reducer.cc

// Copyright (c) 2020 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/imperative/reducer.h"

namespace paddle {
namespace imperative {

#if defined(PADDLE_WITH_NCCL)
std::shared_ptr<Reducer> Reducer::s_instance_ = NULL;

// context is used to select the stream for concat
void Group::ConcatTensors(const platform::CUDADeviceContext &context) {
  switch (dtype_) {
    case framework::proto::VarType::FP16:
      ConcatTensorsForAllReduce<platform::float16>(context, dense_tensors_,
                                                   &dense_contents_);
      break;
    case framework::proto::VarType::FP32:
      ConcatTensorsForAllReduce<float>(context, dense_tensors_,
                                       &dense_contents_);
      break;
    case framework::proto::VarType::FP64:
      ConcatTensorsForAllReduce<double>(context, dense_tensors_,
                                        &dense_contents_);
      break;
    default:
      PADDLE_THROW(platform::errors::Unimplemented(
          "Data type (%s) is not supported when it concats tensors for "
          "allreduce.",
          framework::DataTypeToString(dtype_)));
  }
}

// context is used to select the stream for split
void Group::SplitTensors(const platform::CUDADeviceContext &context) {
  switch (dtype_) {
    case framework::proto::VarType::FP16:
      SplitTensorsForAllReduce<platform::float16>(context, &dense_contents_,
                                                  &dense_tensors_);
      break;
    case framework::proto::VarType::FP32:
      SplitTensorsForAllReduce<float>(context, &dense_contents_,
                                      &dense_tensors_);
      break;
    case framework::proto::VarType::FP64:
      SplitTensorsForAllReduce<double>(context, &dense_contents_,
                                       &dense_tensors_);
      break;
    default:
      PADDLE_THROW(platform::errors::Unimplemented(
          "Data type (%s) is not supported when it splits tensors for "
          "allreduce.",
          framework::DataTypeToString(dtype_)));
  }
}

std::ostream &operator<<(std::ostream &out, const Group &group) {
  const auto &vars = group.variable_indices_;
  out << "numul: " << group.all_length_ << " ;is_sparse: " << group.is_sparse_
      << " ;var number: " << vars.size() << "\n";
  auto begin = vars.begin();
  auto end = vars.end();
  out << "[";
  for (int i = 0; begin != end && i < 100; ++i, ++begin) {
    if (i > 0) out << ' ';
    out << *begin;
  }
  if (begin != end) {
    out << " ...";
  }
  out << "]\n";
  return out;
}

Reducer::Reducer(const std::vector<std::shared_ptr<imperative::VarBase>> &vars,
                 const std::vector<std::vector<size_t>> &group_indices,
                 const std::vector<bool> &is_sparse_gradient,
                 std::shared_ptr<imperative::ParallelContext> parallel_ctx,
                 const std::vector<size_t> &group_size_limits)
    : vars_(vars),
      group_indices_(group_indices),
      is_sparse_gradient_(is_sparse_gradient),
      parallel_ctx_(parallel_ctx),
      group_size_limits_(group_size_limits) {
  VLOG(3) << "Start construct the Reducer ...";
  // initialize groups
  InitializeGroups(group_indices);
  for (size_t global_var_index = 0; global_var_index < vars_.size();
       ++global_var_index) {
    vars_[global_var_index]->SharedVar()->AddGradVarLeafBackwardHook(
        std::unique_ptr<LambdaGradAccumulatorPostHook>(
            new LambdaGradAccumulatorPostHook([=](VariableWrapper *grad) {
              this->AddDistHook(grad, global_var_index);
            })));
  }
  // create streams
  compute_stream_ = static_cast<platform::CUDADeviceContext *>(
                        platform::DeviceContextPool::Instance().Get(place_))
                        ->stream();
  comm_stream_ = platform::NCCLCommContext::Instance().Get(0, place_)->stream();
  // create events
  CreateGroupEvents(group_indices.size());
  comm_enent_ = platform::CudaEventResourcePool::Instance().New(
      BOOST_GET_CONST(platform::CUDAPlace, place_).device);

  std::call_once(once_flag_, []() {
    std::atexit([]() { Reducer::GetInstance()->ReleaseReducer(); });
  });
}

void Reducer::ReleaseReducer() {
  for (auto &event : events_) {
    event.reset();
  }
  comm_enent_.reset();
}

void Reducer::CreateGroupEvents(int group_num) {
  // release old events
  for (auto &event : events_) {
    event.reset();
  }
  events_.clear();
  events_.resize(group_num);
  for (auto &event : events_) {
    event = platform::CudaEventResourcePool::Instance().New(
        BOOST_GET_CONST(platform::CUDAPlace, place_).device);
  }
}

