binary_eval.h

Go to the documentation of this file.

/*
 * This file is a part of TiledArray.
 * Copyright (C) 2013  Virginia Tech
 *
 *  This program is free software: you can redistribute it and/or modify
 *  it under the terms of the GNU General Public License as published by
 *  the Free Software Foundation, either version 3 of the License, or
 *   (at your option) any later version.
 *
 *  This program is distributed in the hope that it will be useful,
 *  but WITHOUT ANY WARRANTY; without even the implied warranty of
 *  MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
 *  GNU General Public License for more details.
 *
 *  You should have received a copy of the GNU General Public License
 *  along with this program.  If not, see <http://www.gnu.org/licenses/>.
 *
 */
 
#ifndef TILEDARRAY_DIST_EVAL_BINARY_EVAL_H__INCLUDED
#define TILEDARRAY_DIST_EVAL_BINARY_EVAL_H__INCLUDED
 
#include <TiledArray/dist_eval/dist_eval.h>
#include <TiledArray/zero_tensor.h>
 
#include <TiledArray/tensor/type_traits.h>
 
namespace TiledArray {
namespace detail {
 
 
template <typename Left, typename Right, typename Op, typename Policy>
class BinaryEvalImpl : public DistEvalImpl<typename Op::result_type, Policy>,
                       public std::enable_shared_from_this<
                           BinaryEvalImpl<Left, Right, Op, Policy>> {
 public:
  typedef BinaryEvalImpl<Left, Right, Op, Policy>
      BinaryEvalImpl_;  
  typedef DistEvalImpl<typename Op::result_type, Policy>
      DistEvalImpl_;  
  typedef typename DistEvalImpl_::TensorImpl_
      TensorImpl_;           
  typedef Left left_type;    
  typedef Right right_type;  
  typedef typename DistEvalImpl_::ordinal_type ordinal_type;  
  typedef typename DistEvalImpl_::range_type range_type;      
  typedef typename DistEvalImpl_::shape_type shape_type;      
  typedef typename DistEvalImpl_::pmap_interface
      pmap_interface;  
  typedef
      typename DistEvalImpl_::trange_type trange_type;    
  typedef typename DistEvalImpl_::value_type value_type;  
  typedef
      typename DistEvalImpl_::eval_type eval_type;  
  typedef Op op_type;  
 
  using std::enable_shared_from_this<BinaryEvalImpl_>::shared_from_this;
 
 private:
  left_type left_;    
  right_type right_;  
  op_type op_;        
 
 public:
 
  template <typename Perm, typename = std::enable_if_t<
                               TiledArray::detail::is_permutation_v<Perm>>>
  BinaryEvalImpl(const left_type& left, const right_type& right, World& world,
                 const trange_type& trange, const shape_type& shape,
                 const std::shared_ptr<pmap_interface>& pmap, const Perm& perm,
                 const op_type& op)
      : DistEvalImpl_(world, trange, shape, pmap, outer(perm)),
        left_(left),
        right_(right),
        op_(op) {
    TA_ASSERT(left.trange() == right.trange());
  }
 
  virtual ~BinaryEvalImpl() {}
 
 
  virtual Future<value_type> get_tile(ordinal_type i) const {
    TA_ASSERT(TensorImpl_::is_local(i));
    TA_ASSERT(!TensorImpl_::is_zero(i));
 
    const auto source_index = DistEvalImpl_::perm_index_to_source(i);
    const ProcessID source =
        left_.owner(source_index);  // Left and right
                                    // should have the same owner
 
    const madness::DistributedID key(DistEvalImpl_::id(), i);
    return TensorImpl_::world().gop.template recv<value_type>(source, key);
  }
 
 
  virtual void discard_tile(ordinal_type i) const { get_tile(i); }
 
 private:
 
