batch_norm_op.cu

/* Copyright (c) 2016 PaddlePaddle Authors. All Rights Reserved.

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 <algorithm>
#include <cfloat>
#include <string>
#include <vector>
#include "cub/cub.cuh"
#include "paddle/fluid/framework/data_layout.h"
#include "paddle/fluid/operators/batch_norm_op.h"
#include "paddle/fluid/operators/math/math_function.h"
#include "paddle/fluid/platform/cudnn_helper.h"
#include "paddle/fluid/platform/float16.h"

DECLARE_bool(cudnn_batchnorm_spatial_persistent);

namespace paddle {
namespace operators {

using Tensor = framework::Tensor;
using DataLayout = framework::DataLayout;
template <typename T>
using CudnnDataType = platform::CudnnDataType<T>;
template <typename T>
using BatchNormParamType = typename CudnnDataType<T>::BatchNormParamType;

template <typename T>
class BatchNormKernel<platform::CUDADeviceContext, T>
    : public framework::OpKernel<T> {
 public:
  void Compute(const framework::ExecutionContext &ctx) const override {
    PADDLE_ENFORCE(platform::is_gpu_place(ctx.GetPlace()),
                   "It must use CUDAPlace.");
    double epsilon = static_cast<double>(ctx.Attr<float>("epsilon"));
    float momentum = ctx.Attr<float>("momentum");
    const bool is_test = ctx.Attr<bool>("is_test");
    const bool use_global_stats = ctx.Attr<bool>("use_global_stats");
    const std::string data_layout_str = ctx.Attr<std::string>("data_layout");
    const DataLayout data_layout =
        framework::StringToDataLayout(data_layout_str);

    // Get the size for each dimension.
    // NCHW [batch_size, in_channels, in_height, in_width]
    const auto *x = ctx.Input<Tensor>("X");
    const auto &x_dims = x->dims();
    PADDLE_ENFORCE(x_dims.size() >= 2 && x_dims.size() <= 5,
                   "The Input dim size should be between 2 and 5");

    auto *y = ctx.Output<Tensor>("Y");
    y->mutable_data<T>(ctx.GetPlace());

    int N, C, H, W, D;
    ExtractNCWHD(x_dims, data_layout, &N, &C, &H, &W, &D);

    auto dtype = platform::CudnnDataType<T>::type;
    const bool fast_nhwc_batch_norm =
        is_test ||
        (dtype == CUDNN_DATA_HALF && FLAGS_cudnn_batchnorm_spatial_persistent);

    auto compute_format =
        fast_nhwc_batch_norm && data_layout == DataLayout::kNHWC
            ? DataLayout::kNHWC
            : DataLayout::kNCHW;

    Tensor transformed_x(x->type());
    Tensor transformed_y(y->type());
    if (data_layout == DataLayout::kNHWC &&
        compute_format == DataLayout::kNCHW && x_dims.size() > 2) {
      VLOG(3) << "Transform input tensor from NHWC to NCHW.";
      ResizeToChannelFirst<platform::CUDADeviceContext, T>(ctx, x,
                                                           &transformed_x);
      TransToChannelFirst<platform::CUDADeviceContext, T>(ctx, x,
                                                          &transformed_x);
      ResizeToChannelFirst<platform::CUDADeviceContext, T>(ctx, y,
                                                           &transformed_y);
    } else {
      transformed_x.ShareDataWith(*x);
      transformed_y.ShareDataWith(*y);
    }

    // ------------------- cudnn descriptors ---------------------
    cudnnTensorDescriptor_t data_desc_;
    cudnnTensorDescriptor_t bn_param_desc_;
    cudnnBatchNormMode_t mode_;

    PADDLE_ENFORCE_CUDA_SUCCESS(
        platform::dynload::cudnnCreateTensorDescriptor(&data_desc_));
    PADDLE_ENFORCE_CUDA_SUCCESS(
        platform::dynload::cudnnCreateTensorDescriptor(&bn_param_desc_));

    if (epsilon <= CUDNN_BN_MIN_EPSILON - FLT_EPSILON) {
      LOG(ERROR) << "Provided epsilon is smaller than "
                 << "CUDNN_BN_MIN_EPSILON. Setting it to "
                 << "CUDNN_BN_MIN_EPSILON instead.";
    }
    epsilon = std::max(epsilon, CUDNN_BN_MIN_EPSILON);
#if CUDNN_VERSION_MIN(7, 0, 0)
    if (FLAGS_cudnn_batchnorm_spatial_persistent) {
      mode_ = CUDNN_BATCHNORM_SPATIAL_PERSISTENT;
    } else {
      mode_ = CUDNN_BATCHNORM_SPATIAL;
    }
#else
    mode_ = CUDNN_BATCHNORM_SPATIAL;
#endif