void Reducer::InitializeDenseGroups(
    const std::vector<size_t> &variable_indices_, Group *p_group) {
  int64_t all_length = 0;
  for (size_t index = 0; index < variable_indices_.size(); ++index) {
    const auto variable_index = variable_indices_[index];
    const auto &var = vars_[variable_index];
    const auto var_name = var->Name();
    PADDLE_ENFORCE_EQ(is_sparse_gradient_[variable_index], false,
                      platform::errors::PreconditionNotMet(
                          "Tensor %s's GRAD must be LoDTensor, but received "
                          "GRAD is SelectedRows",
                          var_name));

    auto lod_tensor = var->MutableVar()->GetMutable<framework::LoDTensor>();
    PADDLE_ENFORCE_EQ(lod_tensor->IsInitialized(), true,
                      platform::errors::PreconditionNotMet(
                          "Tensor %s is not initialized.", var_name));
    auto size = lod_tensor->numel();
    PADDLE_ENFORCE_GT(
        size, 0, platform::errors::PreconditionNotMet(
                     "The number of tensor %s's elements is 0.", var_name));
    all_length += size;

    p_group->length_.push_back(size);
    // for concat operator
    p_group->dense_tensors_.push_back(framework::Tensor());

    // check the dtype and place, it must be same.
    auto dtype = var->DataType();
    auto place = var->Place();
    if (index > 0) {
      PADDLE_ENFORCE_EQ(
          dtype, p_group->dtype_,
          platform::errors::PreconditionNotMet(
              "Tensor %s has different dtype. Expected dtype is %s, but actual "
              "dtype is %s",
              var_name, framework::DataTypeToString(p_group->dtype_),
              framework::DataTypeToString(dtype)));
      PADDLE_ENFORCE_EQ(place, place_,
                        platform::errors::PreconditionNotMet(
                            "Tensor %s has different place. Expected place is "
                            "%s, but actual place is %s",
                            var_name, place_, place));
    } else {
      p_group->dtype_ = dtype;
      place_ = place;
    }
  }
  p_group->all_length_ = all_length;
}

// Each parameter will be initialized according to the group information.
// For the sparse parameter, sparse_contents_ in the group directly points
// to the parameter. For dense parameters, first construct an empty Tensor().
// Then specify the actual memory in MarkVariableReady.
void Reducer::InitializeGroups(
    const std::vector<std::vector<size_t>> &group_indices) {
  VLOG(3) << "Start initialize groups ..";
  // clear the group
  groups_.clear();
  groups_.reserve(group_indices.size());
  variable_locators_.clear();
  variable_locators_.resize(vars_.size());

  auto group_nums = group_indices.size();
  for (size_t group_index = 0; group_index < group_nums; ++group_index) {
    const auto &variable_indices_ = group_indices[group_index];
    PADDLE_ENFORCE_GT(
        variable_indices_.size(), 0,
        platform::errors::PreconditionNotMet(
            "The number of group[%d]'s elements is 0.", group_index));
    Group group;

    // It's just for check the sparse or dense
    auto first_varbase = vars_[variable_indices_.front()];
    if (variable_indices_.size() == 1 &&
        is_sparse_gradient_[variable_indices_.front()]) {
      // process the sparse gradient. one sparse, one group
      group.sparse_contents_ = first_varbase->MutableGradVar();
      group.dtype_ = first_varbase->DataType();
      group.is_sparse_ = true;
    } else {
      // process the dense gradient.
      InitializeDenseGroups(variable_indices_, &group);
      // Alloc the continuous space
      auto tensor = group.dense_contents_.GetMutable<framework::LoDTensor>();
      tensor->Resize(framework::make_ddim({group.all_length_}))
          .mutable_data(place_, group.dtype_);
    }

    // map variables to this group by VariableLocator
    size_t inside_group_index = 0;
    for (const auto var_index : group_indices[group_index]) {
      variable_locators_[var_index] = VariableLocator{
          .group_index = group_index,
          .inside_group_index = inside_group_index++,
      };
    }
    group.variable_indices_ = std::move(variable_indices_);
    groups_.emplace_back(std::move(group));

    // Debug Message For Reducer
    VLOG(3) << "The Group[" << group_index << "]:";
    VLOG(3) << groups_.back();
  }
}

// After each batch is calculated, the counter of each group(group.pending_)
// and allreudce sequence counter(next_group_) will be cleaned up again.
void Reducer::PrepareForBackward() {
  VLOG(3) << "start reseting count..";
  next_group_ = 0;
  std::for_each(groups_.begin(), groups_.end(), [](Group &group) {
    group.pending_ = group.variable_indices_.size();
  });
}

