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              : #ifndef HCCLV2_CCU_CONTEXT_ALL_GATHER_MESH_1D_DETOUR_H
      12              : #define HCCLV2_CCU_CONTEXT_ALL_GATHER_MESH_1D_DETOUR_H
      13              : 
      14              : #include <vector>
      15              : 
      16              : #include <ios>
      17              : #include "log.h"
      18              : #include "ccu_context_alg_base.h"
      19              : #include "ccu_datatype.h"
      20              : 
      21              : namespace Hccl {
      22              : 
      23              : class CcuContextAllGatherMeshDetour1D : public CcuContextAlgBase {
      24              : public:
      25              :     CcuContextAllGatherMeshDetour1D(const CcuCtxArg &arg, const std::vector<CcuTransport*> &transports,
      26              :                               const CcuTransportGroup &group);
      27            0 :     ~CcuContextAllGatherMeshDetour1D() override {}
      28              : 
      29              :     void Algorithm() override;
      30              :     std::vector<uint64_t> GeneArgs(const CcuTaskArg &arg) override;
      31              : private:
      32              :     void ProcessTransports(const std::vector<CcuTransport *> &transports);
      33              :     void AllocDetourRes();
      34              :     void CreateMultiOpBroadcastDetour();
      35              :     void GroupBroadcastDetour(std::vector<CcuRep::Variable> &lengths, std::vector<CcuRep::Memory> &src,
      36              :         std::vector<CcuRep::Memory> &dst);
      37              :     void FirstStep();
      38              :     void SecondStep();
      39              : 
      40              :     // ctx args
      41              :     uint64_t singleTransportSize_{0};  // 每个loop单次传输的总数据量，通信域级别
      42              :     uint64_t detourPathNum_{0};
      43              :     uint64_t pathNumPerPeer_{0};  // 到每个rank有几个transport，包括重复的
      44              :     std::vector<std::vector<CcuTransport*>> detourTransports_;  // 默认是transport的每rankSize-1个为一组，第一组是直连链路
      45              : 
      46              :     // task args
      47              :     CcuRep::Variable input_;
      48              :     std::vector<CcuRep::Variable> output_;
      49              :     std::vector<CcuRep::Variable> token_;
      50              :     CcuRep::Variable baseOffset_;  // rankId*sliceSize，待广播分片的基础偏移
      51              :     CcuRep::Variable tailOffset_;  // 尾块相对偏移，singleTransportSize*128*iterNum
      52              :     CcuRep::Variable loopIterNum_;  // for detour loopgroup only
      53              :     std::vector<CcuRep::Variable> lengths_;  // 每组transport对应一个len
      54              :     GroupOpSize groupOpSize_;  // 只处理尾块
      55              : };
      56              : } // namespace Hccl
      57              : 
      58              : #endif // HCCLV2_CCU_CONTEXT_ALL_GATHER_MESH_1D_DETOUR_H