/home/docs/checkouts/readthedocs.org/user_builds/advanced-micro-devices-composable-kernel/checkouts/docs-6.4.3/include/ck/tensor_operation/gpu/device/impl/device_pool2d_fwd_nhwc_nhwc.hpp Source File

/home/docs/checkouts/readthedocs.org/user_builds/advanced-micro-devices-composable-kernel/checkouts/docs-6.4.3/include/ck/tensor_operation/gpu/device/impl/device_pool2d_fwd_nhwc_nhwc.hpp Source File#

Composable Kernel: /home/docs/checkouts/readthedocs.org/user_builds/advanced-micro-devices-composable-kernel/checkouts/docs-6.4.3/include/ck/tensor_operation/gpu/device/impl/device_pool2d_fwd_nhwc_nhwc.hpp Source File
Go to the documentation of this file.
 // SPDX-License-Identifier: MIT
 // Copyright (c) 2024, Advanced Micro Devices, Inc. All rights reserved.
  
 #pragma once
  
 #include <iostream>
 #include <sstream>
  
 #include "ck/tensor_description/tensor_descriptor.hpp"
 #include "ck/tensor_description/tensor_descriptor_helper.hpp"
 #include "ck/tensor_operation/gpu/device/tensor_layout.hpp"
 #include "ck/tensor_operation/gpu/device/reduction_operator_mapping.hpp"
 #include "ck/tensor_operation/gpu/device/device_pool_fwd.hpp"
 #include "ck/tensor_operation/gpu/device/impl/device_reduce_common.hpp"
 #include "ck/tensor_operation/gpu/grid/gridwise_2d_reduction_threadwise.hpp"
 #include "ck/host_utility/device_prop.hpp"
 #include "ck/host_utility/kernel_launch.hpp"
  
 namespace ck {
 namespace tensor_operation {
 namespace device {
  
 template <typename InDataType,
           typename OutDataType,
           typename IndexDataType, // enable if OutputIndex == true
           typename ComputeDataType,
           ck::ReduceTensorOp ReduceOpId,
           bool OutputIndex,
           ck::index_t BlockSize,
           ck::index_t MThreadClusterSize,
           ck::index_t KThreadClusterSize,
           ck::index_t MThreadSliceSize,
           ck::index_t KThreadSliceSize,
           ck::index_t InSrcOutDstVectorSize>
 struct DevicePool2dFwd_NHWC_NHWC : public DevicePoolFwd<4,
                                                         2,
                                                         InDataType,
                                                         OutDataType,
                                                         IndexDataType,
                                                         tensor_layout::convolution::NHWC,
                                                         tensor_layout::convolution::NHWC,
                                                         ReduceOpId,
                                                         OutputIndex>
 {
     static constexpr auto I0 = Number<0>{};
     static constexpr auto I1 = Number<1>{};
  
     static constexpr index_t InOutRank  = 4;
     static constexpr index_t WindowRank = 2;
  
     using ReduceOperation = typename reduce_binary_operator<ReduceOpId>::opType;
  
     using InElementwiseOperation =
         typename reduce_unary_operator<ReduceOpId, true, true>::InElementwiseOperation;
  
     using AccElementwiseOperation =
         typename reduce_unary_operator<ReduceOpId, true, true>::AccElementwiseOperation;
  
     static constexpr ck::index_t M_BlockTileSize = MThreadClusterSize * MThreadSliceSize;
     static constexpr ck::index_t K_BlockTileSize = KThreadClusterSize * KThreadSliceSize;
  
     static auto MakeABGridDescriptor_A_M_K_B_M(std::vector<ck::index_t> input_nchw_lengths,
                                                std::vector<ck::index_t> output_nchw_lengths,
                                                std::vector<ck::index_t> input_nchw_stride,
                                                std::vector<ck::index_t> output_nchw_stride,
                                                std::vector<ck::index_t> window_spatial_yx_lengths,
                                                std::vector<ck::index_t> window_yx_strides,
                                                std::vector<ck::index_t> window_yx_dilations,
                                                std::vector<ck::index_t> input_left_hw_pads,
                                                std::vector<ck::index_t> input_right_hw_pads)
     {
         const index_t N  = input_nchw_lengths[0];
         const index_t C  = input_nchw_lengths[1];
         const index_t Hi = input_nchw_lengths[2];
         const index_t Wi = input_nchw_lengths[3];
  
         const index_t Ho = output_nchw_lengths[2];
         const index_t Wo = output_nchw_lengths[3];
         const index_t Y  = window_spatial_yx_lengths[0];
         const index_t X  = window_spatial_yx_lengths[1];
  