    VLOG(3) << "Setting descriptors.";
    std::vector<int> dims;
    std::vector<int> strides;
    if (compute_format == DataLayout::kNCHW) {
      dims = {N, C, H, W, D};
      strides = {C * H * W * D, H * W * D, W * D, D, 1};
    } else {
      dims = {N, C, H, W, D};
      strides = {H * W * D * C, 1, W * D * C, D * C, C};
    }
    PADDLE_ENFORCE_CUDA_SUCCESS(platform::dynload::cudnnSetTensorNdDescriptor(
        data_desc_, CudnnDataType<T>::type,
        x_dims.size() > 3 ? x_dims.size() : 4, dims.data(), strides.data()));
    // Note: PERSISTENT not implemented for inference
    PADDLE_ENFORCE_CUDA_SUCCESS(
        platform::dynload::cudnnDeriveBNTensorDescriptor(
            bn_param_desc_, data_desc_,
            is_test ? CUDNN_BATCHNORM_SPATIAL : mode_));

    const auto *scale = ctx.Input<Tensor>("Scale");
    const auto *bias = ctx.Input<Tensor>("Bias");

    auto &dev_ctx = ctx.template device_context<platform::CUDADeviceContext>();

    auto handle = dev_ctx.cudnn_handle();

    // Now, depending on whether we are running test or not, we have two paths.
    if (is_test || use_global_stats) {
      // only when test we use input to do computation.
      const auto *est_mean = ctx.Input<Tensor>("Mean");
      const auto *est_var = ctx.Input<Tensor>("Variance");
      // Run inference mode.
      PADDLE_ENFORCE_EQ(est_mean->dims().size(), 1UL);
      PADDLE_ENFORCE_EQ(est_var->dims().size(), 1UL);
      PADDLE_ENFORCE_EQ(est_mean->dims()[0], C);
      PADDLE_ENFORCE_EQ(est_var->dims()[0], C);

      PADDLE_ENFORCE_CUDA_SUCCESS(
          platform::dynload::cudnnBatchNormalizationForwardInference(
              handle,
              // Note: PERSISTENT not implemented for inference
              CUDNN_BATCHNORM_SPATIAL, CudnnDataType<T>::kOne(),
              CudnnDataType<T>::kZero(), data_desc_,
              transformed_x.template data<T>(), data_desc_,
              transformed_y.template mutable_data<T>(ctx.GetPlace()),
              bn_param_desc_, scale->template data<BatchNormParamType<T>>(),
              bias->template data<BatchNormParamType<T>>(),
              est_mean->template data<BatchNormParamType<T>>(),
              est_var->template data<BatchNormParamType<T>>(), epsilon));
    } else {
      // if MomentumTensor is set, use MomentumTensor value, momentum
      // is only used in this training branch
      if (ctx.HasInput("MomentumTensor")) {
        const auto *mom_tensor = ctx.Input<Tensor>("MomentumTensor");
        Tensor mom_cpu;
        TensorCopySync(*mom_tensor, platform::CPUPlace(), &mom_cpu);
        momentum = mom_cpu.data<float>()[0];
      }

      // Run training mode.
      // obtain running mean and running inv var, and see if we need to
      // initialize them.

      auto *mean_out = ctx.Output<Tensor>("MeanOut");
      auto *variance_out = ctx.Output<Tensor>("VarianceOut");
      mean_out->mutable_data<BatchNormParamType<T>>(ctx.GetPlace());
      variance_out->mutable_data<BatchNormParamType<T>>(ctx.GetPlace());

      auto *saved_mean = ctx.Output<Tensor>("SavedMean");
      auto *saved_variance = ctx.Output<Tensor>("SavedVariance");
      saved_mean->mutable_data<BatchNormParamType<T>>(ctx.GetPlace());
      saved_variance->mutable_data<BatchNormParamType<T>>(ctx.GetPlace());
      math::SetConstant<platform::CUDADeviceContext, BatchNormParamType<T>>
          functor;
      functor(dev_ctx, saved_mean, static_cast<BatchNormParamType<T>>(0));
      functor(dev_ctx, saved_variance, static_cast<BatchNormParamType<T>>(0));

