    typename ElementC_,
    typename LayoutC_,
    typename Operator_,
    int Stages>
struct DefaultMmaCore<Shape_, WarpShape_, InstructionShape_, ElementA_,
                      layout::ColumnMajor, ElementB_, layout::ColumnMajor,
                      ElementC_, LayoutC_, arch::OpClassWmmaTensorOp, Stages,
                      Operator_> {
  using Shape = Shape_;
  using WarpShape = WarpShape_;
  using InstructionShape = InstructionShape_;
  using ElementA = ElementA_;
  using LayoutA = layout::ColumnMajor;
  using ElementB = ElementB_;
  using LayoutB = layout::ColumnMajor;
  using ElementC = ElementC_;
  using LayoutC = LayoutC_;
  using OperatorClass = arch::OpClassWmmaTensorOp;

  using WarpCount =
      GemmShape<Shape::kM / WarpShape::kM, Shape::kN / WarpShape::kN,
                Shape::kK / WarpShape::kK>;

  // Divisility requirements
  static_assert(
      !(Shape::kM % WarpShape::kM) && !(Shape::kN % WarpShape::kN),
      "Threadblock-scoped GEMM should be divisible by warp-scoped GEMM size.");

  static int const kWarpSize = warp::WarpSize<arch::OpClassWmmaTensorOp>::value;

  static int const kThreads = WarpCount::kCount * kWarpSize;

  static int const kAccessSizeInBits = 128;

  using Operator = Operator_; 

  // Warp thread arrangement 
  static int const kWarpThreadArrangementContiguousB =
      Shape::kK / (kAccessSizeInBits / sizeof_bits<ElementA>::value);

  static int const kWarpThreadArrangementStridedB =
      kWarpSize / kWarpThreadArrangementContiguousB;

  //
  // Shared memory layouts
  //

  // shared memory layout for wmma is same as the operands' layout in global memory
  using SmemLayoutA = LayoutA;
  using SmemLayoutB = LayoutB;

  // Pad shared memory to avoid bank conflicts
  static int const kPaddingA = 128 / sizeof_bits<ElementA>::value;
  static int const kPaddingB = 128 / sizeof_bits<ElementB>::value;
  
  //
  // Iterators to write to shared memory
  //

  using IteratorThreadMapA = transform::PitchLinearStripminedThreadMap<
    layout::PitchLinearShape<Shape::kM, Shape::kK>,
    kThreads,
    kAccessSizeInBits / sizeof_bits<ElementA>::value
  >;

  using SmemIteratorA = transform::threadblock::RegularTileIterator<
      MatrixShape<Shape::kM, Shape::kK>, ElementA, SmemLayoutA, 1,
      IteratorThreadMapA>;

  using IteratorThreadMapB =  transform::PitchLinearStripminedThreadMap<
    layout::PitchLinearShape<Shape::kK, Shape::kN>,
    kThreads,
    kAccessSizeInBits / sizeof_bits<ElementB>::value
  >;

  using SmemIteratorB = transform::threadblock::RegularTileIterator<
      MatrixShape<Shape::kK, Shape::kN>, ElementB, SmemLayoutB, 0,
      IteratorThreadMapB>;

  //
  // Warp-level matrix multiply operator
  //

  // Define the warp-level tensor op
  using Policy = cutlass::gemm::warp::MmaTensorOpPolicy<
    cutlass::arch::Wmma<
      InstructionShape,
      ElementA,
      LayoutA,
      ElementB,
      LayoutB,
      ElementC,
      LayoutC,
      Operator
    >,
    cutlass::MatrixShape<1, 1>
  >;

  using MmaTensorOp = cutlass::gemm::warp::MmaTensorOpWmma<
    WarpShape,
    ElementA,
    SmemLayoutA,
    ElementB,
    SmemLayoutB,
    ElementC,
    LayoutC,
    Policy
  >;

  using MmaPolicy = MmaPolicy<
    MmaTensorOp,
    MatrixShape<kPaddingA, 0>,
    MatrixShape<kPaddingB, 0>,
    WarpCount::kK
  >;
};

} // namespace threadblock
} // namespace gemm
} // namespace cutlass

#endif // defined(CUTLASS_ARCH_WMMA_ENABLED)