conv_cudnn_op.cu 12.5 KB
Newer Older
武毅 已提交
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240 241 242 243 244 245 246 247 248 249 250 251 252 253 254 255 256 257 258 259 260 261 262 263 264 265 266 267 268 269 270 271 272 273 274 275 276 277
/* Copyright (c) 2016 PaddlePaddle Authors All Rights Reserve.

   Licensed under the Apache License, Version 2.0 (the "License");
   you may not use this file except in compliance with the License.
   You may obtain a copy of the License at

   http://www.apache.org/licenses/LICENSE-2.0

   Unless required by applicable law or agreed to in writing, software
   distributed under the License is distributed on an "AS IS" BASIS,
   WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
   See the License for the specific language governing permissions and
   limitations under the License. */

#include "paddle/framework/eigen.h"
#include "paddle/framework/op_registry.h"
#include "paddle/memory/memory.h"
#include "paddle/operators/conv2d_op.h"
#include "paddle/platform/assert.h"
#include "paddle/platform/cudnn_helper.h"

namespace paddle {
namespace operators {

using Tensor = framework::Tensor;
using ScopedTensorDescriptor = platform::ScopedTensorDescriptor;
using ScopedFilterDescriptor = platform::ScopedFilterDescriptor;
using ScopedConvolutionDescriptor = platform::ScopedConvolutionDescriptor;
using DataLayout = platform::DataLayout;
using CUDADeviceContext = platform::CUDADeviceContext;

static constexpr size_t kCONV_CUDNN_WORKSPACE_LIMIT_BYTES = 1024 * 1024 * 1024;

// NOTE: framework::vectorize converts to type int64_t
//       which does not fit cudnn inputs.
std::vector<int> Dims2Vector(const framework::DDim& dims) {
  std::vector<int> ret;
  for (int i = 0; i < dims.size(); i++) {
    ret.push_back(dims[i]);
  }
  return ret;
}

template <typename T>
class CudnnConvOpKernel : public framework::OpKernel<T> {
 public:
  void Compute(const framework::ExecutionContext& ctx) const override {
    PADDLE_ENFORCE(platform::is_gpu_place(ctx.GetPlace()),
                   "It must use GPUPlace.");
    auto* input = ctx.Input<Tensor>("Input");
    auto* filter = ctx.Input<Tensor>("Filter");
    auto* output = ctx.Output<Tensor>("Output");

    std::vector<int> strides = ctx.Attr<std::vector<int>>("strides");
    std::vector<int> paddings = ctx.Attr<std::vector<int>>("paddings");
    std::vector<int> dilations = ctx.Attr<std::vector<int>>("dilations");
    int groups = ctx.Attr<int>("groups");
    int user_workspace_size = ctx.Attr<int>("workspace_size_MB");

    const T* input_data = input->data<T>();
    const T* filter_data = filter->data<T>();
    T* output_data = output->mutable_data<T>(ctx.GetPlace());

    // ------------------- cudnn descriptors ---------------------
    ScopedTensorDescriptor input_desc;
    ScopedTensorDescriptor output_desc;
    ScopedFilterDescriptor filter_desc;
    ScopedConvolutionDescriptor conv_desc;
    DataLayout layout = DataLayout::kNCHW;

    cudnnTensorDescriptor_t cudnn_input_desc =
        input_desc.descriptor<T>(layout, Dims2Vector(input->dims()), groups);
    cudnnTensorDescriptor_t cudnn_output_desc =
        output_desc.descriptor<T>(layout, Dims2Vector(output->dims()), groups);
    cudnnFilterDescriptor_t cudnn_filter_desc =
        filter_desc.descriptor<T>(layout, Dims2Vector(filter->dims()), groups);
    cudnnConvolutionDescriptor_t cudnn_conv_desc =
        conv_desc.descriptor<T>(paddings, strides, dilations);

    int input_channels = input->dims()[1];
    int input_height = input->dims()[2];
    int input_width = input->dims()[3];
    int output_channels = output->dims()[1];
    int output_height = output->dims()[2];
    int output_width = output->dims()[3];

    int group_offset_in = input_channels / groups * input_height * input_width;
    int group_offset_out =
        output_channels / groups * output_height * output_width;
    int group_offset_filter = filter->numel() / groups;
    // ------------------- cudnn conv workspace ---------------------
    void* cudnn_workspace = nullptr;
    size_t workspace_size_in_bytes;  // final workspace to allocate.
    size_t workspace_size_limit = kCONV_CUDNN_WORKSPACE_LIMIT_BYTES;
    if (user_workspace_size > 0) {
      workspace_size_limit = user_workspace_size * 1024 * 1024;
    }
    // ------------------- cudnn conv algorithm ---------------------
    cudnnConvolutionFwdAlgo_t algo;
    auto handle = ctx.cuda_device_context().cudnn_handle();

