chromium/third_party/gemmlowp/src/internal/unpack.h

// Copyright 2015 The Gemmlowp 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.

// unpack.h: unpacking the result blocks computed by compute.h,
// storing them into the destination matrix.

#ifndef GEMMLOWP_INTERNAL_UNPACK_H_
#define GEMMLOWP_INTERNAL_UNPACK_H_

#include "allocator.h"
#include "block_params.h"
#include "output.h"
#include "pack.h"

#include <cmath>

namespace gemmlowp {

class PackedResult { … };

struct MatrixBlockBounds { … };

template <int Rows, int Cols, typename SrcMapType>
void PrefetchResultBlock(const SrcMapType& src,
                         const VectorMap<const std::int32_t, VectorShape::Col>&
                             lhs_sums_of_each_slice,
                         int src_row, int src_col) { … }

template <typename KernelFormat, typename RegisterBlockType,
          typename SrcMapType, typename LhsOffset, typename RhsOffset,
          typename OutputPipelineExecutorType, typename DstType>
void UnpackResultBlock(const SrcMapType& src,
                       const OutputPipelineExecutorType& executor, DstType* dst,
                       const VectorMap<const std::int32_t, VectorShape::Col>&
                           lhs_sums_of_each_slice,
                       const VectorMap<const std::int32_t, VectorShape::Row>&
                           rhs_sums_of_each_slice,
                       const LhsOffset& lhs_offset, const RhsOffset& rhs_offset,
                       int depth, int src_row, int src_col, int src_global_row,
                       int src_global_col, int dst_row, int dst_col) { … }

template <typename KernelFormat, typename ResultBlockType,
          typename PackedResultType, typename LhsOffset, typename RhsOffset,
          typename OutputPipelineType>
void UnpackResult(ResultBlockType* dst, const MatrixBlockBounds& dst_block,
                  const PackedResultType& src, int depth,
                  const std::int32_t* lhs_sums_of_each_slice_ptr,
                  const std::int32_t* rhs_sums_of_each_slice_ptr,
                  const LhsOffset& lhs_offset, const RhsOffset& rhs_offset,
                  const OutputPipelineType& output_pipeline) { … }

}  // end namespace gemmlowp

#endif  // GEMMLOWP_INTERNAL_UNPACK_H_