// Add hook function to each leaf node. When the gradient of a leaf node is
// generated, if it is the sparse parameter, it will directly execute allreduce,
// if it is the dense parameter, it will execute three steps: 1,
// MarkVariableReady. Find the position of the corresponding group
// through var_index, share the gradient memory and the group dense_tensors,
// the group counter is reduced by 1. 2, MarkGroupReady: When the group
// counter is 0, it means that allreduce can be emitted, and
// concat + allreduce + split is emitted in turn according to next_group_.
// 3, FinalizeBackward: after the end, synchronize each stream.
void Reducer::AddDistHook(VariableWrapper *var_warpper, size_t var_index) {
  const auto &var_locator = variable_locators_[var_index];
  auto group_index = var_locator.group_index;
  auto &group = groups_[group_index];

  if (!has_rebuilt_group_) {
    rebuild_vars_.push_back(vars_[var_index]);
    rebuild_var_indices_.push_back(var_index);
  }

  if (!group.is_sparse_) {
    // Only dense_contents_ need memory copy
    MarkVariableReady(var_index, var_warpper);
  }
  if (--group.pending_ == 0) {
    // can start allreduce
    MarkGroupReady(group_index);
  }

  if (next_group_ == groups_.size()) {
    FinalizeBackward();
  }
}

void Reducer::MarkVariableReady(size_t var_index,
                                VariableWrapper *var_warpper) {
  const auto &var_locator = variable_locators_[var_index];
  auto group_index = var_locator.group_index;
  auto inside_group_index = var_locator.inside_group_index;
  auto &group = groups_[group_index];
  auto length = group.length_[inside_group_index];

  auto tensor = var_warpper->MutableVar()->GetMutable<framework::LoDTensor>();
  group.dense_tensors_[inside_group_index].ShareDataWith(*tensor).Resize(
      {static_cast<int64_t>(length)});
}

void Reducer::MarkGroupReady(size_t group_index) {
  if (group_index > next_group_) {
    VLOG(3) << "It will adjust the order of group in next batch automatically";
    return;
  }

  PADDLE_ENFORCE_CUDA_SUCCESS(
      cudaEventRecord(events_[group_index].get(), compute_stream_));
  PADDLE_ENFORCE_CUDA_SUCCESS(
      cudaStreamWaitEvent(comm_stream_, events_[group_index].get(), 0));

  for (; next_group_ < groups_.size() && groups_[next_group_].pending_ == 0;
       ++next_group_) {
    auto &group = groups_[next_group_];
    if (group.is_sparse_) {
      VLOG(3) << "sparse group [" << next_group_ << "] start allreduce...";
      parallel_ctx_->AllReduceByStream(*group.sparse_contents_,
                                       group.sparse_contents_, 0, false);
    } else {
      VLOG(3) << "dense group [" << next_group_ << "] start allreduce...";
      // Select common commstream to concat tensors
      // group.dense_tensors ---> group.dense_contents_
      group.ConcatTensors(*parallel_ctx_->GetDeviceContext(0));

      // Start allreduce
      parallel_ctx_->AllReduceByStream(group.dense_contents_,
                                       &(group.dense_contents_), 0, false);
      // Select common commstream to split tensors
      // group.dense_contents_ ---> group.dense_tensors
      group.SplitTensors(*parallel_ctx_->GetDeviceContext(0));
    }
  }
}

std::vector<std::vector<size_t>> Reducer::RebuildGruops() {
  std::reverse(rebuild_vars_.begin(), rebuild_vars_.end());
  std::reverse(rebuild_var_indices_.begin(), rebuild_var_indices_.end());
  auto rebuild_group_indices =
      AssignGroupBySize(rebuild_vars_, is_sparse_gradient_, group_size_limits_,
                        rebuild_var_indices_);
  has_rebuilt_group_ = true;
  rebuild_vars_.clear();
  rebuild_var_indices_.clear();
  std::reverse(rebuild_group_indices.begin(), rebuild_group_indices.end());
  return rebuild_group_indices;
}

void Reducer::FinalizeBackward() {
  PADDLE_ENFORCE_CUDA_SUCCESS(cudaEventRecord(comm_enent_.get(), comm_stream_));
  PADDLE_ENFORCE_CUDA_SUCCESS(
      cudaStreamWaitEvent(compute_stream_, comm_enent_.get(), 0));
  if (!has_rebuilt_group_) {
    VLOG(3) << "Start rebuilding the groups";
    auto rebuild_group_indices = RebuildGruops();
    auto rebuild_group_number = rebuild_group_indices.size();
    group_indices_ = std::move(rebuild_group_indices);
    CreateGroupEvents(rebuild_group_number);
    InitializeGroups(group_indices_);
  }
  VLOG(3) << "In the batch, Reducer is finished...";
}