      if ((N * H * W * D) == 1) {
        // Only 1 element in normalization dimension,
        // skip the batch norm calculation, let y = x.
        framework::TensorCopy(*x, ctx.GetPlace(), y);
      } else {
        double this_factor = 1. - momentum;

        bool called = false;
#if CUDNN_VERSION_MIN(7, 4, 1)
        if (compute_format == DataLayout::kNHWC) {
          called = true;
          size_t workspace_size = 0;
          size_t reserve_space_size = 0;
          void *reserve_space_ptr = nullptr;
          void *workspace_ptr = nullptr;
          Tensor workspace_tensor;
          // Create reserve space and workspace for batch norm.
          // Create tensor for each batchnorm op, it will be used in the
          // backward. Thus this tensor shouldn't be temp.
          auto *reserve_space = ctx.Output<Tensor>("ReserveSpace");
          PADDLE_ENFORCE_NOT_NULL(
              reserve_space,
              platform::errors::NotFound(
                  "The argument ReserveSpace of batch_norm op is not found."));

          // --------------- cudnn batchnorm workspace ---------------
          PADDLE_ENFORCE_CUDA_SUCCESS(
              platform::dynload::
                  cudnnGetBatchNormalizationForwardTrainingExWorkspaceSize(
                      /*handle=*/handle,
                      /*mode=*/mode_,
                      /*bnIps=*/CUDNN_BATCHNORM_OPS_BN,
                      /*xDesc=*/data_desc_,
                      /*zDesc=*/nullptr,
                      /*yDesc=*/data_desc_,
                      /*bnScaleBiasMeanVarDesc=*/bn_param_desc_,
                      /*activationDesc=*/nullptr,
                      /*sizeInBytes=*/&workspace_size));

          // -------------- cudnn batchnorm reserve space --------------
          PADDLE_ENFORCE_CUDA_SUCCESS(
              platform::dynload::
                  cudnnGetBatchNormalizationTrainingExReserveSpaceSize(
                      /*handle=*/handle,
                      /*mode=*/mode_,
                      /*bnOps=*/CUDNN_BATCHNORM_OPS_BN,
                      /*activationDesc=*/nullptr,
                      /*xDesc=*/data_desc_,
                      /*sizeInBytes=*/&reserve_space_size));

          reserve_space_ptr = reserve_space->mutable_data(
              ctx.GetPlace(), transformed_x.type(), reserve_space_size);
          workspace_ptr = workspace_tensor.mutable_data(
              ctx.GetPlace(), transformed_x.type(), workspace_size);
          PADDLE_ENFORCE_CUDA_SUCCESS(
              platform::dynload::cudnnBatchNormalizationForwardTrainingEx(
                  handle, mode_, CUDNN_BATCHNORM_OPS_BN,
                  CudnnDataType<T>::kOne(), CudnnDataType<T>::kZero(),
                  data_desc_, transformed_x.template data<T>(), nullptr,
                  nullptr, data_desc_, transformed_y.template data<T>(),
                  bn_param_desc_, scale->template data<BatchNormParamType<T>>(),
                  bias->template data<BatchNormParamType<T>>(), this_factor,
                  mean_out->template mutable_data<BatchNormParamType<T>>(
                      ctx.GetPlace()),
                  variance_out->template mutable_data<BatchNormParamType<T>>(
                      ctx.GetPlace()),
                  epsilon,
                  saved_mean->template mutable_data<BatchNormParamType<T>>(
                      ctx.GetPlace()),
                  saved_variance->template mutable_data<BatchNormParamType<T>>(
                      ctx.GetPlace()),
                  nullptr, workspace_ptr, workspace_size, reserve_space_ptr,
                  reserve_space_size));
        }
#endif
        if (!called) {
          PADDLE_ENFORCE_CUDA_SUCCESS(
              platform::dynload::cudnnBatchNormalizationForwardTraining(
                  handle, mode_, CudnnDataType<T>::kOne(),
                  CudnnDataType<T>::kZero(), data_desc_,
                  transformed_x.template data<T>(), data_desc_,
                  transformed_y.template mutable_data<T>(ctx.GetPlace()),
                  bn_param_desc_, scale->template data<BatchNormParamType<T>>(),
                  bias->template data<BatchNormParamType<T>>(), this_factor,
                  mean_out->template mutable_data<BatchNormParamType<T>>(
                      ctx.GetPlace()),
                  variance_out->template mutable_data<BatchNormParamType<T>>(
                      ctx.GetPlace()),
                  epsilon,
                  saved_mean->template mutable_data<BatchNormParamType<T>>(
                      ctx.GetPlace()),
                  saved_variance->template mutable_data<BatchNormParamType<T>>(
                      ctx.GetPlace())));
        }
      }
    }

