* 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 kernel_operator_vec_bilinearinterpolation_intf_impl.h
* \brief
*/
#ifndef ASCENDC_MODULE_OPERATOR_VEC_BILINEARINTERPOLATION_INTERFACE_IMPL_H
#define ASCENDC_MODULE_OPERATOR_VEC_BILINEARINTERPOLATION_INTERFACE_IMPL_H
#include "kernel_tensor.h"
#include "kernel_check.h"
#if __NPU_ARCH__ == 1001
#include "dav_c100/kernel_operator_vec_bilinearinterpolation_impl.h"
#elif __NPU_ARCH__ == 2002
#include "dav_m200/kernel_operator_vec_bilinearinterpolation_impl.h"
#elif __NPU_ARCH__ == 2201
#include "dav_c220/kernel_operator_vec_bilinearinterpolation_impl.h"
#elif __NPU_ARCH__ == 3002
#include "dav_m300/kernel_operator_vec_bilinearinterpolation_impl.h"
#elif __NPU_ARCH__ == 3102
#include "dav_m310/kernel_operator_vec_bilinearinterpolation_impl.h"
#elif __NPU_ARCH__ == 3101
#include "dav_c310/kernel_operator_vec_bilinearinterpolation_impl.h"
#elif (__NPU_ARCH__ == 5102)
#include "dav_m510/kernel_operator_vec_bilinearinterpolation_impl.h"
#elif __NPU_ARCH__ == 3113
#include "dav_l311/kernel_operator_vec_bilinearinterpolation_impl.h"
#endif
#pragma begin_pipe(V)
namespace AscendC {
template <typename T>
__aicore__ inline void BilinearInterpolationCheck(const LocalTensor<T> &src0, const LocalTensor<uint32_t> &src0Offset,
const LocalTensor<T> &src1, uint16_t vROffset, const LocalTensor<uint8_t> &sharedTmpBuffer)
{
#if ASCENDC_CPU_DEBUG
uint32_t sharedTmpBufferSize = sharedTmpBuffer.GetSize();
#if __NPU_ARCH__ == 2201
uint32_t expectedTmpBufferSize = (src0.GetSize() + src1.GetSize()) * 32;
#elif (__NPU_ARCH__ == 3101) || (__NPU_ARCH__ == 5102)
uint32_t expectedTmpBufferSize = 0;
#else
uint32_t expectedTmpBufferSize = src0Offset.GetSize() * sizeof(uint32_t);
#endif
ASCENDC_ASSERT((sharedTmpBufferSize >= expectedTmpBufferSize), { KERNEL_LOG(KERNEL_ERROR, "Failed to check "
"sharedTmpBuffer size in BilinearInterpolation, its expected size is at least %u, current size is %u",
expectedTmpBufferSize, sharedTmpBufferSize);});
#endif
ASCENDC_ASSERT((SupportType<PrimT<T>, half>()), {KERNEL_LOG(KERNEL_ERROR,
"Failed to check dtype in BilinearInterpolation,"
" current api support dtype combination is src: half, dst: half");});
ASCENDC_CHECK_VALUE_RANGE(vROffset, 128, UINT16_MAX, "vROffset", "BilinearInterpolation");
ASCENDC_ASSERT((vROffset % AscendCUtils::GetC0Count(sizeof(PrimT<T>)) == 0), { KERNEL_LOG(KERNEL_ERROR,
"Failed to check vROffset value in BilinearInterpolation, vROffset * sizeof(T) must be 32B aligned, "
"current vROffset value is %u. In NPU mode, no error is reported. The value is rounded down by 32B.",
vROffset); });
}
template <typename T>
__aicore__ inline void BilinearInterpolation(const LocalTensor<T> &dst, const LocalTensor<T> &src0,
const LocalTensor<uint32_t> &src0Offset, const LocalTensor<T> &src1, uint64_t mask, uint8_t hRepeat,
bool repeatMode, uint16_t dstBlkStride, uint16_t vROffset, uint8_t vRepeat,
const LocalTensor<uint8_t> &sharedTmpBuffer)
{
#if ASCENDC_CPU_DEBUG
if (!CheckFuncBilinearInterpolation(dst, src0, src0Offset, src1, mask, hRepeat, repeatMode,
dstBlkStride, vROffset, vRepeat, "BilinearInterpolation")) {
ASCENDC_REPORT_CHECK_ERROR("BilinearInterpolation", KernelFuncType::MASK_COUNT_MODE);
}
#endif
BilinearInterpolationCheck(src0, src0Offset, src1, vROffset, sharedTmpBuffer);
BilinearInterpolationCalc(dst, src0, src0Offset, src1, mask, hRepeat,
repeatMode, dstBlkStride, vROffset, vRepeat, sharedTmpBuffer);
}
template <typename T>
__aicore__ inline void BilinearInterpolation(const LocalTensor<T> &dst, const LocalTensor<T> &src0,
const LocalTensor<uint32_t> &src0Offset, const LocalTensor<T> &src1, uint64_t mask[], uint8_t hRepeat,
bool repeatMode, uint16_t dstBlkStride, uint16_t vROffset, uint8_t vRepeat,
const LocalTensor<uint8_t> &sharedTmpBuffer)
{
#if ASCENDC_CPU_DEBUG
if (!CheckFuncBilinearInterpolation(dst, src0, src0Offset, src1, mask, hRepeat, repeatMode,
dstBlkStride, vROffset, vRepeat, "BilinearInterpolation")) {
ASCENDC_REPORT_CHECK_ERROR("BilinearInterpolation", KernelFuncType::MASK_BIT_MODE);
}
#endif
BilinearInterpolationCheck(src0, src0Offset, src1, vROffset, sharedTmpBuffer);
BilinearInterpolationCalc(dst, src0, src0Offset, src1, mask, hRepeat,
repeatMode, dstBlkStride, vROffset, vRepeat, sharedTmpBuffer);
}
}
#pragma end_pipe
#endif