#ifdef TILEDARRAY_HAS_CUDA
  template <typename L, typename R, typename U = value_type>
  std::enable_if_t<!detail::is_cuda_tile_v<U>, void> eval_tile(
      const ordinal_type i, L left, R right) {
    DistEvalImpl_::set_tile(i, op_(left, right));
  }
 
  template <typename L, typename R, typename U = value_type>
  std::enable_if_t<detail::is_cuda_tile_v<U>, void> eval_tile(
      const ordinal_type i, L left, R right) {
    // TODO avoid copy the Op object
    auto result_tile =
        madness::add_cuda_task(DistEvalImpl_::world(), op_, left, right);
    DistEvalImpl_::set_tile(i, result_tile);
  }
#else
  template <typename L, typename R>
  void eval_tile(const ordinal_type i, L left, R right) {
    DistEvalImpl_::set_tile(i, op_(left, right));
  }
#endif
 
  virtual int internal_eval() {
    // Evaluate child tensors
    left_.eval();
    right_.eval();
 
    // Task function argument types
    typedef typename std::conditional<
        op_type::left_is_consumable, typename left_type::value_type,
        const typename left_type::value_type>::type& left_argument_type;
    typedef typename std::conditional<
        op_type::right_is_consumable, typename right_type::value_type,
        const typename right_type::value_type>::type& right_argument_type;
 
    ordinal_type task_count = 0ul;
 
    // Construct local iterator
    TA_ASSERT(left_.pmap() == right_.pmap());
    std::shared_ptr<BinaryEvalImpl_> self = shared_from_this();
    typename pmap_interface::const_iterator it = left_.pmap()->begin();
    const typename pmap_interface::const_iterator end = left_.pmap()->end();
 
    if (left_.is_dense() && right_.is_dense() && TensorImpl_::is_dense()) {
      // Evaluate tiles where both arguments and the result are dense
      for (; it != end; ++it) {
        // Get tile indices
        const auto source_index = *it;
        const auto target_index =
            DistEvalImpl_::perm_index_to_target(source_index);
 
        // Schedule tile evaluation task
        TensorImpl_::world().taskq.add(
            self,
            &BinaryEvalImpl_::template eval_tile<left_argument_type,
                                                 right_argument_type>,
            target_index, left_.get(source_index), right_.get(source_index));
 
        ++task_count;
      }
    } else {
      // Evaluate tiles where the result or one of the arguments is sparse
      for (; it != end; ++it) {
        // Get tile indices
        const auto index = *it;
        const auto target_index = DistEvalImpl_::perm_index_to_target(index);
 
        if (!TensorImpl_::is_zero(target_index)) {
          // Schedule tile evaluation task
          if (left_.is_zero(index)) {
            TensorImpl_::world().taskq.add(
                self,
                &BinaryEvalImpl_::template eval_tile<const ZeroTensor,
                                                     right_argument_type>,
                target_index, ZeroTensor(), right_.get(index));
          } else if (right_.is_zero(index)) {
            TensorImpl_::world().taskq.add(
                self,
                &BinaryEvalImpl_::template eval_tile<left_argument_type,
                                                     const ZeroTensor>,
                target_index, left_.get(index), ZeroTensor());
          } else {
            TensorImpl_::world().taskq.add(
                self,
                &BinaryEvalImpl_::template eval_tile<left_argument_type,
                                                     right_argument_type>,
                target_index, left_.get(index), right_.get(index));
          }
 
          ++task_count;
        } else {
          // Cleanup unused tiles
          if (!left_.is_zero(index)) left_.discard(index);
          if (!right_.is_zero(index)) right_.discard(index);
        }
      }
    }
 
    // Wait for child tensors to be evaluated, and process tasks while waiting.
    left_.wait();
    right_.wait();
 
    return task_count;
  }
 
};  // class BinaryEvalImpl
 
}  // namespace detail
}  // namespace TiledArray
 
#endif  // TILEDARRAY_DIST_EVAL_BINARY_EVAL_H__INCLUDED