    if (data_layout == DataLayout::kNHWC &&
        compute_format == DataLayout::kNCHW && x_dims.size() > 2) {
      VLOG(3) << "Transform batchnorm output from NCHW to NHWC";
      TransToChannelLast<paddle::platform::CUDADeviceContext, T>(
          ctx, &transformed_y, y);
    }
    // clean when exit.
    PADDLE_ENFORCE_CUDA_SUCCESS(
        platform::dynload::cudnnDestroyTensorDescriptor(data_desc_));
    PADDLE_ENFORCE_CUDA_SUCCESS(
        platform::dynload::cudnnDestroyTensorDescriptor(bn_param_desc_));
  }
};

template <typename T, int BlockDim, framework::DataLayout layout>
static __global__ void KeBNBackwardScaleBias(
    const T *dy, const T *x, const BatchNormParamType<T> *mean,
    const BatchNormParamType<T> *variance, const double epsilon, const int N,
    const int C, const int HxW, BatchNormParamType<T> *dscale,
    BatchNormParamType<T> *dbias) {
  const int outer_size = C;
  const int inner_size = N * HxW;
  typedef cub::BlockReduce<BatchNormParamType<T>, BlockDim> BlockReduce;
  __shared__ typename BlockReduce::TempStorage ds_storage;
  __shared__ typename BlockReduce::TempStorage db_storage;

  for (int i = blockIdx.x; i < outer_size; i += gridDim.x) {
    BatchNormParamType<T> ds_sum = static_cast<BatchNormParamType<T>>(0);
    BatchNormParamType<T> db_sum = static_cast<BatchNormParamType<T>>(0);

    BatchNormParamType<T> inv_var_i = 1.0 / sqrt(variance[i] + epsilon);
    BatchNormParamType<T> mean_i = mean[i];
    for (int j = threadIdx.x; j < inner_size; j += blockDim.x) {
      const int index = layout == framework::DataLayout::kNCHW
                            ? (j / HxW * C + i) * HxW + j % HxW
                            : j * outer_size + i;
      ds_sum += static_cast<BatchNormParamType<T>>(dy[index]) *
                (static_cast<BatchNormParamType<T>>(x[index]) - mean_i);
      db_sum += static_cast<BatchNormParamType<T>>(dy[index]);
    }
    ds_sum = BlockReduce(ds_storage).Reduce(ds_sum, cub::Sum());
    db_sum = BlockReduce(db_storage).Reduce(db_sum, cub::Sum());
    if (threadIdx.x == 0) {
      dscale[i] = ds_sum * inv_var_i;
      dbias[i] = db_sum;
    }
    __syncthreads();
  }
}

template <typename T, framework::DataLayout layout>
static __global__ void KeBNBackwardData(const T *dy,
                                        const BatchNormParamType<T> *scale,
                                        const BatchNormParamType<T> *variance,
                                        const double epsilon, const int C,
                                        const int HxW, const int num, T *dx) {
  int gid = blockIdx.x * blockDim.x + threadIdx.x;
  int stride = blockDim.x * gridDim.x;
  for (int i = gid; i < num; i += stride) {
    const int c = layout == framework::DataLayout::kNCHW ? i / HxW % C : i % C;
    BatchNormParamType<T> inv_var = 1.0 / sqrt(variance[c] + epsilon);
    dx[i] = static_cast<T>(static_cast<BatchNormParamType<T>>(dy[i]) *
                           scale[c] * inv_var);
  }
}

template <typename T, int BlockDim, framework::DataLayout layout>
static __global__ void BNBackwardData(const T *dy,
                                      const BatchNormParamType<T> *scale,
                                      const BatchNormParamType<T> *mean,
                                      const T *x,
                                      const BatchNormParamType<T> *variance,
                                      const int C, const int N, const int HxW,
                                      T *dx) {
  const int outer_size = C;
  const int inner_size = N * HxW;
  typedef cub::BlockReduce<BatchNormParamType<T>, BlockDim> BlockReduce;
  __shared__ typename BlockReduce::TempStorage dy_storage;
  __shared__ typename BlockReduce::TempStorage dy_x_sub_mean_storage;
  __shared__ BatchNormParamType<T> dy_sum_val;
  __shared__ BatchNormParamType<T> dy_x_sub_mean_sum_val;

