reduce_op.cu 2.2 KB
Newer Older
G
guosheng 已提交
1 2
/* Copyright (c) 2016 PaddlePaddle Authors. All Rights Reserve.

L
Luo Tao 已提交
3 4 5
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
G
guosheng 已提交
6

L
Luo Tao 已提交
7
    http://www.apache.org/licenses/LICENSE-2.0
G
guosheng 已提交
8

L
Luo Tao 已提交
9 10 11 12 13
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. */
G
guosheng 已提交
14 15 16 17 18 19

#define EIGEN_USE_GPU
#include "paddle/operators/reduce_op.h"

namespace ops = paddle::operators;

Q
QI JUN 已提交
20 21 22
#define REGISTER_REDUCE_GPU_KERNEL(reduce_type, functor, grad_functor)    \
  REGISTER_OP_CUDA_KERNEL(                                                \
      reduce_type, ops::ReduceKernel<paddle::platform::CUDADeviceContext, \
23 24 25 26 27 28 29
                                     float, ops::functor>,                \
      ops::ReduceKernel<paddle::platform::CUDADeviceContext, double,      \
                        ops::functor>,                                    \
      ops::ReduceKernel<paddle::platform::CUDADeviceContext, int,         \
                        ops::functor>,                                    \
      ops::ReduceKernel<paddle::platform::CUDADeviceContext, int64_t,     \
                        ops::functor>);                                   \
Q
QI JUN 已提交
30 31 32
  REGISTER_OP_CUDA_KERNEL(                                                \
      reduce_type##_grad,                                                 \
      ops::ReduceGradKernel<paddle::platform::CUDADeviceContext, float,   \
33 34 35 36 37 38
                            ops::grad_functor>,                           \
      ops::ReduceGradKernel<paddle::platform::CUDADeviceContext, double,  \
                            ops::grad_functor>,                           \
      ops::ReduceGradKernel<paddle::platform::CUDADeviceContext, int,     \
                            ops::grad_functor>,                           \
      ops::ReduceGradKernel<paddle::platform::CUDADeviceContext, int64_t, \
Q
QI JUN 已提交
39
                            ops::grad_functor>);
40 41

FOR_EACH_KERNEL_FUNCTOR(REGISTER_REDUCE_GPU_KERNEL);