Line data    Source code

       1              : /**
       2              :  * Copyright (c) 2025 Huawei Technologies Co., Ltd.
       3              :  * This program is free software, you can redistribute it and/or modify it under the terms and conditions of
       4              :  * CANN Open Software License Agreement Version 2.0 (the "License").
       5              :  * Please refer to the License for details. You may not use this file except in compliance with the License.
       6              :  * THIS SOFTWARE IS PROVIDED ON AN "AS IS" BASIS, WITHOUT WARRANTIES OF ANY KIND, EITHER EXPRESS OR IMPLIED,
       7              :  * INCLUDING BUT NOT LIMITED TO NON-INFRINGEMENT, MERCHANTABILITY, OR FITNESS FOR A PARTICULAR PURPOSE.
       8              :  * See LICENSE in the root of the software repository for the full text of the License.
       9              :  */
      10              : 
      11              : #include "comm_halving_doubling.h"
      12              : 
      13              : namespace hccl {
      14            0 : CommHalvingDoubling::CommHalvingDoubling(const std::string &collectiveId,
      15              :                                          const u32 userRank, const u32 userRankSize,
      16              :                                          const u32 rank, const u32 rankSize, const TopoType topoFlag,
      17              :                                          const HcclDispatcher dispatcher,
      18              :                                          const std::unique_ptr<NotifyPool> &notifyPool,
      19              :                                          std::map<HcclIpAddress, HcclNetDevCtx> &netDevCtxMap,
      20              :                                          const IntraExchanger &exchanger,
      21              :                                          const std::vector<RankInfo> paraVector,
      22              :                                          const DeviceMem& inputMem, const DeviceMem& outputMem,
      23              :                                          const bool isUsedRdmaLevel0,
      24              :                                          const std::string &tag, const NICDeployment nicDeployInner,
      25              :                                          const u32 subUserRankRoot, HalvingDoublingType halvingDoublingType,
      26            0 :                                          const bool isHaveCpuRank, const bool useSuperPodMode)
      27              :     : CommBase(collectiveId, userRank, userRankSize, rank, rankSize, paraVector, topoFlag,
      28              :                dispatcher, notifyPool, netDevCtxMap, exchanger, inputMem, outputMem, isUsedRdmaLevel0,
      29              :                tag, nicDeployInner, false, false, false,
      30              :                INVALID_UINT, isHaveCpuRank, useSuperPodMode),
      31            0 :       subUserRankRoot_(subUserRankRoot), halvingDoublingType_(halvingDoublingType)
      32              : {
      33            0 : }
      34              : 
      35            0 : CommHalvingDoubling::~CommHalvingDoubling()
      36              : {
      37            0 : }
      38              : 
      39            0 : HcclResult CommHalvingDoubling::CalcLink()
      40              : {
      41            0 :     u32 subRoot = INVALID_VALUE_RANKID;
      42            0 :     HcclResult ret = HCCL_SUCCESS;
      43            0 :     if (subUserRankRoot_ != INVALID_VALUE_RANKID) {
      44            0 :         CHK_RET(GetRankByUserRank(subUserRankRoot_, subRoot));
      45              :     }
      46              : 
      47            0 :     std::vector<bool> linkRelation = AlgTemplateBase::CalcLinksRelation(rank_, rankSize_, subRoot,
      48            0 :                                                                      halvingDoublingType_);
      49            0 :     if (linkRelation.size() == 0) {
      50            0 :         HCCL_ERROR("[Calc][Link]comm halving doubling calculate link relation failed!");
      51            0 :         return HCCL_E_INTERNAL;
      52              :     }
      53              : 
      54            0 :     for (u32 dstRank = 0; dstRank < rankSize_; dstRank++) {
      55            0 :         if (!linkRelation[dstRank]) {
      56            0 :             HCCL_DEBUG("comm halving doubling rank[%u] don't need create link with dst_rank[%u]", rank_, dstRank);
      57            0 :             continue;
      58              :         }
      59              : 
      60            0 :         if (rank_ < dstRank) {
      61            0 :             ret = CalcLinksNum(MachineType::MACHINE_SERVER_TYPE, dstRank);
      62            0 :             CHK_PRT_RET(ret != HCCL_SUCCESS,
      63              :                 HCCL_ERROR("[Calc][Link]comm halving doubling calc links num failed, type[%d], "\
      64              :                 "dstRank[%u]", static_cast<int32_t>(MachineType::MACHINE_SERVER_TYPE), dstRank), ret);
      65              :         }
      66              : 
      67            0 :         if (rank_ > dstRank) {
      68            0 :             ret = CalcLinksNum(MachineType::MACHINE_CLIENT_TYPE, dstRank);
      69            0 :             CHK_PRT_RET(ret != HCCL_SUCCESS,
      70              :                 HCCL_ERROR("[Calc][Link]comm halving doubling calc links num failed, type[%d], "\
      71              :                 "dstRank[%u]", static_cast<int32_t>(MachineType::MACHINE_CLIENT_TYPE), dstRank), ret);
      72              :         }
      73              :     }
      74              : 
      75            0 :     return HCCL_SUCCESS;
      76            0 : }
      77            0 : bool CommHalvingDoubling::NeedDataReceivedAck()
      78              : {
      79            0 :     HCCL_INFO("having-doubling need make a comm supporting DataReceivedAck");
      80            0 :     return true;
      81              : }
      82              : }  // namespace hccl