    PADDLE_ENFORCE(platform::dynload::cudnnGetConvolutionForwardAlgorithm(
        handle, cudnn_input_desc, cudnn_filter_desc, cudnn_conv_desc,
        cudnn_output_desc, CUDNN_CONVOLUTION_FWD_SPECIFY_WORKSPACE_LIMIT,
        workspace_size_limit, &algo));
    // get workspace size able to allocate
    PADDLE_ENFORCE(platform::dynload::cudnnGetConvolutionForwardWorkspaceSize(
        handle, cudnn_input_desc, cudnn_filter_desc, cudnn_conv_desc,
        cudnn_output_desc, algo, &workspace_size_in_bytes));
    // Allocate on GPU memory
    platform::GPUPlace gpu = boost::get<platform::GPUPlace>(ctx.GetPlace());
    cudnn_workspace = paddle::memory::Alloc(gpu, workspace_size_in_bytes);
    // ------------------- cudnn conv forward ---------------------
    T alpha = 1.0f, beta = 0.0f;
    for (int i = 0; i < groups; i++) {
      PADDLE_ENFORCE(platform::dynload::cudnnConvolutionForward(
          handle, &alpha, cudnn_input_desc, input_data + i * group_offset_in,
          cudnn_filter_desc, filter_data + i * group_offset_filter,
          cudnn_conv_desc, algo, cudnn_workspace, workspace_size_in_bytes,
          &beta, cudnn_output_desc, output_data + i * group_offset_out));
    }
    // Release the cudnn workspace
    paddle::memory::Free(gpu, cudnn_workspace);
  }
};

template <typename T>
class CudnnConvGradOpKernel : public framework::OpKernel<T> {
 public:
  void Compute(const framework::ExecutionContext& ctx) const override {
    PADDLE_ENFORCE(platform::is_gpu_place(ctx.GetPlace()),
                   "It must use GPUPlace.");
    auto input = ctx.Input<Tensor>("Input");
    auto filter = ctx.Input<Tensor>("Filter");
    auto output_grad = ctx.Input<Tensor>(framework::GradVarName("Output"));
    auto input_grad = ctx.Output<Tensor>(framework::GradVarName("Input"));
    auto filter_grad = ctx.Output<Tensor>(framework::GradVarName("Filter"));

    const T* input_data = input->data<T>();
    const T* output_grad_data = output_grad->data<T>();
    const T* filter_data = filter->data<T>();

    std::vector<int> strides = ctx.Attr<std::vector<int>>("strides");
    std::vector<int> paddings = ctx.Attr<std::vector<int>>("paddings");
    std::vector<int> dilations = ctx.Attr<std::vector<int>>("dilations");
    int groups = ctx.Attr<int>("groups");
    int user_workspace_size = ctx.Attr<int>("workspace_size_MB");

    // ------------------- cudnn descriptors ---------------------
    ScopedTensorDescriptor input_desc;
    ScopedTensorDescriptor output_grad_desc;
    ScopedTensorDescriptor input_grad_desc;

    ScopedFilterDescriptor filter_desc;
    ScopedFilterDescriptor filter_grad_desc;
    ScopedConvolutionDescriptor conv_desc;
    DataLayout layout = DataLayout::kNCHW;

    cudnnTensorDescriptor_t cudnn_input_desc =
        input_desc.descriptor<T>(layout, Dims2Vector(input->dims()), groups);
    cudnnTensorDescriptor_t cudnn_output_grad_desc =
        output_grad_desc.descriptor<T>(layout, Dims2Vector(output_grad->dims()),
                                       groups);
    cudnnFilterDescriptor_t cudnn_filter_desc =
        filter_desc.descriptor<T>(layout, Dims2Vector(filter->dims()), groups);
    cudnnTensorDescriptor_t cudnn_input_grad_desc = nullptr;
    cudnnFilterDescriptor_t cudnn_filter_grad_desc = nullptr;

    cudnnConvolutionDescriptor_t cudnn_conv_desc =
        conv_desc.descriptor<T>(paddings, strides, dilations);

    int input_channels = input->dims()[1];
    int input_height = input->dims()[2];
    int input_width = input->dims()[3];
    int output_grad_channels = filter->dims()[0];
    int output_grad_height = output_grad->dims()[2];
    int output_grad_width = output_grad->dims()[3];

    int group_offset_in = input_channels / groups * input_height * input_width;
    int group_offset_out =
        output_grad_channels / groups * output_grad_height * output_grad_width;
    int group_offset_filter = filter->numel() / groups;
    // ------------------- cudnn backward algorithm ---------------------
    cudnnConvolutionBwdDataAlgo_t data_algo;
    cudnnConvolutionBwdFilterAlgo_t filter_algo;
    size_t workspace_size_in_bytes = 0, tmp_size = 0;
    size_t workspace_size_limit = kCONV_CUDNN_WORKSPACE_LIMIT_BYTES;
    if (user_workspace_size > 0) {
      workspace_size_limit = user_workspace_size * 1024 * 1024;
    }