  for (int i = blockIdx.x; i < outer_size; i += gridDim.x) {
    BatchNormParamType<T> inv_var_i = variance[i];
    BatchNormParamType<T> mean_i = mean[i];
    BatchNormParamType<T> dy_sum = static_cast<BatchNormParamType<T>>(0);
    BatchNormParamType<T> dy_x_sub_mean_sum =
        static_cast<BatchNormParamType<T>>(0);
    for (int j = threadIdx.x; j < inner_size; j += blockDim.x) {
      const int index = layout == framework::DataLayout::kNCHW
                            ? (j / HxW * C + i) * HxW + j % HxW
                            : j * outer_size + i;
      BatchNormParamType<T> dy_i =
          static_cast<BatchNormParamType<T>>(dy[index]);
      dy_sum += dy_i;
      dy_x_sub_mean_sum +=
          dy_i * (static_cast<BatchNormParamType<T>>(x[index]) - mean_i);
    }

    dy_sum = BlockReduce(dy_storage).Reduce(dy_sum, cub::Sum());
    dy_x_sub_mean_sum = BlockReduce(dy_x_sub_mean_storage)
                            .Reduce(dy_x_sub_mean_sum, cub::Sum());

    if (threadIdx.x == 0) {
      dy_sum_val = dy_sum;
      dy_x_sub_mean_sum_val = dy_x_sub_mean_sum;
    }
    __syncthreads();

    for (int j = threadIdx.x; j < inner_size; j += blockDim.x) {
      const int index = layout == framework::DataLayout::kNCHW
                            ? (j / HxW * C + i) * HxW + j % HxW
                            : j * outer_size + i;
      dx[index] =
          (static_cast<BatchNormParamType<T>>(dy[index]) -
           dy_sum_val / static_cast<BatchNormParamType<T>>(inner_size) -
           (static_cast<BatchNormParamType<T>>(x[index]) - mean_i) *
               dy_x_sub_mean_sum_val * inv_var_i * inv_var_i / inner_size) *
          scale[i] * inv_var_i;
    }
  }
}

template <typename T>
class BatchNormGradKernel<platform::CUDADeviceContext, T>
    : public framework::OpKernel<T> {
 public:
  void Compute(const framework::ExecutionContext &ctx) const override {
    PADDLE_ENFORCE(platform::is_gpu_place(ctx.GetPlace()),
                   "It must use CUDAPlace.");
    double epsilon = static_cast<double>(ctx.Attr<float>("epsilon"));
    const std::string data_layout_str = ctx.Attr<std::string>("data_layout");
    const bool use_global_stats = ctx.Attr<bool>("use_global_stats");

    const DataLayout data_layout =
        framework::StringToDataLayout(data_layout_str);
    const auto *x = ctx.Input<Tensor>("X");
    const auto *d_y = ctx.Input<Tensor>(framework::GradVarName("Y"));
    const auto *scale = ctx.Input<Tensor>("Scale");
    const bool is_test = ctx.Attr<bool>("is_test");
    PADDLE_ENFORCE_EQ(
        is_test, false,
        platform::errors::InvalidArgument(
            "`is_test = True` CANNOT be used in train program. If "
            "you want to use global status in pre_train model, "
            "please set `use_global_stats = True`"));

    const auto &x_dims = x->dims();

    PADDLE_ENFORCE(x_dims.size() >= 2 && x_dims.size() <= 5,
                   "The Input dim size should be between 2 and 5");
    int N, C, H, W, D;
    ExtractNCWHD(x_dims, data_layout, &N, &C, &H, &W, &D);

    // init output
    auto *d_x = ctx.Output<Tensor>(framework::GradVarName("X"));
    auto *d_scale = ctx.Output<Tensor>(framework::GradVarName("Scale"));
    auto *d_bias = ctx.Output<Tensor>(framework::GradVarName("Bias"));

    d_x->mutable_data<T>(ctx.GetPlace());
    if (d_scale && d_bias) {
      d_scale->mutable_data<BatchNormParamType<T>>(ctx.GetPlace());
      d_bias->mutable_data<BatchNormParamType<T>>(ctx.GetPlace());
    }
    PADDLE_ENFORCE_EQ(scale->dims().size(), 1UL);
    PADDLE_ENFORCE_EQ(scale->dims()[0], C);

    auto dtype = platform::CudnnDataType<T>::type;
    const auto *reserve_space = ctx.Input<Tensor>("ReserveSpace");
    const bool fast_nhwc_batch_norm =
        dtype == CUDNN_DATA_HALF && FLAGS_cudnn_batchnorm_spatial_persistent &&
        reserve_space != nullptr;
    auto compute_format =
        fast_nhwc_batch_norm && data_layout == DataLayout::kNHWC
            ? DataLayout::kNHWC
            : DataLayout::kNCHW;

