* -------------------------------------------------------------------------
* This file is part of the MindStudio project.
* Copyright (c) 2025 Huawei Technologies Co.,Ltd.
*
* MindStudio is licensed under Mulan PSL v2.
* You can use this software according to the terms and conditions of the Mulan PSL v2.
* You may obtain a copy of Mulan PSL v2 at:
*
* http://license.coscl.org.cn/MulanPSL2
*
* 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 FIT FOR A PARTICULAR PURPOSE.
* See the Mulan PSL v2 for more details.
* -------------------------------------------------------------------------
*/
#ifndef PROFILER_SERVER_RLMICROBATCHFSDPCLASSIFIER_H
#define PROFILER_SERVER_RLMICROBATCHFSDPCLASSIFIER_H
#include "RLMicroBatchClassifierBase.h"
#include "PythonApiRepo.h"
namespace Dic::Module::RL {
using namespace Dic::Module::Timeline;
class RLMicroBatchFSDPClassifier : public RLMicroBatchClassifierBase {
public:
virtual ~RLMicroBatchFSDPClassifier() = default;
protected:
std::vector<Protocol::RLPipelineNode> QueryMicroBatchSlices(
const std::string &fileId, const RLMstxConfig &config, const Protocol::RLPipelineNode &taskNode) override;
std::vector<Protocol::RLPipelineNode> MicroBatchClassifier(std::vector<Protocol::RLPipelineNode> &nodes) override;
std::vector<RLPipelineNode> QueryFPSlices(const std::string &rankId, const RLPipelineNode &taskNode);
std::vector<RLPipelineNode> QueryBPSlices(const std::string &rankId, const RLPipelineNode &taskNode);
std::vector<RLPipelineNode> TransSliceToNodes(const std::vector<CompeteSliceDomain> &slices,
const RLPipelineNode &task, const std::string &name, const std::string &nodeType);
std::vector<RLPipelineNode> NodeSortMerge(
const std::vector<RLPipelineNode> &left, const std::vector<RLPipelineNode> &right);
bool IsHappenBefore(
std::vector<RLPipelineNode>::const_iterator left, std::vector<RLPipelineNode>::const_iterator right);
void ProcessInitState(const RLPipelineNode &node, std::vector<RLPipelineNode> &res);
void ProcessFPState(const RLPipelineNode &node, std::vector<RLPipelineNode> &res);
void ProcessBPState(const RLPipelineNode &node, std::vector<RLPipelineNode> &res);
void ProcessCommunicationState(const RLPipelineNode &node, std::vector<RLPipelineNode> &res);
RLPipelineNode current;
State state{State::Init};
inline static std::string FP_MICRO_BATCH_NAME = "FullyShardedDataParallel.forward";
inline static std::string BP_MICRO_BATCH_NAME = "autograd::engine";
inline static std::string BP_MICRO_BATCH_END_NAME = "Enqueue@record_event";
};
}
#endif