TensorMultiplyAdd.hpp

Go to the documentation of this file.

 
//~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~//
// Copyright (c) Lawrence Livermore National Security, LLC and other
// RAJA Project Developers. See top-level LICENSE and COPYRIGHT
// files for dates and other details. No copyright assignment is required
// to contribute to RAJA.
//
// SPDX-License-Identifier: (BSD-3-Clause)
//~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~//
 
#ifndef RAJA_pattern_tensor_ET_TensorMultiplyAddAdd_HPP
#define RAJA_pattern_tensor_ET_TensorMultiplyAddAdd_HPP
 
#include "RAJA/config.hpp"
 
#include "RAJA/util/macros.hpp"
 
#include "RAJA/pattern/tensor/internal/ET/ExpressionTemplateBase.hpp"
#include "RAJA/pattern/tensor/internal/ET/MultiplyOperator.hpp"
 
namespace RAJA
{
namespace internal
{
namespace expt
{
 
 
namespace ET
{
 
 
template<typename LEFT_OPERAND_TYPE,
         typename RIGHT_OPERAND_TYPE,
         typename ADD_OPERAND_TYPE>
class TensorMultiplyAdd
    : public TensorExpressionBase<TensorMultiplyAdd<LEFT_OPERAND_TYPE,
                                                    RIGHT_OPERAND_TYPE,
                                                    ADD_OPERAND_TYPE>>
{
public:
  using self_type          = TensorMultiplyAdd<LEFT_OPERAND_TYPE,
                                               RIGHT_OPERAND_TYPE,
                                               ADD_OPERAND_TYPE>;
  using left_operand_type  = LEFT_OPERAND_TYPE;
  using right_operand_type = RIGHT_OPERAND_TYPE;
  using add_operand_type   = ADD_OPERAND_TYPE;
  using multiply_op = MultiplyOperator<LEFT_OPERAND_TYPE, RIGHT_OPERAND_TYPE>;
 
  using element_type = typename LEFT_OPERAND_TYPE::element_type;
  using index_type   = typename LEFT_OPERAND_TYPE::index_type;
 
  using result_type                       = typename multiply_op::result_type;
  static constexpr camp::idx_t s_num_dims = multiply_op::s_num_dims;
 
private:
  left_operand_type m_left_operand;
  right_operand_type m_right_operand;
  add_operand_type m_add_operand;
 
public:
  RAJA_INLINE
 
  RAJA_HOST_DEVICE
  TensorMultiplyAdd(left_operand_type const& left_operand,
                    right_operand_type const& right_operand,
                    add_operand_type const& add_operand)
      : m_left_operand {left_operand},
        m_right_operand {right_operand},
        m_add_operand {add_operand}
  {}
 
  template<typename TILE_TYPE>
  RAJA_INLINE RAJA_HOST_DEVICE auto eval(TILE_TYPE const& tile) const
      -> decltype(multiply_op::multiply_add(tile,
                                            m_left_operand,
                                            m_right_operand,
                                            m_add_operand))
  {
    return multiply_op::multiply_add(tile, m_left_operand, m_right_operand,
                                     m_add_operand);
  }
 
  RAJA_INLINE
 
  RAJA_HOST_DEVICE
  void print_ast() const
  {
    printf("MultiplyAdd[");
    multiply_op::print_ast();
    printf("](");
    m_left_operand.print_ast();
    printf(", ");
    m_right_operand.print_ast();
    printf(", ");
    m_add_operand.print_ast();
    printf(")");
  }
};
 
 
}  // namespace ET
 
}  // namespace expt
}  // namespace internal
 
}  // namespace RAJA
 
 
#endif