    Tensor transformed_x(x->type());
    Tensor transformed_d_y(d_y->type());
    Tensor transformed_d_x(d_x->type());
    if (data_layout == DataLayout::kNHWC &&
        compute_format == DataLayout::kNCHW) {
      VLOG(3) << "Transform input tensor from NHWC to NCHW.";
      ResizeToChannelFirst<platform::CUDADeviceContext, T>(ctx, x,
                                                           &transformed_x);
      TransToChannelFirst<platform::CUDADeviceContext, T>(ctx, x,
                                                          &transformed_x);
      ResizeToChannelFirst<platform::CUDADeviceContext, T>(ctx, d_y,
                                                           &transformed_d_y);
      TransToChannelFirst<platform::CUDADeviceContext, T>(ctx, d_y,
                                                          &transformed_d_y);
      ResizeToChannelFirst<platform::CUDADeviceContext, T>(ctx, d_x,
                                                           &transformed_d_x);
    } else {
      transformed_x.ShareDataWith(*x);
      transformed_d_y.ShareDataWith(*d_y);
      transformed_d_x.ShareDataWith(*d_x);
    }

    std::vector<int> dims;
    std::vector<int> strides;
    if (compute_format == DataLayout::kNCHW) {
      dims = {N, C, H, W, D};
      strides = {C * H * W * D, H * W * D, W * D, D, 1};
    } else {
      dims = {N, C, H, W, D};
      strides = {H * W * C * D, 1, W * D * C, D * C, C};
    }

    auto &dev_ctx = ctx.template device_context<platform::CUDADeviceContext>();
    const int num = transformed_x.numel();
    const int block = 512;
    int max_threads = dev_ctx.GetMaxPhysicalThreadCount();
    const int max_blocks = std::max(max_threads / block, 1);
    int grid1 = (num + block - 1) / block;
    int grid2 = std::min(C, max_blocks);

    if (!use_global_stats) {
      if ((N * H * W * D) == 1) {
        framework::TensorCopy(*d_y, ctx.GetPlace(), d_x);
        math::SetConstant<platform::CUDADeviceContext, BatchNormParamType<T>>
            functor;
        functor(dev_ctx, d_scale, static_cast<BatchNormParamType<T>>(0));
        functor(dev_ctx, d_bias, static_cast<BatchNormParamType<T>>(0));
        return;
      }

      // ------------------- cudnn descriptors ---------------------
      cudnnTensorDescriptor_t data_desc_;
      cudnnTensorDescriptor_t bn_param_desc_;
      cudnnBatchNormMode_t mode_;

      PADDLE_ENFORCE_CUDA_SUCCESS(
          platform::dynload::cudnnCreateTensorDescriptor(&data_desc_));
      PADDLE_ENFORCE_CUDA_SUCCESS(
          platform::dynload::cudnnCreateTensorDescriptor(&bn_param_desc_));
      if (epsilon <= CUDNN_BN_MIN_EPSILON - FLT_EPSILON) {
        LOG(ERROR) << "Provided epsilon is smaller than "
                   << "CUDNN_BN_MIN_EPSILON. Setting it to "
                   << "CUDNN_BN_MIN_EPSILON instead.";
      }
      epsilon = std::max(epsilon, CUDNN_BN_MIN_EPSILON);
#if CUDNN_VERSION_MIN(7, 0, 0)
      if (FLAGS_cudnn_batchnorm_spatial_persistent) {
        mode_ = CUDNN_BATCHNORM_SPATIAL_PERSISTENT;
      } else {
        mode_ = CUDNN_BATCHNORM_SPATIAL;
      }
#else
      mode_ = CUDNN_BATCHNORM_SPATIAL;
#endif