// According to the size of each parameter, it is allocated to different groups.
// The sparse parameter occupies a group exclusively. The dense parameters of
// the same data type are assigned to the same group. When dividing groups, the
// size of each group will be limited according to each value in
// group_size_limits in turn. When it is not enough, it will be divided
// by the last value of group_size_limits. The limit value is 0, which
// means that the parameter will monopolize the group.
std::vector<std::vector<size_t>> AssignGroupBySize(
    const std::vector<std::shared_ptr<imperative::VarBase>> &vars,
    const std::vector<bool> &is_sparse_gradient,
    const std::vector<size_t> &group_size_limits,
    const std::vector<int64_t> &tensor_indices) {
  PADDLE_ENFORCE_EQ(vars.size(), is_sparse_gradient.size(),
                    platform::errors::PreconditionNotMet(
                        "vars len must be equal to is_sparse_gradient len, but "
                        "[%lu] != [%lu]",
                        vars.size(), is_sparse_gradient.size()));
  auto check_perm = [](const std::vector<int64_t> &x) -> bool {
    size_t len = x.size();
    std::vector<size_t> cnt(len, 0);
    for (size_t i = 0; i < len; ++i) {
      if (x[i] >= static_cast<int64_t>(len) || x[i] < 0 || cnt[x[i]]) {
        return false;
      }
      cnt[x[i]]++;
    }
    return true;
  };
  PADDLE_ENFORCE_EQ(true, check_perm(tensor_indices),
                    platform::errors::PreconditionNotMet(
                        "tensor_indices must be a permutation from 0 to %lu",
                        tensor_indices.size()));
  // the return vector
  std::vector<std::vector<size_t>> res;

  // Key: the var type
  // Value: should use which index in group_size_limits for group size limit
  std::unordered_map<std::string, size_t> group_limit_index;

  // Key: the var type
  // Value: <the var index in input tensors, total numel in this group>
  std::unordered_map<std::string, std::pair<std::vector<size_t>, size_t>>
      next_group;

  for (size_t i = 0; i < vars.size(); ++i) {
    const auto &var = vars[i];

    size_t tensor_real_index = i;
    if (!tensor_indices.empty()) {
      tensor_real_index = tensor_indices[i];
    }

    if (is_sparse_gradient[tensor_real_index]) {
      // we keep sparse var a single group
      res.push_back({tensor_real_index});
      continue;
    }

    const auto &var_dtype = var->DataType();
    const auto var_dtype_str = framework::DataTypeToString(var_dtype);
    VLOG(3) << "var[" << var->GradVarName() << "] 's type is "
            << var->DataType();
    auto &group_info = next_group[var_dtype_str];
    int64_t var_size = -1;
    if (var->Var().IsType<framework::LoDTensor>()) {
      var_size = var->Var().Get<framework::LoDTensor>().numel();
    } else {
      VLOG(3) << "var " << var->Name()
              << " is not tensor or selected_rows, so skip it";
      continue;
    }
    group_info.first.push_back(tensor_real_index);
    group_info.second += framework::SizeOfType(var_dtype) * var_size;

    if (group_limit_index.find(var_dtype_str) == group_limit_index.end()) {
      // means it is the first var of var_dtype
      group_limit_index[var_dtype_str] = 0;
    }
    auto &cur_limit_index = group_limit_index[var_dtype_str];
    if (group_info.second >= group_size_limits[cur_limit_index]) {
      // exceed group capacity and create a new group
      res.emplace_back(std::move(group_info.first));
      group_info = std::pair<std::vector<size_t>, size_t>();
      cur_limit_index =
          (std::min)(cur_limit_index + 1, group_size_limits.size() - 1);
    }
  }

  // add the final groups
  for (auto &e : next_group) {
    auto &group_info = e.second;
    if (!group_info.first.empty()) {
      res.emplace_back(std::move(group_info.first));
    }
  }

  for (const auto &group_index : res) {
    PADDLE_ENFORCE_NE(
        group_index.empty(), true,
        platform::errors::PreconditionNotMet(
            "AssignGroupBySize construct empty group, please check."));
  }
  if (tensor_indices.empty()) {
    std::sort(res.begin(), res.end(),
              [](const std::vector<size_t> &x, const std::vector<size_t> &y) {
                return x.front() < y.front();
              });
  }
  return res;
}
#endif

}  // namespace imperative
}  // namespace paddle