* 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 fmod_common_impl.h
* \brief
*/
#if !defined(__ASCENDC_INCLUDE_INTERNAL_HEADERS__)
#pragma message( \
"impl/adv_api/detail/math/fmod/fmod_common_impl.h is an internal header file and must not be used directly. Functions or variables defined in this file may be removed in the future. Please use \"#include \"adv_api/math/fmod.h\"\" and use public functions or variables defined in interface headers files.")
#define __ASCENDC_INCLUDE_INTERNAL_HEADERS__
#define __UNDEF_ASCENDC_INCLUDE_INTERNAL_HEADERS_MATH_FMOD_FMOD_COMMON_IMPL_H__
#endif
#ifndef IMPL_MATH_FMOD_FMOD_COMMON_IMPL_H
#define IMPL_MATH_FMOD_FMOD_COMMON_IMPL_H
#include "kernel_tensor.h"
#include "kernel_basic_intf.h"
#include "include/adv_api/math/trunc.h"
#include "../../common/check.h"
#ifdef ASCENDC_CPU_DEBUG
#include "../../api_check/kernel_check/math/fmod/fmod_check.h"
#endif
#include "../../api_check/kernel_api_check.h"
namespace AscendC {
namespace {
constexpr uint32_t SRC0_IDX = 1;
constexpr uint32_t SRC1_IDX = 2;
constexpr uint32_t TRUNC_IDX = 3;
}
__aicore__ inline void FmodCompute(
const LocalTensor<float>& dstTensor, const LocalTensor<float>& src0Tensor, const LocalTensor<float>& src1Tensor,
const LocalTensor<uint8_t>& sharedTmpBuffer, const uint32_t stackSize, const uint32_t calCount)
{
PipeBarrier<PIPE_V>();
Div(dstTensor, src0Tensor, src1Tensor, calCount);
PipeBarrier<PIPE_V>();
Trunc(dstTensor, dstTensor, sharedTmpBuffer, calCount);
PipeBarrier<PIPE_V>();
Mul(dstTensor, dstTensor, src1Tensor, calCount);
PipeBarrier<PIPE_V>();
Sub(dstTensor, src0Tensor, dstTensor, calCount);
PipeBarrier<PIPE_V>();
}
__aicore__ inline void FmodCompute(
const LocalTensor<half>& dstTensor, const LocalTensor<half>& src0Tensor, const LocalTensor<half>& src1Tensor,
const LocalTensor<uint8_t>& sharedTmpBuffer, const uint32_t stackSize, const uint32_t calCount)
{
LocalTensor<float> floatTmpTensor = sharedTmpBuffer.ReinterpretCast<float>();
LocalTensor<float> tmpSrc0 = floatTmpTensor[SRC0_IDX * stackSize];
LocalTensor<float> tmpSrc1 = floatTmpTensor[SRC1_IDX * stackSize];
PipeBarrier<PIPE_V>();
Cast<float, half>(tmpSrc0, src0Tensor, RoundMode::CAST_NONE, calCount);
Cast<float, half>(tmpSrc1, src1Tensor, RoundMode::CAST_NONE, calCount);
PipeBarrier<PIPE_V>();
FmodCompute(
floatTmpTensor, tmpSrc0, tmpSrc1, sharedTmpBuffer[TRUNC_IDX * stackSize * sizeof(float)], stackSize, calCount);
Cast<half, float>(dstTensor, floatTmpTensor, RoundMode::CAST_NONE, calCount);
PipeBarrier<PIPE_V>();
}
template <typename T, bool isReuseSource = false>
__aicore__ inline void FmodImpl(
const LocalTensor<T>& dstTensor, const LocalTensor<T>& src0Tensor, const LocalTensor<T>& src1Tensor,
const LocalTensor<uint8_t>& sharedTmpBuffer, const uint32_t calCount)
{
if ASCEND_IS_AIC {
return;
}
CHECK_FUNC_HIGHLEVEL_API(Fmod, (T, isReuseSource), (dstTensor, src0Tensor, src1Tensor, sharedTmpBuffer, calCount));
if constexpr (sizeof(T) == sizeof(float)) {
FmodCompute(dstTensor, src0Tensor, src1Tensor, sharedTmpBuffer, src0Tensor.GetSize(), calCount);
return;
}
constexpr uint32_t maxLiveNodeCount =
8;
uint32_t bufferSize = sharedTmpBuffer.GetSize();
uint32_t stackSize =
bufferSize / sizeof(T) / maxLiveNodeCount / ONE_BLK_SIZE * ONE_BLK_SIZE;
CheckTmpBufferSize(stackSize, 0, bufferSize);
stackSize = stackSize > src0Tensor.GetSize() ? src0Tensor.GetSize() : stackSize;
const uint32_t round = calCount / stackSize;
const uint32_t tail = calCount % stackSize;
for (uint32_t i = 0; i < round; ++i) {
FmodCompute(
dstTensor[i * stackSize], src0Tensor[i * stackSize], src1Tensor[i * stackSize], sharedTmpBuffer, stackSize,
stackSize);
}
if (tail > 0) {
FmodCompute(
dstTensor[round * stackSize], src0Tensor[round * stackSize], src1Tensor[round * stackSize], sharedTmpBuffer,
stackSize, tail);
}
}
}
#endif
#if defined(__UNDEF_ASCENDC_INCLUDE_INTERNAL_HEADERS_MATH_FMOD_FMOD_COMMON_IMPL_H__)
#undef __ASCENDC_INCLUDE_INTERNAL_HEADERS__
#undef __UNDEF_ASCENDC_INCLUDE_INTERNAL_HEADERS_MATH_FMOD_FMOD_COMMON_IMPL_H__
#endif