* 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.
*/
#ifndef GE_HYBRID_EXECUTOR_MODEL_HYBRID_MODEL_ASYNC_EXECUTOR_H_
#define GE_HYBRID_EXECUTOR_MODEL_HYBRID_MODEL_ASYNC_EXECUTOR_H_
#include <atomic>
#include <mutex>
#include <future>
#include "ge/ge_api_error_codes.h"
#include "ge/ge_api_types.h"
#include "common/dump/opdebug_register.h"
#include "graph/load/model_manager/data_inputer.h"
#include "common/dump/data_dumper.h"
#include "hybrid/executor/hybrid_model_executor.h"
#include "hybrid/executor/hybrid_model_rt_v1_executor.h"
#include "hybrid/executor/hybrid_model_pipeline_executor.h"
#include "hybrid/executor/hybrid_model_rt_v2_executor.h"
#include "acl/acl_rt.h"
namespace ge {
namespace hybrid {
class HybridModelAsyncExecutor {
public:
struct DefaultStreamGuarder {
aclrtStream default_stream = nullptr;
uint32_t stream_ref_count = 0U;
std::mutex mu;
};
explicit HybridModelAsyncExecutor(HybridModel *const model);
~HybridModelAsyncExecutor();
Status Init(const aclrtStream stream = nullptr);
Status Execute(const std::vector<DataBuffer> &inputs,
const std::vector<GeTensorDesc> &input_desc,
std::vector<DataBuffer> &outputs,
std::vector<GeTensorDesc> &output_desc,
aclrtStream stream = nullptr);
Status Execute(const std::vector<gert::Tensor> &inputs, std::vector<gert::Tensor> &outputs);
Status ExecuteWithStreamAsync(const std::vector<GeTensor> &inputs, std::vector<GeTensor> &outputs,
aclrtStream stream = nullptr);
Status ExecuteWithStreamAsync(const std::vector<gert::Tensor> &inputs,
std::vector<gert::Tensor> &outputs,
aclrtStream stream = nullptr);
Status Start(const std::shared_ptr<ModelListener> &listener);
void SetDeviceId(const uint32_t device_id);
void SetModelId(const uint32_t model_id);
Status Stop();
Status EnqueueData(const std::shared_ptr<RunArgs> &args);
uint32_t GetDataInputerSize() const { return data_inputer_->Size(); }
bool GetRunningFlag() const { return running_flag_; }
void SetRunningFlag(const bool flag) { running_flag_ = flag; }
const GraphExecutionContext *GeContext() const { return executor_->GetContext(); }
GraphExecutionContext *GeContext() {return executor_->GetContext(); }
private:
Status RunInternal();
Status BuildExecutor();
DefaultStreamGuarder &GetDefaultStreamGuarder() const;
static std::map<std::pair<uint32_t, uint32_t>, DefaultStreamGuarder> default_stream_by_dev_;
static std::mutex mu_for_guarder_;
HybridModel *model_;
uint32_t device_id_ = 0U;
uint32_t model_id_ = 0U;
std::atomic_bool run_flag_;
bool running_flag_ = false;
std::unique_ptr<DataInputer> data_inputer_;
std::unique_ptr<HybridModelExecutor> executor_;
std::future<Status> future_;
aclrtStream stream_ = nullptr;
bool owner_stream_ = false;
std::shared_ptr<ModelListener> listener_;
std::vector<TensorValue> output_cache_;
};
}
}
#endif