path: root/tensorflow/core/kernels/batch_matmul_op.cc

/* Copyright 2015 Google Inc. 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.
==============================================================================*/

// See docs in ../ops/math_ops.cc.

#define EIGEN_USE_THREADS

#include <vector>
#include "third_party/eigen3/unsupported/Eigen/CXX11/Tensor"
#include "tensorflow/core/framework/op.h"
#include "tensorflow/core/framework/op_kernel.h"
#include "tensorflow/core/framework/tensor.h"
#include "tensorflow/core/framework/tensor_shape.h"
#include "tensorflow/core/framework/types.h"
#include "tensorflow/core/kernels/fill_functor.h"
#include "tensorflow/core/platform/logging.h"
#include "tensorflow/core/platform/types.h"
#include "tensorflow/core/util/work_sharder.h"

#if GOOGLE_CUDA
#include "tensorflow/core/platform/stream_executor.h"
#endif  // GOOGLE_CUDA

namespace tensorflow {

typedef Eigen::ThreadPoolDevice CPUDevice;
typedef Eigen::GpuDevice GPUDevice;

template <typename Device, typename Scalar>
struct LaunchBatchMatMul;

template <typename Scalar>
struct LaunchBatchMatMul<CPUDevice, Scalar> {
  static void Launch(OpKernelContext* context, const Tensor& in_x,
                     const Tensor& in_y, bool adj_x, bool adj_y, Tensor* out) {
    auto Tx = in_x.tensor<Scalar, 3>();
    auto Ty = in_y.tensor<Scalar, 3>();
    auto Tz = out->tensor<Scalar, 3>();

    // Shards "n"-matmuls into "num" shards. Each shard is
    // dispatched to a thread.
    auto worker_threads = *(context->device()->tensorflow_cpu_worker_threads());
    const int64 num_units = in_x.dim_size(0);
    const int64 cost_per_unit =
        in_x.dim_size(0) * in_x.dim_size(1) * out->dim_size(2);
    Shard(worker_threads.num_threads, worker_threads.workers, num_units,
          cost_per_unit, [&Tx, &Ty, adj_x, adj_y, &Tz](int start, int limit) {
            LaunchBatchMatMul<CPUDevice, Scalar>::Run(Tx, Ty, adj_x, adj_y, Tz,
                                                      start, limit);
          });
  }

  template <typename In, typename Out>
  static void Run(In Tx, In Ty, bool adj_x, bool adj_y, Out Tz, int start,
                  int limit) {
    Eigen::array<Eigen::IndexPair<Eigen::DenseIndex>, 1> contract_pairs;

    Eigen::internal::scalar_conjugate_op<Scalar> conj;
    if (!adj_x && !adj_y) {
      for (int i = start; i < limit; ++i) {
        auto x = Tx.template chip<0>(i);
        auto y = Ty.template chip<0>(i);
        auto z = Tz.template chip<0>(i);
        contract_pairs[0] = Eigen::IndexPair<Eigen::DenseIndex>(1, 0);
        z = x.contract(y, contract_pairs);  // matmul
      }
    } else if (!adj_x && adj_y) {
      for (int i = start; i < limit; ++i) {
        auto x = Tx.template chip<0>(i);
        auto y = Ty.template chip<0>(i).unaryExpr(conj);
        auto z = Tz.template chip<0>(i);
        contract_pairs[0] = Eigen::IndexPair<Eigen::DenseIndex>(1, 1);
        z = x.contract(y, contract_pairs);  // matmul
      }
    } else if (adj_x && !adj_y) {
      for (int i = start; i < limit; ++i) {
        auto x = Tx.template chip<0>(i).unaryExpr(conj);
        auto y = Ty.template chip<0>(i);
        auto z = Tz.template chip<0>(i);
        contract_pairs[0] = Eigen::IndexPair<Eigen::DenseIndex>(0, 0);
        z = x.contract(y, contract_pairs);  // matmul
      }
    } else {
      for (int i = start; i < limit; ++i) {
        auto x = Tx.template chip<0>(i).unaryExpr(conj);
        auto y = Ty.template chip<0>(i).unaryExpr(conj);
        auto z = Tz.template chip<0>(i);
        contract_pairs[0] = Eigen::IndexPair<Eigen::DenseIndex>(0, 1);
        z = x.contract(y, contract_pairs);  // matmul
      }
    }
  }
};

