============================= test session starts ============================== platform linux -- Python 3.7.5, pytest-5.4.3, py-1.8.1, pluggy-0.13.1 rootdir: /home/jenkins/mindspore/testcases/testcases/tests/st/sparse, inifile: /home/jenkins/sault/virtual_test/virtualenv_001/sault/config/pytest.ini plugins: anyio-3.7.1, forked-1.1.3, xdist-1.32.0 [INFO] ATRACE(5051,python3.7):2024-01-11-05:49:52.885.081 [trace_attr.c:105](tid:5051) platform is 1. [INFO] ATRACE(5051,python3.7):2024-01-11-05:49:52.885.280 [trace_recorder.c:114](tid:5051) use root path: /home/jenkins/ascend/atrace [INFO] ATRACE(5051,python3.7):2024-01-11-05:49:52.885.311 [trace_signal.c:133](tid:5051) register signal handler for signo 2 succeed. [INFO] ATRACE(5051,python3.7):2024-01-11-05:49:52.885.322 [trace_signal.c:133](tid:5051) register signal handler for signo 15 succeed. [INFO] RUNTIME(5051,python3.7):2024-01-11-05:49:53.297.244 [runtime.cc:1159] 5051 GetAicoreNumByLevel: workingDev_=0 [INFO] RUNTIME(5051,python3.7):2024-01-11-05:49:53.297.317 [runtime.cc:4719] 5051 GetVisibleDevices: ASCEND_RT_VISIBLE_DEVICES param was not set collected 1 item test_coo.py [INFO] TDT(5051,python3.7):2024-01-11-05:49:57.756.143 [process_mode_manager.cpp:109][OpenProcess][tid:5051] [ProcessModeManager] enter into open process deviceId[0] rankSize[0] [INFO] TDT(5051,python3.7):2024-01-11-05:49:57.757.604 [process_mode_manager.cpp:379][InitTsdClient][tid:5051] [TsdClient] deviceId[0] begin to init hdc client [INFO] TDT(5051,python3.7):2024-01-11-05:49:57.757.753 [version_verify.cpp:34][SetVersionInfo][tid:5051] VersionVerify: send client version to server [INFO] TDT(5051,python3.7):2024-01-11-05:49:57.757.798 [version_verify.cpp:50][SetVersionInfo][tid:5051] send feature_info:{msg_type:35, features:{check before send aicpu package,}} [INFO] TDT(5051,python3.7):2024-01-11-05:49:57.757.817 [version_verify.cpp:50][SetVersionInfo][tid:5051] send feature_info:{msg_type:37, features:{check before send open qs message,}} [INFO] TDT(5051,python3.7):2024-01-11-05:49:57.758.143 [version_verify.cpp:66][PeerVersionCheck][tid:5051] VersionVerify: Check client version info, server[1230], client[1230] [INFO] TDT(5051,python3.7):2024-01-11-05:49:57.758.159 [version_verify.cpp:87][ParseVersionInfo][tid:5051] VersionVerify: pass client version info success [INFO] TDT(5051,python3.7):2024-01-11-05:49:57.758.168 [hdc_client.cpp:276][CheckHdcConnection][tid:5051] Service[2] create hdc success [INFO] TDT(5051,python3.7):2024-01-11-05:49:57.758.183 [version_verify.cpp:120][SpecialFeatureCheck][tid:5051] VersionVerify: new type[35], supported [INFO] TDT(5051,python3.7):2024-01-11-05:49:57.758.227 [process_mode_manager.cpp:748][GetDeviceCheckCode][tid:5051] [TsdClient][deviceId=0] [sessionId=1] wait package info respond [INFO] TDT(5051,python3.7):2024-01-11-05:49:57.758.357 [process_mode_manager.cpp:379][InitTsdClient][tid:5051] [TsdClient] deviceId[0] begin to init hdc client [INFO] TDT(5051,python3.7):2024-01-11-05:49:57.758.497 [version_verify.cpp:34][SetVersionInfo][tid:5051] VersionVerify: send client version to server [INFO] TDT(5051,python3.7):2024-01-11-05:49:57.758.509 [version_verify.cpp:50][SetVersionInfo][tid:5051] send feature_info:{msg_type:35, features:{check before send aicpu package,}} [INFO] TDT(5051,python3.7):2024-01-11-05:49:57.758.519 [version_verify.cpp:50][SetVersionInfo][tid:5051] send