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_COLL_ALG_BASE_SELECTOR
      12              : #define HCCLV2_COLL_ALG_BASE_SELECTOR
      13              : 
      14              : #include <string>
      15              : #include "coll_alg_params.h"
      16              : #include "coll_operator.h"
      17              : #include "virtual_topo.h"
      18              : #include "log.h"
      19              : #include "env_config.h"
      20              : 
      21              : namespace Hccl {
      22              : constexpr u32 CCU_MS_MODE = 2;
      23              : 
      24              : constexpr u32 BIG_CLOS_RANGE = 8;
      25              : 
      26              : enum class SelectorStatus { MATCH, NOT_MATCH };
      27              : enum class Level0Shape {
      28              :     MESH_1D = 1,
      29              :     MESH_2D = 2,
      30              :     CLOS = 3,
      31              :     MESH_1D_CLOS = 4,
      32              : };
      33              : 
      34              : 
      35              : const std::map<OpType, std::string> OP_TYPE_TO_AICPU_SOLE_ALG_MAP = {
      36              :     {OpType::ALLGATHER, "InsAllGatherMesh"},
      37              :     {OpType::REDUCESCATTER, "InsReduceScatterNHR"},
      38              :     {OpType::ALLREDUCE, "InsAllReduceNHR"},
      39              :     {OpType::ALLTOALL, "InsAlltoAllMesh"},
      40              :     {OpType::ALLTOALLV, "InsAlltoAllvMesh"},
      41              :     {OpType::ALLTOALLVC, "InsAlltoAllvcMesh"},
      42              : };
      43              : 
      44              : class BaseSelector {
      45              : public:
      46            0 :     virtual ~BaseSelector() {};
      47              :     BaseSelector &SetVirtualTopo(RankGraph *rankGraph);
      48              :     BaseSelector &SetDevType(DevType devType);
      49              :     BaseSelector &SetMyRank(RankId myRank);
      50              :     BaseSelector &SetRankSize(u32 rankSize);
      51              :     BaseSelector &SetSeverId(std::string severId);
      52              :     BaseSelector &SetDeviceNumPerSever(u32 deviceNumPerSever);
      53              :     BaseSelector &SetServerNum(u32 serverNum);
      54              :     BaseSelector &SetOpConfig(OpExecuteConfig opConfig);
      55              :     BaseSelector &SetIsMc2(bool isMc2);
      56              : 
      57              :     RankGraph *GetVirtualTopo();
      58              :     DevType      GetDevType();
      59              :     RankId       GetMyRank() const;
      60              :     u32          GetRankSize() const;
      61              :     std::string  GetSeverId();
      62              :     u32          GetDeviceNumPerSever() const;
      63              :     u32          GetServerNum() const;
      64              : 
      65              :     bool IsInputOutputOverlap(const std::shared_ptr<Buffer> &inputMem, const std::shared_ptr<Buffer> &outputMem) const;
      66              :     virtual SelectorStatus Select(const CollAlgOperator &op, CollAlgParams &params, std::string &primQueueGenName)
      67              :      = 0;
      68              : 
      69              : protected:
      70              :     struct NetLayerDetails {
      71              :         u32 netLayerNum;
      72              :         std::set<u32> netLayers;
      73              :         std::vector<u32> netInstNumOfLayer;
      74              :         std::vector<std::vector<u32>> instSizeListOfLayer;
      75              :         std::vector<u32> localNetInsSizeOfLayer;
      76              :     };
      77              :     struct TopoInstDetails {
      78              :         u32 topoInstNum;
      79              :         std::vector<u32> sizeOfTopo;
      80              :         std::vector<TopoType> typeOfTopo;
      81              :         std::vector<std::vector<u32>> ranksInTopo;
      82              :         std::map<TopoType, std::vector<u32>> rankNumForTopoType;
      83              :     };
      84              :     struct TopoInfo {
      85              :         u32 levelNum;
      86              :         Level0Shape level0Shape;
      87              :         NetLayerDetails netLayerDetails;
      88              :         std::vector<TopoInstDetails> topoInstDetailsOfLayer;
      89              : 
      90              :         bool Level0Nhr{false};
      91              :         bool Level1Nhr{false};
      92              : 
      93              :         bool level0PcieMix{false};
      94              :         bool level0BigClosRange{false};
      95              :     };
      96              :     u32 Gcd(u32 a, u32 b) const;  // 自定义实现的 gcd 函数（兼容旧版本 C++）
      97              :     u32 GcdOfArray(const std::vector<u32> &numbers) const;  // 计算数组中所有元素的最大公约数
      98              :     u32 GetLevel0Gcd();
      99              :     void CalcTopoShape(TopoInfo &topoInfo) const;
     100              :     bool IsAsymmetricTopoShapeLevel1Nhr(const std::vector<std::vector<u32>> &localIdPerBoard, u32 gcdRankSizeLevel0) const;
     101              :     bool IsTopoShapeLevel0Regular(const std::vector<std::vector<u32>> &localIdPerBoard) const;
     102              :     bool IsLayerAllConnetedWithTopo(const TopoInfo &topoInfo, const u32 netLayer, const TopoType topoType) const;
     103              :     HcclResult CalcLevel0TopoShape(TopoInfo &topoInfo) const;
     104              :     HcclResult ExtractNetLayerDetails(TopoInfo &topoInfo) const;
     105              :     HcclResult ExtractTopoDetails(TopoInfo &topoInfo) const;
     106              :     bool Is2DieFullMesh() const;
     107              :     bool IsLevel0PcieMix() const;
     108              :     RankGraph *rankGraph_ = nullptr;
     109              :     OpExecuteConfig opConfig_;
     110              :     DevType      devType_;
     111              :     RankId       myRank_;
     112              :     u32          rankSize_;
     113              :     std::string  severId_;
     114              :     u32          deviceNumPerSever_;
     115              :     u32          serverNum_;
     116              :     bool         isMc2_{false};
     117              : };
     118              : 
     119              : } // namespace Hccl
     120              : #endif