#if GOOGLE_CUDA

namespace {
template <typename T>
perftools::gputools::DeviceMemory<T> AsDeviceMemory(const T* cuda_memory) {
  perftools::gputools::DeviceMemoryBase wrapped(const_cast<T*>(cuda_memory));
  perftools::gputools::DeviceMemory<T> typed(wrapped);
  return typed;
}

class CublasScratchAllocator : public perftools::gputools::ScratchAllocator {
 public:
  using Stream = ::perftools::gputools::Stream;
  using DeviceMemoryBytes = ::perftools::gputools::DeviceMemory<uint8>;

  CublasScratchAllocator(OpKernelContext* context) : context_(context) {}

  int64 GetMemoryLimitInBytes(Stream* stream) override { return -1; }

  perftools::gputools::port::StatusOr<DeviceMemoryBytes> AllocateBytes(
      Stream* stream, int64 byte_size) override {
    Tensor temporary_memory;

    Status allocation_status(context_->allocate_temp(
        DT_UINT8, TensorShape({byte_size}), &temporary_memory));
    if (!allocation_status.ok()) {
      return perftools::gputools::port::StatusOr<DeviceMemoryBytes>(
          DeviceMemoryBytes::MakeFromByteSize(nullptr, 0));
    }
    // Hold the reference of the allocated tensors until the end of the
    // allocator.
    allocated_tensors_.push_back(temporary_memory);
    return perftools::gputools::port::StatusOr<DeviceMemoryBytes>(
        DeviceMemoryBytes::MakeFromByteSize(
            temporary_memory.flat<uint8>().data(),
            temporary_memory.flat<uint8>().size()));
  }

 private:
  OpKernelContext* context_;
  std::vector<Tensor> allocated_tensors_;
};
}  // namespace

template <typename Scalar>
struct LaunchBatchMatMul<GPUDevice, Scalar> {
  static void Launch(OpKernelContext* context, const Tensor& in_x,
                     const Tensor& in_y, bool adj_x, bool adj_y, Tensor* out) {
    perftools::gputools::blas::Transpose trans[] = {
        perftools::gputools::blas::Transpose::kNoTranspose,
        perftools::gputools::blas::Transpose::kTranspose};
    const uint64 m = in_x.dim_size(adj_x ? 2 : 1);
    const uint64 k = in_x.dim_size(adj_x ? 1 : 2);
    const uint64 n = in_y.dim_size(adj_y ? 1 : 2);
    const uint64 batch_size = in_x.dim_size(0);
    auto blas_transpose_a = trans[adj_x];
    auto blas_transpose_b = trans[adj_y];

    auto* stream = context->op_device_context()->stream();
    OP_REQUIRES(context, stream, errors::Internal("No GPU stream available."));

    typedef perftools::gputools::DeviceMemory<Scalar> DeviceMemoryType;
    std::vector<DeviceMemoryType> a_device_memory;
    std::vector<DeviceMemoryType> b_device_memory;
    std::vector<DeviceMemoryType> c_device_memory;
    std::vector<DeviceMemoryType*> a_ptrs;
    std::vector<DeviceMemoryType*> b_ptrs;
    std::vector<DeviceMemoryType*> c_ptrs;
    a_device_memory.reserve(batch_size);
    b_device_memory.reserve(batch_size);
    c_device_memory.reserve(batch_size);
    a_ptrs.reserve(batch_size);
    b_ptrs.reserve(batch_size);
    c_ptrs.reserve(batch_size);
    auto* a_base_ptr = in_x.template flat<Scalar>().data();
    auto* b_base_ptr = in_y.template flat<Scalar>().data();
    auto* c_base_ptr = out->template flat<Scalar>().data();
    for (int64 i = 0; i < batch_size; ++i) {
      a_device_memory.push_back(AsDeviceMemory(a_base_ptr + i * m * k));
      b_device_memory.push_back(AsDeviceMemory(b_base_ptr + i * k * n));
      c_device_memory.push_back(AsDeviceMemory(c_base_ptr + i * m * n));
      a_ptrs.push_back(&a_device_memory.back());
      b_ptrs.push_back(&b_device_memory.back());
      c_ptrs.push_back(&c_device_memory.back());
    }

    // Cublas does
    // C = A x B
    // where A, B and C are assumed to be in column major.
    // We want the output to be in row-major, so we can compute
    // C' = B' x A' (' stands for transpose)
    CublasScratchAllocator scratch_allocator(context);
    bool blas_launch_status =
        stream
            ->ThenBlasGemmBatchedWithScratch(
                blas_transpose_b, blas_transpose_a, n, m, k,
                static_cast<Scalar>(1.0), b_ptrs, adj_y ? k : n, a_ptrs,
                adj_x ? m : k, static_cast<Scalar>(0.0), c_ptrs, n, batch_size,
                &scratch_allocator)
            .ok();
    if (!blas_launch_status) {
      context->SetStatus(errors::Internal(
          "Blas SGEMMBatched launch failed : a.shape=",
          in_x.shape().DebugString(), ", b.shape=", in_y.shape().DebugString(),
          ", m=", m, ", n=", n, ", k=", k, ", batch_size=", batch_size));
    }
  }
};