      PADDLE_ENFORCE_CUDA_SUCCESS(platform::dynload::cudnnSetTensorNdDescriptor(
          data_desc_, CudnnDataType<T>::type,
          x_dims.size() > 3 ? x_dims.size() : 4, dims.data(), strides.data()));
      PADDLE_ENFORCE_CUDA_SUCCESS(
          platform::dynload::cudnnDeriveBNTensorDescriptor(bn_param_desc_,
                                                           data_desc_, mode_));

      const auto *saved_mean = ctx.Input<Tensor>("SavedMean");
      const auto *saved_var = ctx.Input<Tensor>("SavedVariance");
      const auto *saved_mean_data =
          saved_mean->template data<BatchNormParamType<T>>();
      const auto *saved_var_data =
          saved_var->template data<BatchNormParamType<T>>();

      if (d_scale && d_bias) {
        bool called = false;
#if CUDNN_VERSION_MIN(7, 4, 1)
        if (compute_format == DataLayout::kNHWC) {
          called = true;
          size_t workspace_size = 0;
          void *workspace_ptr = nullptr;
          Tensor workspace_tensor;
          auto reserve_space_size = reserve_space->memory_size();
          // --------------- cudnn batchnorm workspace ---------------
          PADDLE_ENFORCE_CUDA_SUCCESS(
              platform::dynload::
                  cudnnGetBatchNormalizationBackwardExWorkspaceSize(
                      /*handle=*/dev_ctx.cudnn_handle(),
                      /*mode=*/mode_,
                      /*bnIps=*/CUDNN_BATCHNORM_OPS_BN,
                      /*xDesc=*/data_desc_,
                      /*yDesc=*/data_desc_,
                      /*dyDesc=*/data_desc_,
                      /*dzDesc=*/nullptr,
                      /*dxDesc=*/data_desc_,
                      /*bnScaleBiasMeanVarDesc=*/bn_param_desc_,
                      /*activationDesc=*/nullptr,
                      /*sizeInBytes=*/&workspace_size));

          workspace_ptr = workspace_tensor.mutable_data(
              ctx.GetPlace(), transformed_x.type(), workspace_size);

          PADDLE_ENFORCE_CUDA_SUCCESS(
              platform::dynload::cudnnBatchNormalizationBackwardEx(
                  /*handle=*/dev_ctx.cudnn_handle(),
                  /*mode=*/mode_,
                  /*bnOps=*/CUDNN_BATCHNORM_OPS_BN,
                  /*alphaDataDiff=*/CudnnDataType<T>::kOne(),
                  /*betaDataDiff=*/CudnnDataType<T>::kZero(),
                  /*alphaParamDiff=*/CudnnDataType<T>::kOne(),
                  /*betaParamDiff=*/CudnnDataType<T>::kZero(),
                  /*xDesc=*/data_desc_,
                  /*xData=*/transformed_x.template data<T>(),
                  /*yDesc=*/nullptr,
                  /*yData=*/nullptr,
                  /*dyDesc=*/data_desc_,
                  /*dyData=*/transformed_d_y.template data<T>(),
                  /*dzDesc=*/nullptr,
                  /*dzData=*/nullptr,
                  /*dxDesc=*/data_desc_,
                  /*dxData=*/transformed_d_x.template mutable_data<T>(
                      ctx.GetPlace()),
                  /*dBnScaleBiasDesc=*/bn_param_desc_,
                  /*bnScaleData=*/scale->template data<BatchNormParamType<T>>(),
                  /*bnBiasData=*/nullptr,
                  /*dBnScaleData=*/d_scale
                      ->template mutable_data<BatchNormParamType<T>>(
                          ctx.GetPlace()),
                  /*dBnBiasData=*/d_bias
                      ->template mutable_data<BatchNormParamType<T>>(
                          ctx.GetPlace()),
                  /*epsilon=*/epsilon,
                  /*savedMean=*/saved_mean_data,
                  /*savedInvVariance=*/saved_var_data,
                  /*activationDesc=*/nullptr,
                  /*workspace=*/workspace_ptr,
                  /*workSpaceSizeInBytes=*/workspace_size,
                  /*reserveSpace=*/const_cast<T *>(
                      reserve_space->template data<T>()),
                  /*reserveSpaceSizeInBytes=*/reserve_space_size));
        }
#endif
        if (!called) {
          PADDLE_ENFORCE_CUDA_SUCCESS(
              platform::dynload::cudnnBatchNormalizationBackward(
                  dev_ctx.cudnn_handle(), mode_, CudnnDataType<T>::kOne(),
                  CudnnDataType<T>::kZero(), CudnnDataType<T>::kOne(),
                  CudnnDataType<T>::kZero(), data_desc_,
                  transformed_x.template data<T>(), data_desc_,
                  transformed_d_y.template data<T>(), data_desc_,
                  transformed_d_x.template mutable_data<T>(ctx.GetPlace()),
                  bn_param_desc_, scale->template data<BatchNormParamType<T>>(),
                  d_scale->template mutable_data<BatchNormParamType<T>>(
                      ctx.GetPlace()),
                  d_bias->template mutable_data<BatchNormParamType<T>>(
                      ctx.GetPlace()),
                  epsilon, saved_mean_data, saved_var_data));
        }

