c_alltoall_op.cu.cc

/* Copyright (c) 2019 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 "paddle/fluid/operators/collective/c_alltoall_op.h"

#if defined(PADDLE_WITH_NCCL)
#include "paddle/fluid/platform/collective_helper.h"
#include "paddle/fluid/platform/nccl_helper.h"
#endif

namespace paddle {
namespace operators {

template <typename T>
class CAllToAllOpCUDAKernel : public framework::OpKernel<T> {
 public:
  void Compute(const framework::ExecutionContext& ctx) const override {
#if defined(PADDLE_WITH_NCCL)
    auto x = ctx.Input<framework::LoDTensor>("X");
    auto out = ctx.Output<framework::LoDTensor>("Out");
    int send_numel = x->numel();
    ncclDataType_t dtype = platform::ToNCCLDataType(x->type());

    int ring_id = ctx.Attr<int>("ring_id");
    auto place = ctx.GetPlace();
    auto comm = platform::NCCLCommContext::Instance().Get(ring_id, place);
    int nranks = comm->nranks();
    PADDLE_ENFORCE_GE(
        ring_id, 0,
        platform::errors::InvalidArgument(
            "The ring_id (%d) for c_scatter_op must be non-negative.",
            ring_id));

    cudaStream_t stream = nullptr;
    if (ctx.Attr<bool>("use_calc_stream")) {
      auto dev_ctx = platform::DeviceContextPool::Instance().Get(place);
      stream = static_cast<platform::CUDADeviceContext*>(dev_ctx)->stream();
    } else {
      stream = comm->stream();
    }

    framework::DDim x_dims = x->dims();
    framework::DDim out_dims(x_dims);
    auto send_buf = x->data<T>();
    auto recv_buf = out->mutable_data<T>(out_dims, place);
    size_t offset = 0;
    send_numel /= nranks;
    PADDLE_ENFORCE_CUDA_SUCCESS(platform::dynload::ncclGroupStart());
    for (auto i = 0; i < nranks; ++i) {
      PADDLE_ENFORCE_CUDA_SUCCESS(platform::dynload::ncclSend(
          send_buf + offset, send_numel, dtype, i, comm->comm(), stream));
      PADDLE_ENFORCE_CUDA_SUCCESS(platform::dynload::ncclRecv(
          recv_buf + offset, send_numel, dtype, i, comm->comm(), stream));
      offset += send_numel;
    }
    PADDLE_ENFORCE_CUDA_SUCCESS(platform::dynload::ncclGroupEnd());
#else
    PADDLE_ENFORCE_EQ(
        true, false,
        platform::errors::Unavailable("PaddlePaddle should compile with GPU."));
#endif
  }
};

}  // namespace operators
}  // namespace paddle

namespace ops = paddle::operators;
namespace plat = paddle::platform;

REGISTER_OP_CUDA_KERNEL(c_alltoall, ops::CAllToAllOpCUDAKernel<float>,
                        ops::CAllToAllOpCUDAKernel<double>,
                        ops::CAllToAllOpCUDAKernel<int>,
                        ops::CAllToAllOpCUDAKernel<int64_t>,
                        ops::CAllToAllOpCUDAKernel<plat::float16>);