#endif  // GOOGLE_CUDA

template <typename Device, typename Scalar>
class BatchMatMul : public OpKernel {
 public:
  explicit BatchMatMul(OpKernelConstruction* context) : OpKernel(context) {
    OP_REQUIRES_OK(context, context->GetAttr("adj_x", &adj_x_));
    OP_REQUIRES_OK(context, context->GetAttr("adj_y", &adj_y_));
  }

  virtual ~BatchMatMul() {}

  void Compute(OpKernelContext* ctx) override {
    const Tensor& in0 = ctx->input(0);
    const Tensor& in1 = ctx->input(1);
    OP_REQUIRES(ctx, in0.dims() == in1.dims(),
                errors::InvalidArgument("In[0] and In[1] has different ndims: ",
                                        in0.shape().DebugString(), " vs. ",
                                        in1.shape().DebugString()));
    const int ndims = in0.dims();
    OP_REQUIRES(
        ctx, ndims >= 3,
        errors::InvalidArgument("In[0] and In[1] ndims must be >= 3: ", ndims));
    TensorShape out_shape;
    for (int i = 0; i < ndims - 2; ++i) {
      OP_REQUIRES(ctx, in0.dim_size(i) == in1.dim_size(i),
                  errors::InvalidArgument("In[0].dim(", i, ") and In[1].dim(",
                                          i, ") must be the same: ",
                                          in0.shape().DebugString(), " vs ",
                                          in1.shape().DebugString()));
      out_shape.AddDim(in0.dim_size(i));
    }
    auto n = out_shape.num_elements();
    auto d0 = in0.dim_size(ndims - 2);
    auto d1 = in0.dim_size(ndims - 1);
    Tensor in0_reshaped;
    CHECK(in0_reshaped.CopyFrom(in0, TensorShape({n, d0, d1})));
    auto d2 = in1.dim_size(ndims - 2);
    auto d3 = in1.dim_size(ndims - 1);
    Tensor in1_reshaped;
    CHECK(in1_reshaped.CopyFrom(in1, TensorShape({n, d2, d3})));
    if (adj_x_) std::swap(d0, d1);
    if (adj_y_) std::swap(d2, d3);
    OP_REQUIRES(ctx, d1 == d2,
                errors::InvalidArgument(
                    "In[0] mismatch In[1] shape: ", d1, " vs. ", d2, ": ",
                    in0.shape().DebugString(), " ", in1.shape().DebugString(),
                    " ", adj_x_, " ", adj_y_));
    out_shape.AddDim(d0);
    out_shape.AddDim(d3);
    Tensor* out = nullptr;
    OP_REQUIRES_OK(ctx, ctx->allocate_output(0, out_shape, &out));
    if (out->NumElements() == 0) {
      return;
    }
    if (in0.NumElements() == 0 || in1.NumElements() == 0) {
      functor::SetZeroFunctor<Device, Scalar> f;
      f(ctx->eigen_device<Device>(), out->flat<Scalar>());
      return;
    }
    Tensor out_reshaped;
    CHECK(out_reshaped.CopyFrom(*out, TensorShape({n, d0, d3})));
    LaunchBatchMatMul<Device, Scalar>::Launch(ctx, in0_reshaped, in1_reshaped,
                                              adj_x_, adj_y_, &out_reshaped);
  }

 private:
  bool adj_x_;
  bool adj_y_;
};

#define REGISTER_CPU(TYPE)                                              \
  REGISTER_KERNEL_BUILDER(                                              \
      Name("BatchMatMul").Device(DEVICE_CPU).TypeConstraint<TYPE>("T"), \
      BatchMatMul<CPUDevice, TYPE>)

#define REGISTER_GPU(TYPE)                                              \
  REGISTER_KERNEL_BUILDER(                                              \
      Name("BatchMatMul").Device(DEVICE_GPU).TypeConstraint<TYPE>("T"), \
      BatchMatMul<GPUDevice, TYPE>)

REGISTER_CPU(float);
REGISTER_CPU(double);
REGISTER_CPU(int32);
REGISTER_CPU(complex64);

#ifdef GOOGLE_CUDA
REGISTER_GPU(float);
#endif  // GOOGLE_CUDA

#undef REGISTER_CPU
#undef REGISTER_GPU
}  // end namespace tensorflow