        if (data_layout == DataLayout::kNHWC &&
            compute_format == DataLayout::kNCHW) {
          VLOG(3) << "Transform batchnorm output from NCHW to NHWC";
          TransToChannelLast<paddle::platform::CUDADeviceContext, T>(
              ctx, &transformed_d_x, d_x);
        }
      } else {
        if (compute_format == DataLayout::kNCHW) {
          if (d_x) {
            BNBackwardData<T, block, framework::DataLayout::kNCHW><<<
                grid2, block, 0, dev_ctx.stream()>>>(
                d_y->data<T>(), scale->data<BatchNormParamType<T>>(),
                saved_mean_data, x->data<T>(), saved_var_data, C, N, H * W * D,
                d_x->data<T>());
          }
        } else {
          if (d_x) {
            BNBackwardData<T, block, framework::DataLayout::kNHWC><<<
                grid2, block, 0, dev_ctx.stream()>>>(
                d_y->data<T>(), scale->data<BatchNormParamType<T>>(),
                saved_mean_data, x->data<T>(), saved_var_data, C, N, H * W * D,
                d_x->data<T>());
          }
        }
      }

      // clean when exit.
      PADDLE_ENFORCE_CUDA_SUCCESS(
          platform::dynload::cudnnDestroyTensorDescriptor(data_desc_));
      PADDLE_ENFORCE_CUDA_SUCCESS(
          platform::dynload::cudnnDestroyTensorDescriptor(bn_param_desc_));
    } else {
      const auto *running_mean = ctx.Input<Tensor>("Mean");
      const auto *running_var = ctx.Input<Tensor>("Variance");

      const auto *running_mean_data =
          running_mean->template data<BatchNormParamType<T>>();
      const auto *running_var_data =
          running_var->template data<BatchNormParamType<T>>();

      if (compute_format == DataLayout::kNCHW) {
        if (d_x) {
          KeBNBackwardData<T, framework::DataLayout::kNCHW><<<
              grid1, block, 0, dev_ctx.stream()>>>(
              d_y->data<T>(), scale->data<BatchNormParamType<T>>(),
              running_var_data, epsilon, C, H * W, num, d_x->data<T>());
        }
        if (d_scale && d_bias) {
          KeBNBackwardScaleBias<T, block, framework::DataLayout::kNCHW><<<
              grid2, block, 0, dev_ctx.stream()>>>(
              d_y->data<T>(), x->data<T>(), running_mean_data, running_var_data,
              epsilon, N, C, H * W * D, d_scale->data<BatchNormParamType<T>>(),
              d_bias->data<BatchNormParamType<T>>());
        }
      } else {
        if (d_x) {
          KeBNBackwardData<T, framework::DataLayout::kNHWC><<<
              grid1, block, 0, dev_ctx.stream()>>>(
              d_y->data<T>(), scale->data<BatchNormParamType<T>>(),
              running_var_data, epsilon, C, H * W, num, d_x->data<T>());
        }
        if (d_scale && d_bias) {
          KeBNBackwardScaleBias<T, block, framework::DataLayout::kNHWC><<<
              grid2, block, 0, dev_ctx.stream()>>>(
              d_y->data<T>(), x->data<T>(), running_mean_data, running_var_data,
              epsilon, N, C, H * W * D, d_scale->data<BatchNormParamType<T>>(),
              d_bias->data<BatchNormParamType<T>>());
        }
      }
    }
  }
};

}  // namespace operators
}  // namespace paddle

namespace ops = paddle::operators;
namespace plat = paddle::platform;
REGISTER_OP_CUDA_KERNEL(
    batch_norm, ops::BatchNormKernel<plat::CUDADeviceContext, float>,
    ops::BatchNormKernel<plat::CUDADeviceContext, double>,
    ops::BatchNormKernel<plat::CUDADeviceContext, plat::float16>);
REGISTER_OP_CUDA_KERNEL(
    batch_norm_grad, ops::BatchNormGradKernel<plat::CUDADeviceContext, float>,
    ops::BatchNormGradKernel<plat::CUDADeviceContext, double>,
    ops::BatchNormGradKernel<plat::CUDADeviceContext, plat::float16>);