         const index_t WindowStrideH = window_yx_strides[0];
         const index_t WindowStrideW = window_yx_strides[1];
  
         const index_t WindowDilationH = window_yx_dilations[0];
         const index_t WindowDilationW = window_yx_dilations[1];
  
         const index_t InLeftPadH = input_left_hw_pads[0];
         const index_t InLeftPadW = input_left_hw_pads[1];
  
         const index_t InRightPadH = input_right_hw_pads[0];
         const index_t InRightPadW = input_right_hw_pads[1];
  
         const index_t MRaw = N * Ho * Wo * C;
         const index_t MPad = math::integer_least_multiple(MRaw, M_BlockTileSize) - MRaw;
  
         const index_t KRaw = Y * X;
         const index_t KPad = math::integer_least_multiple(KRaw, K_BlockTileSize) - KRaw;
  
         // A[ReduceM, ReduceK]
         const index_t Ni_stride = input_nchw_stride[0];
         const index_t Ci_stride = input_nchw_stride[1];
         const index_t Hi_stride = input_nchw_stride[2];
         const index_t Wi_stride = input_nchw_stride[3];
  
         const auto in_grid_desc_n_hi_wi_c = make_naive_tensor_descriptor(
             make_tuple(N, Hi, Wi, C), make_tuple(Ni_stride, Hi_stride, Wi_stride, Ci_stride));
  
         const auto in_grid_desc_n_hip_wip_c = transform_tensor_descriptor(
             in_grid_desc_n_hi_wi_c,
             make_tuple(make_pass_through_transform(N),
                        make_pad_transform(Hi, InLeftPadH, InRightPadH),
                        make_pad_transform(Wi, InLeftPadW, InRightPadW),
                        make_pass_through_transform(C)),
             make_tuple(Sequence<0>{}, Sequence<1>{}, Sequence<2>{}, Sequence<3>{}),
             make_tuple(Sequence<0>{}, Sequence<1>{}, Sequence<2>{}, Sequence<3>{}));
  
         const auto in_grid_desc_n_y_ho_x_wo_c = transform_tensor_descriptor(
             in_grid_desc_n_hip_wip_c,
             make_tuple(
                 make_pass_through_transform(N),
                 make_embed_transform(make_tuple(Y, Ho), make_tuple(WindowDilationH, WindowStrideH)),
                 make_embed_transform(make_tuple(X, Wo), make_tuple(WindowDilationW, WindowStrideW)),
                 make_pass_through_transform(C)),
             make_tuple(Sequence<0>{}, Sequence<1>{}, Sequence<2>{}, Sequence<3>{}),
             make_tuple(Sequence<0>{}, Sequence<1, 2>{}, Sequence<3, 4>{}, Sequence<5>{}));
  
         const auto in_grid_desc_reducemraw_reducekraw =
             transform_tensor_descriptor(in_grid_desc_n_y_ho_x_wo_c,
                                         make_tuple(make_merge_transform(make_tuple(N, Ho, Wo, C)),
                                                    make_merge_transform(make_tuple(Y, X))),
                                         make_tuple(Sequence<0, 2, 4, 5>{}, Sequence<1, 3>{}),
                                         make_tuple(Sequence<0>{}, Sequence<1>{}));
  
         const auto in_grid_desc_reducem_reducek = transform_tensor_descriptor(
             in_grid_desc_reducemraw_reducekraw,
             make_tuple(make_right_pad_transform(MRaw, MPad), make_right_pad_transform(KRaw, KPad)),
             make_tuple(Sequence<0>{}, Sequence<1>{}),
             make_tuple(Sequence<0>{}, Sequence<1>{}));
  
         // B[ReduceM]
         const index_t No_stride = output_nchw_stride[0];
         const index_t Co_stride = output_nchw_stride[1];
         const index_t Ho_stride = output_nchw_stride[2];
         const index_t Wo_stride = output_nchw_stride[3];
  
         const auto out_grid_desc_n_ho_wo_c = make_naive_tensor_descriptor(
             make_tuple(N, Hi, Wi, C), make_tuple(No_stride, Ho_stride, Wo_stride, Co_stride));
  
         const auto out_grid_desc_reducemraw =
             transform_tensor_descriptor(out_grid_desc_n_ho_wo_c,
                                         make_tuple(make_merge_transform(make_tuple(N, Ho, Wo, C))),
                                         make_tuple(Sequence<0, 1, 2, 3>{}),
                                         make_tuple(Sequence<0>{}));
  
         const auto out_grid_desc_reducem =
             transform_tensor_descriptor(out_grid_desc_reducemraw,
                                         make_tuple(make_right_pad_transform(MRaw, MPad)),
                                         make_tuple(Sequence<0>{}),
                                         make_tuple(Sequence<0>{}));
  