    auto handle = ctx.cuda_device_context().cudnn_handle();
    if (input_grad) {
      cudnn_input_grad_desc = input_grad_desc.descriptor<T>(
          layout, Dims2Vector(input_grad->dims()), groups);
      PADDLE_ENFORCE(
          platform::dynload::cudnnGetConvolutionBackwardDataAlgorithm(
              handle, cudnn_filter_desc,
              // dyDesc: Handle to the previously initialized input differential
              // tensor descriptor.
              cudnn_output_grad_desc, cudnn_conv_desc,
              // dxDesc: Handle to the previously initialized output tensor
              // descriptor.
              cudnn_input_grad_desc,
              CUDNN_CONVOLUTION_BWD_DATA_SPECIFY_WORKSPACE_LIMIT,
              workspace_size_limit, &data_algo));
      PADDLE_ENFORCE(
          platform::dynload::cudnnGetConvolutionBackwardDataWorkspaceSize(
              handle, cudnn_filter_desc, cudnn_output_grad_desc,
              cudnn_conv_desc, cudnn_input_grad_desc, data_algo, &tmp_size));
      workspace_size_in_bytes = std::max(workspace_size_in_bytes, tmp_size);
    }

    if (filter_grad) {
      cudnn_filter_grad_desc = filter_grad_desc.descriptor<T>(
          layout, Dims2Vector(filter_grad->dims()), groups);
      PADDLE_ENFORCE(
          platform::dynload::cudnnGetConvolutionBackwardFilterAlgorithm(
              handle, cudnn_input_desc, cudnn_output_grad_desc, cudnn_conv_desc,
              cudnn_filter_desc,
              CUDNN_CONVOLUTION_BWD_FILTER_SPECIFY_WORKSPACE_LIMIT,
              workspace_size_limit, &filter_algo));

      PADDLE_ENFORCE(
          platform::dynload::cudnnGetConvolutionBackwardFilterWorkspaceSize(
              handle, cudnn_input_desc, cudnn_output_grad_desc, cudnn_conv_desc,
              cudnn_filter_desc, filter_algo, &tmp_size));
      workspace_size_in_bytes = std::max(workspace_size_in_bytes, tmp_size);
    }
    // ------------------- cudnn conv workspace ---------------------
    // Already on GPU
    void* cudnn_workspace = nullptr;
    platform::GPUPlace gpu = boost::get<platform::GPUPlace>(ctx.GetPlace());
    cudnn_workspace = paddle::memory::Alloc(gpu, workspace_size_in_bytes);
    // ------------------- cudnn conv backward data ---------------------
    // FIXME(typhoonzero): template type T may not be the same as cudnn call.
    T alpha = 1.0f, beta = 0.0f;
    if (input_grad) {
      T* input_grad_data = input_grad->mutable_data<T>(ctx.GetPlace());
      auto t = framework::EigenVector<T>::Flatten(*input_grad);
      t.device(ctx.GetEigenDevice<platform::GPUPlace>()) =
          t.constant(static_cast<T>(0));
      for (int i = 0; i < groups; i++) {
        PADDLE_ENFORCE(platform::dynload::cudnnConvolutionBackwardData(
            handle, &alpha, cudnn_filter_desc,
            filter_data + i * group_offset_filter, cudnn_output_grad_desc,
            output_grad_data + i * group_offset_out, cudnn_conv_desc, data_algo,
            cudnn_workspace, workspace_size_in_bytes, &beta,
            cudnn_input_grad_desc, input_grad_data + i * group_offset_in));
      }
    }
    // ------------------- cudnn conv backward filter ---------------------
    if (filter_grad) {
      T* filter_grad_data = filter_grad->mutable_data<T>(ctx.GetPlace());
      auto t = framework::EigenVector<T>::Flatten(*filter_grad);
      t.device(ctx.GetEigenDevice<platform::GPUPlace>()) =
          t.constant(static_cast<T>(0));
      for (int i = 0; i < groups; i++) {
        PADDLE_ENFORCE(platform::dynload::cudnnConvolutionBackwardFilter(
            handle, &alpha, cudnn_input_desc, input_data + i * group_offset_in,
            cudnn_output_grad_desc, output_grad_data + i * group_offset_out,
            cudnn_conv_desc, filter_algo, cudnn_workspace,
            workspace_size_in_bytes, &beta, cudnn_filter_grad_desc,
            filter_grad_data + i * group_offset_filter));
      }
    }
    // Release the cudnn workspace
    paddle::memory::Free(gpu, cudnn_workspace);
  }
};

}  // namespace operators
}  // namespace paddle

REGISTER_OP_GPU_KERNEL(conv_cudnn, paddle::operators::CudnnConvOpKernel<float>);
REGISTER_OP_GPU_KERNEL(conv_cudnn_grad,
                       paddle::operators::CudnnConvGradOpKernel<float>);