* 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_sync_impl.h
* \brief
*/
#if !defined(__ASCENDC_INCLUDE_INTERNAL_HEADERS__)
#pragma message("impl/basic_api/dav_c100/kernel_operator_sync_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 \"basic_api/kernel_operator_intf.h\"\" and use public functions or variables defined in interface headers files.")
#define __ASCENDC_INCLUDE_INTERNAL_HEADERS__
#define __UNDEF_ASCENDC_INCLUDE_INTERNAL_HEADERS_KERNEL_OPERATOR_SYNC_IMPL_H__
#endif
#ifndef ASCENDC_MODULE_OPERATOR_SYNC_IMPL_H
#define ASCENDC_MODULE_OPERATOR_SYNC_IMPL_H
namespace AscendC {
__aicore__ inline void ClcSyncCount(__gm__ int32_t* localSyncGM, __ubuf__ int32_t* ubWorkspaceAddr,
const int32_t blockIdx, const int32_t totalBlocks, bool isFirst, int32_t& count)
{
if (isFirst) {
__ubuf__ int32_t* localUbAddr = ubWorkspaceAddr + (blockIdx * DEFAULT_BLK_NUM);
*(reinterpret_cast<__ubuf__ int32_t*>(localUbAddr)) = 1;
event_t eventIdSToMte3 = static_cast<event_t>(GetTPipePtr()->FetchEventID(HardEvent::S_MTE3));
SetFlag<HardEvent::S_MTE3>(eventIdSToMte3);
WaitFlag<HardEvent::S_MTE3>(eventIdSToMte3);
copy_ubuf_to_gm(static_cast<__gm__ void*>(localSyncGM), static_cast<__ubuf__ void*>(localUbAddr), 0, 1, 1, 0,
0);
event_t eventIDMTE3ToMTE2 = static_cast<event_t>(GetTPipePtr()->FetchEventID(HardEvent::MTE3_MTE2));
SetFlag<HardEvent::MTE3_MTE2>(eventIDMTE3ToMTE2);
WaitFlag<HardEvent::MTE3_MTE2>(eventIDMTE3ToMTE2);
count = 1;
for (int32_t i = 0; i < totalBlocks; i++) {
if (i != blockIdx) {
count += *(reinterpret_cast<__ubuf__ int32_t*>(ubWorkspaceAddr) + i * ONE_BLK_FLOAT_NUM);
}
}
} else {
for (int32_t i = 0; i < totalBlocks; i++) {
count += *(reinterpret_cast<__ubuf__ int32_t*>(ubWorkspaceAddr) + i * ONE_BLK_FLOAT_NUM);
}
}
}
template <bool isAIVOnly = true>
__aicore__ inline void SoftSyncAllImpl(__gm__ int32_t* gmWorkspaceAddr, __ubuf__ int32_t* ubWorkspaceAddr,
const int usedCores)
{
PipeBarrier<PIPE_ALL>();
int32_t totalBlocks = isAIVOnly ? get_block_num() : (GetTaskRationImpl() * get_block_num());
totalBlocks = usedCores != 0 ? usedCores : totalBlocks;
int32_t blockIdx = isAIVOnly ? get_block_idx() : GetBlockIdxImpl();
__gm__ int32_t* localSyncGM = gmWorkspaceAddr + (blockIdx * DEFAULT_BLK_NUM);
__ubuf__ int32_t* localUbAddr = ubWorkspaceAddr + (blockIdx * DEFAULT_BLK_NUM);
copy_gm_to_ubuf(static_cast<__ubuf__ void*>(localUbAddr), static_cast<__gm__ void*>(localSyncGM), 0, 1, 1, 0, 0);
event_t eventIdMte2ToS = static_cast<event_t>(GetTPipePtr()->FetchEventID(HardEvent::MTE2_S));
SetFlag<HardEvent::MTE2_S>(eventIdMte2ToS);
WaitFlag<HardEvent::MTE2_S>(eventIdMte2ToS);