         return make_tuple(in_grid_desc_reducem_reducek, out_grid_desc_reducem);
     }
  
     using ABGridDescs =
         decltype(MakeABGridDescriptor_A_M_K_B_M({}, {}, {}, {}, {}, {}, {}, {}, {}));
  
     using AGridDesc_M_K = remove_cvref_t<decltype(ABGridDescs{}[I0])>;
     using BGridDesc_M   = remove_cvref_t<decltype(ABGridDescs{}[I1])>;
  
     struct Argument : public BaseArgument
     {
         Argument(const InDataType* p_in_dev,
                  OutDataType* p_out_dev,
                  IndexDataType* p_out_indices_dev,
                  std::vector<ck::index_t>& input_nchw_lengths,
                  std::vector<ck::index_t>& output_nchw_lengths,
                  std::vector<ck::index_t>& input_nchw_stride,
                  std::vector<ck::index_t>& output_nchw_stride,
                  std::vector<ck::index_t>&, // indices_nchw_stride
                  std::vector<ck::index_t>& window_spatial_yx_lengths,
                  std::vector<ck::index_t>& window_yx_strides,
                  std::vector<ck::index_t>& window_yx_dilations,
                  std::vector<ck::index_t>& input_left_hw_pads,
                  std::vector<ck::index_t>& input_right_hw_pads)
             : p_in_dev_{p_in_dev},
               p_out_dev_{p_out_dev},
               p_out_indices_dev_{p_out_indices_dev},
               a_grid_desc_m_k_{},
               b_grid_desc_m_{},
               input_nchw_lengths_{input_nchw_lengths},
               output_nchw_lengths_{output_nchw_lengths},
               input_nchw_stride_{input_nchw_stride},
               output_nchw_stride_{output_nchw_stride}
         {
             const auto descs = MakeABGridDescriptor_A_M_K_B_M(input_nchw_lengths,
                                                               output_nchw_lengths,
                                                               input_nchw_stride,
                                                               output_nchw_stride,
                                                               window_spatial_yx_lengths,
                                                               window_yx_strides,
                                                               window_yx_dilations,
                                                               input_left_hw_pads,
                                                               input_right_hw_pads);
  
             a_grid_desc_m_k_ = descs[I0];
             b_grid_desc_m_   = descs[I1];
  
             int32_t reduceLength = window_spatial_yx_lengths[0] * window_spatial_yx_lengths[1];
  
             std::tie(in_element_op_, acc_element_op_) =
                 reduce_unary_operator<ReduceOpId, true, true>::GetElementwiseOperator(reduceLength);
         }
  
         const InDataType* p_in_dev_;
         OutDataType* p_out_dev_;
         IndexDataType* p_out_indices_dev_;
         AGridDesc_M_K a_grid_desc_m_k_;
         BGridDesc_M b_grid_desc_m_;
  
         InElementwiseOperation in_element_op_;
         AccElementwiseOperation acc_element_op_;
  
         // for checking vector load/store
         std::vector<ck::index_t> input_nchw_lengths_;
         std::vector<ck::index_t> output_nchw_lengths_;
         std::vector<ck::index_t> input_nchw_stride_;
         std::vector<ck::index_t> output_nchw_stride_;
     };
  
     struct Invoker : public BaseInvoker
     {
         float Run(const Argument& arg, const StreamConfig& stream_config = StreamConfig{})
         {
             // for NHWC, the dim C is the fastest dimension, and is not reduced.
             // Hence, it is in M dimension for reduction kernel.
             static constexpr index_t InSrcOutDstVectorDim = 0; // 0: M, 1: K
  
             using gridwise_reduce =
                 GridwiseReduction_mk_to_m_threadwise<InDataType,
                                                      OutDataType,
                                                      ComputeDataType,
                                                      IndexDataType,
                                                      AGridDesc_M_K,
                                                      BGridDesc_M,
                                                      ReduceOperation,
                                                      InElementwiseOperation,
                                                      AccElementwiseOperation,
                                                      InMemoryDataOperationEnum::Set,
                                                      false, // propagate_nan
                                                      BlockSize,
                                                      MThreadSliceSize,
                                                      KThreadSliceSize,
                                                      InSrcOutDstVectorDim,
                                                      InSrcOutDstVectorSize,
                                                      InSrcOutDstVectorSize>;
  
             const auto kernel =
                 kernel_reduce_threadwise<gridwise_reduce,
                                          OutputIndex,
                                          true,  // pooling need to return global index
                                          false, // don't have index input
                                          InDataType,
                                          OutDataType,
                                          ComputeDataType,
                                          IndexDataType,
                                          AGridDesc_M_K,
                                          BGridDesc_M,
                                          InElementwiseOperation,
                                          AccElementwiseOperation>;
  
