/**
 * Copyright (c) 2025 Huawei Technologies Co., Ltd.
 * This program is free software, you can redistribute it and/or modify it under the terms and conditions of
 * CANN Open Software License Agreement Version 2.0 (the "License").
 * Please refer to the License for details. You may not use this file except in compliance with the License.
 * THIS SOFTWARE IS PROVIDED ON AN "AS IS" BASIS, WITHOUT WARRANTIES OF ANY KIND, EITHER EXPRESS OR IMPLIED,
 * INCLUDING BUT NOT LIMITED TO NON-INFRINGEMENT, MERCHANTABILITY, OR FITNESS FOR A PARTICULAR PURPOSE.
 * See LICENSE in the root of the software repository for the full text of the License.
 */

/*!
 * \file vector_dup.h
 * \brief
 */

#ifndef TILEOP_TILE_OPERATOR_VECTOR_DUP__H
#define TILEOP_TILE_OPERATOR_VECTOR_DUP__H
#include "utils/layout.h"
#include "utils/tile_tensor.h"

#define OP_TILE_OP_VEC_DUP TVecDup
template <typename Scalar, typename T0>
TILEOP void TVecDup(T0 dst, Scalar src)
{
    constexpr auto shapeSize = Std::tuple_size<typename T0::Shape>::value;
    constexpr size_t expectSize = 5;
    const auto dstLayout = dst.GetLayout();
    using T1 = std::conditional_t<std::is_same_v<typename T0::Type, bool>, uint8_t, typename T0::Type>;
    auto shape0 = dstLayout.template GetShapeDim<0, expectSize>();
    auto shape1 = dstLayout.template GetShapeDim<1, expectSize>();
    auto shape2 = dstLayout.template GetShapeDim<2, expectSize>();
    auto shape3 = dstLayout.template GetShapeDim<3, expectSize>();
    auto shape4 = dstLayout.template GetShapeDim<4, expectSize>();
    auto dstStride0 = dstLayout.template GetStrideDim<0, expectSize>();
    auto dstStride1 = dstLayout.template GetStrideDim<1, expectSize>();
    auto dstStride2 = dstLayout.template GetStrideDim<2, expectSize>();
    constexpr auto dstTileH = TileOp::GetTensorTileShapeDim<T0, 3, 5>();
    constexpr auto dstTileW = TileOp::GetTensorTileShapeDim<T0, 4, 5>();
    constexpr auto dstTypeSize = sizeof(typename T0::Type);
    for (LoopVar n0Index = 0; n0Index < shape0; ++n0Index) {
        for (LoopVar n1Index = 0; n1Index < shape1; ++n1Index) {
            for (LoopVar n2Index = 0; n2Index < shape2; ++n2Index) {
                using TileDefineDst = pto::Tile<
                    pto::TileType::Vec, T1, dstTileH, dstTileW, pto::BLayout::RowMajor, -1, -1>;
                TileDefineDst dstTile(shape3, shape4);
                auto dstOffset = n0Index * dstStride0 + n1Index * dstStride1 + n2Index * dstStride2;
                pto::TASSIGN(dstTile, (uint64_t)(dst.GetAddr() + dstOffset * dstTypeSize));
                pto::TEXPANDS(dstTile, src);
            }
        }
    }
}
#endif