int32_t curValue = *(reinterpret_cast<__ubuf__ int32_t*>(localUbAddr)) + 1;
bool isFirst = curValue == 1 ? true : false;
*(reinterpret_cast<__ubuf__ int32_t*>(localUbAddr)) = curValue;
event_t eventIdSToMte3 = static_cast<event_t>(GetTPipePtr()->FetchEventID(HardEvent::S_MTE3));
SetFlag<HardEvent::S_MTE3>(eventIdSToMte3);
WaitFlag<HardEvent::S_MTE3>(eventIdSToMte3);
copy_ubuf_to_gm(static_cast<__gm__ void*>(localSyncGM), static_cast<__ubuf__ void*>(localUbAddr), 0, 1, 1, 0, 0);
event_t eventIDMTE3ToMTE2 = static_cast<event_t>(GetTPipePtr()->FetchEventID(HardEvent::MTE3_MTE2));
SetFlag<HardEvent::MTE3_MTE2>(eventIDMTE3ToMTE2);
WaitFlag<HardEvent::MTE3_MTE2>(eventIDMTE3ToMTE2);
int32_t totalBlockCount = ONE_BLK_FLOAT_NUM * totalBlocks;
uint16_t blockLen = totalBlockCount / AscendCUtils::GetC0Count(sizeof(int32_t));
while (true) {
copy_gm_to_ubuf(static_cast<__ubuf__ void*>(ubWorkspaceAddr), static_cast<__gm__ void*>(gmWorkspaceAddr), 0, 1,
blockLen, 0, 0);
SetFlag<HardEvent::MTE2_S>(eventIdMte2ToS);
WaitFlag<HardEvent::MTE2_S>(eventIdMte2ToS);
int32_t count = 0;
ClcSyncCount(localSyncGM, ubWorkspaceAddr, blockIdx, totalBlocks, isFirst, count);
event_t eventIdSToMte2 = static_cast<event_t>(GetTPipePtr()->FetchEventID(HardEvent::S_MTE2));
SetFlag<HardEvent::S_MTE2>(eventIdSToMte2);
WaitFlag<HardEvent::S_MTE2>(eventIdSToMte2);
if (count >= (totalBlocks * curValue)) {
break;
}
}
}
template<pipe_t AIV_PIPE = PIPE_MTE3, pipe_t AIC_PIPE = PIPE_MTE3, bool FORCE = false>
__aicore__ inline void SetNextTaskStartImpl()
{
ASSERT(0 && "unsupported SetNextTaskStart on current device");
}
template<int8_t earlyStartConfig = -1, bool FORCE = false>
__aicore__ inline void WaitPreTaskEndImpl()
{
ASSERT(0 && "unsupported WaitPreTaskEnd on current device");
}
template<bool isAIVOnly = true>
__aicore__ inline void SyncAllImpl()
{
ASSERT(0 && "unsupported hardware syncAll on current device");
}
template <uint8_t modeId, pipe_t pipe>
__aicore__ inline void NotifyEventImpl(uint16_t flagId)
{
ASCENDC_ASSERT((false), "CrossCoreSetFlag is not supported on current device");
}
template <uint8_t modeId, pipe_t pipe>
__aicore__ inline void WaitEventImpl(uint16_t flagId)
{
ASCENDC_ASSERT((false), "CrossCoreWaitFlag is not supported on current device");
}
__aicore__ inline void SetSyncBaseAddr(uint64_t config)
{
ASCENDC_ASSERT((false), { KERNEL_LOG(KERNEL_ERROR, "SetSyncBaseAddr is not supported on current device"); });
}
}
#endif
#if defined(__UNDEF_ASCENDC_INCLUDE_INTERNAL_HEADERS_KERNEL_OPERATOR_SYNC_IMPL_H__)
#undef __ASCENDC_INCLUDE_INTERNAL_HEADERS__
#undef __UNDEF_ASCENDC_INCLUDE_INTERNAL_HEADERS_KERNEL_OPERATOR_SYNC_IMPL_H__
#endif