             ck::index_t M = arg.a_grid_desc_m_k_.GetLength(I0);
  
             const index_t grid_size = (M / M_BlockTileSize);
  
             return launch_and_time_kernel(stream_config,
                                           kernel,
                                           dim3(grid_size),
                                           dim3(BlockSize),
                                           0,
                                           arg.a_grid_desc_m_k_,
                                           arg.b_grid_desc_m_,
                                           arg.in_element_op_,
                                           arg.acc_element_op_,
                                           float(1),
                                           arg.p_in_dev_,
                                           nullptr,
                                           float(0),
                                           arg.p_out_dev_,
                                           arg.p_out_indices_dev_);
         }
  
         float Run(const BaseArgument* p_arg,
                   const StreamConfig& stream_config = StreamConfig{}) override
         {
             return Run(*dynamic_cast<const Argument*>(p_arg), stream_config);
         }
     };
  
     bool IsSupportedArgument(const BaseArgument* p_arg) override
     {
         const Argument* pArg = dynamic_cast<const Argument*>(p_arg);
  
         // C should be fastest dimension
         if(pArg->input_nchw_stride_[1] != 1)
             return false;
  
         for(int i = 0; i < InOutRank; ++i)
         {
             if(pArg->input_nchw_stride_[i] == 1 &&
                pArg->input_nchw_lengths_[i] % InSrcOutDstVectorSize != 0)
                 return false;
  
             if(pArg->output_nchw_stride_[i] == 1 &&
                pArg->output_nchw_lengths_[i] % InSrcOutDstVectorSize != 0)
                 return false;
         }
  
         return true;
     }
  
     virtual std::unique_ptr<BaseArgument>
     MakeArgumentPointer(const void* p_in_dev,
                         void* p_out_dev,
                         void* p_out_indices_dev,
                         std::vector<ck::index_t> input_nchw_lengths,
                         std::vector<ck::index_t> window_yx_lengths,
                         std::vector<ck::index_t> output_nchw_lengths,
                         std::vector<ck::index_t> input_nchw_stride,
                         std::vector<ck::index_t> output_nchw_stride,
                         std::vector<ck::index_t> indices_nchw_stride,
                         std::vector<ck::index_t> window_yx_strides,
                         std::vector<ck::index_t> window_yx_dilations,
                         std::vector<ck::index_t> input_left_hw_pads,
                         std::vector<ck::index_t> input_right_hw_pads,
                         std::vector<ck::index_t> pooling_dims) override
     {
         if(input_nchw_lengths.size() != InOutRank || window_yx_lengths.size() != WindowRank ||
            input_nchw_lengths.size() != InOutRank || window_yx_strides.size() != WindowRank ||
            window_yx_dilations.size() != WindowRank || input_left_hw_pads.size() != WindowRank ||
            input_right_hw_pads.size() != WindowRank)
             throw std::runtime_error("dimension is incorrect");
  
         if(pooling_dims != std::vector<ck::index_t>{2, 3})
             throw std::runtime_error("pooling_dims only support {2, 3} in pool2d so far");
  
         if(output_nchw_stride != indices_nchw_stride)
             throw std::runtime_error(
                 "output_nchw_stride need to be equal to indices_nchw_stride for now");
  
         return std::make_unique<Argument>(static_cast<const InDataType*>(p_in_dev),
                                           static_cast<OutDataType*>(p_out_dev),
                                           static_cast<IndexDataType*>(p_out_indices_dev),
                                           input_nchw_lengths,
                                           output_nchw_lengths,
                                           input_nchw_stride,
                                           output_nchw_stride,
                                           indices_nchw_stride,
                                           window_yx_lengths,
                                           window_yx_strides,
                                           window_yx_dilations,
                                           input_left_hw_pads,
                                           input_right_hw_pads);
     }
  
     std::unique_ptr<BaseInvoker> MakeInvokerPointer() override
     {
         return std::make_unique<Invoker>(Invoker{});
     }
  
     std::string GetTypeString() const override
     {
         auto str = std::stringstream();
  
         // clang-format off
         str << "DevicePool2dFwd_NHWC_NHWC<" << BlockSize << ",";
         str << "M_C" << MThreadClusterSize << "_S" << MThreadSliceSize << ",";
         str << "K_C" << KThreadClusterSize << "_S" << KThreadSliceSize << ",";
         str <<"InSrcOutDstVectorSize_" << InSrcOutDstVectorSize << ">";
         // clang-format on
  
         return str.str();
     }
 };
  
 } // namespace device
 } // namespace tensor_operation
 } // namespace ck