feature_info:{msg_type:37, features:{check before send open qs message,}} [INFO] TDT(5051,python3.7):2024-01-11-05:49:57.758.657 [version_verify.cpp:66][PeerVersionCheck][tid:5051] VersionVerify: Check client version info, server[1230], client[1230] [INFO] TDT(5051,python3.7):2024-01-11-05:49:57.758.668 [version_verify.cpp:87][ParseVersionInfo][tid:5051] VersionVerify: pass client version info success [INFO] TDT(5051,python3.7):2024-01-11-05:49:57.758.677 [hdc_client.cpp:276][CheckHdcConnection][tid:5051] Service[2] create hdc success [INFO] TDT(5051,python3.7):2024-01-11-05:49:57.758.687 [process_mode_manager.cpp:426][ConstructOpenMsg][tid:5051] [TsdClient] tsd get process sign successfully, procpid[5051] signSize[48] [INFO] TDT(5051,python3.7):2024-01-11-05:49:57.758.714 [version_verify.cpp:112][SpecialFeatureCheck][tid:5051] VersionVerify: previous type[6], supported [INFO] TDT(5051,python3.7):2024-01-11-05:49:57.758.732 [process_mode_manager.cpp:126][OpenProcess][tid:5051] [ProcessModeManager] deviceId[0] sessionId[1] rankSize[0], wait sub process start respond [INFO] TDT(5051,python3.7):2024-01-11-05:49:58.082.879 [stub_process_mode_nowin.cpp:63][ProcessQueueForMdc][tid:5051] [TsdClient] it is unnecessary of current mode[0] chiptype[1] to grant queue auth to aicpusd [INFO] TDT(5051,python3.7):2024-01-11-05:49:58.082.935 [stub_process_mode_nowin.cpp:101][OpenInHost][tid:5051] enter into OpenInHost deviceid[0] [INFO] TDT(5051,python3.7):2024-01-11-05:49:58.082.946 [stub_process_mode_nowin.cpp:105][OpenInHost][tid:5051] host cpu not support [INFO] TDT(5051,python3.7):2024-01-11-05:49:58.082.954 [process_mode_manager.cpp:156][OpenProcess][tid:5051] [TsdClient][deviceId=0] [sessionId=1] start hccp and computer process success [INFO] RUNTIME(5051,python3.7):2024-01-11-05:49:58.085.643 [device.cc:340] 5051 Init: isDoubledie:0, topologytype:0 [INFO] RUNTIME(5051,python3.7):2024-01-11-05:49:58.098.226 [npu_driver.cc:5428] 6046 GetDeviceStatus: GetDeviceStatus status=1. [INFO] ATRACE(5051,python3.7):2024-01-11-05:49:58.098.272 [atrace_api.c:28](tid:5051) AtraceCreate start [INFO] ATRACE(5051,python3.7):2024-01-11-05:49:58.098.386 [trace_rb_log.c:84](tid:5051) [RUNTIME_ATRACE_DEV0_TS0] create ring buffer success, buffer size : 131152. [INFO] ATRACE(5051,python3.7):2024-01-11-05:49:58.098.403 [atrace_api.c:32](tid:5051) AtraceCreate end [INFO] TDT(5051,python3.7):2024-01-11-05:49:58.098.427 [client_manager.cpp:157][SetProfilingCallback][tid:5051] [TsdClient] set profiling callback success [TRACE] GE(5051,python3.7):2024-01-11-05:49:58.176.808 [status:INIT] [ge_api.cc:144]5051 GEInitializeImpl:GEInitialize start [INFO] PROFILING(5051,python3.7):2024-01-11-05:49:58.393.347 [msprofiler_impl.cpp:156] >>> (tid:5051) ProfNotifySetDevice called, is open: 1, devId: 0 [INFO] PROFILING(5051,python3.7):2024-01-11-05:49:58.393.491 [platform.cpp:38] >>> (tid:5051) Profiling platform version: 1.0. [INFO] PROFILING(5051,python3.7):2024-01-11-05:49:58.393.507 [ai_drv_dev_api.cpp:384] >>> (tid:5051) Succeeded to DrvGetApiVersion version: 0x72313 [TRACE] GE(5051,python3.7):2024-01-11-05:49:58.441.186 [status:RUNNING] [ge_api.cc:211]5051 GEInitializeImpl:Initializing environment [INFO] GE(5051,python3.7):2024-01-11-05:49:58.441.274 [gelib.cc:98][EVENT]5051 Initialize:[GEPERFTRACE] GE Init Start [INFO] GE(5051,python3.7):2024-01-11-05:49:58.441.613 [gelib.cc:307][EVENT]5051 SystemInitialize:Online infer init GELib success, device id :0 [INFO] DVPP(5051,python3.7):2024-01-11-05:49:58.786.680 [dvpp_engine.cc:41][ENGINE][Initialize:41][tid:5051]dvpp engine do not support [INFO] TUNE(5051,python3.7):2024-01-11-05:49:58.790.920 [cann_kb_pyfunc_mgr.cpp:72][CANNKB][Tid:5051]"CannKbPyfuncMgr: Enter PyObjectInit, reference_ is 0!" [INFO] TUNE(5051,python3.7):2024-01-11-05:49:58.790.966 [handle_manager.cpp:115][CANNKB][Tid:5051]"Start to run init functions to load dynamic python lib!" [INFO] TUNE(5051,python3.7):2024-01-11-05:49:58.791.026 [handle_manager.cpp:407][CANNKB][Tid:5051]"Init functions of loading dynamic python lib end!" [INFO] TUNE(5051,python3.7):2024-01-11-05:49:58.791.036 [cann_kb_pyfunc_mgr.cpp:24][CANNKB][Tid:5051]"CANN_KB_Py has already been initialized." [INFO] TUNE(5051,python3.7):2024-01-11-05:49:58.791.098 [cann_kb_pyfunc_mgr.cpp:117][CANNKB][Tid:5051]"CannKbPyfuncMgr: Run PyObjectInit successfully!" [INFO] HCCL(5051,python3.7):2024-01-11-05:50:11.122.072 [plugin_manager.cc:42][5051]hcom running normal mode. [INFO] DVPP(5051,python3.7):2024-01-11-05:50:11.122.777 [dvpp_engine.cc:92][ENGINE][GetOpsKernelInfoStores:92][tid:5051]dvpp ops kernel info store do not support [INFO] DVPP(5051,python3.7):2024-01-11-05:50:11.122.921 [dvpp_engine.cc:69][ENGINE][GetGraphOptimizerObjs:69][tid:5051]dvpp graph optimizer do not support [INFO] DVPP(5051,python3.7):2024-01-11-05:50:11.657.688 [dvpp_ops_kernel_builder.cc:48][ENGINE][Initialize:48][tid:5051]dvpp ops kernel builder do not support [INFO] GE(5051,python3.7):2024-01-11-05:50:11.665.959 [gelib.cc:169][EVENT]5051 Initialize:[GEPERFTRACE] The time cost of GELib::Initialize is [13224574] micro second. [TRACE] GE(5051,python3.7):2024-01-11-05:50:11.759.515 [status:STOP] [ge_api.cc:255]5051 GEInitializeImpl:GEInitialize finished [TRACE] GE(5051,python3.7):2024-01-11-05:50:11.759.652 [status:INIT] [ge_api.cc:398]5051 Session:Start to construct session. [TRACE] GE(5051,python3.7):2024-01-11-05:50:11.759.671 [status:RUNNING] [ge_api.cc:408]5051 Session:Creating session [INFO] GE(5051,python3.7):2024-01-11-05:50:11.760.200 [graph_var_manager.cc:1445][EVENT]5051 SetMemoryMallocSize:Total memory size is 34357641216 [INFO] GE(5051,python3.7):2024-01-11-05:50:11.760.218 [graph_var_manager.cc:1424][EVENT]5051 SetAllMemoryMaxValue:The graph_mem_max_size is 27915583488 and the var_mem_max_size is 5368381440 [INFO] PROFILING(5051,python3.7):2024-01-11-05:50:11.760.634 [msprofiler_impl.cpp:156] >>> (tid:5051) ProfNotifySetDevice called, is open: 1, devId: 0 [TRACE] GE(5051,python3.7):2024-01-11-05:50:11.761.593 [status:RUNNING] [ge_api.cc:411]5051 Session:Session id is 0 [TRACE] GE(5051,python3.7):2024-01-11-05:50:11.761.616 [status:STOP] [ge_api.cc:420]5051 Session:Session Constructor finished [INFO] PROFILING(5051,python3.7):2024-01-11-05:50:11.771.365 [platform.cpp:38] >>> (tid:5051) Profiling platform version: 1.0. [INFO] PROFILING(5051,python3.7):2024-01-11-05:50:11.771.394 [ai_drv_dev_api.cpp:384] >>> (tid:5051) Succeeded to DrvGetApiVersion version: 0x72313 [TRACE] GE(5051,python3.7):2024-01-11-05:50:11.771.642 [status:INIT] [ge_api.cc:144]5051 GEInitializeImpl:GEInitialize start [INFO] GE(5051,python3.7):2024-01-11-05:50:11.896.318 [scalable_config.cc:55][EVENT]5051 ScalableConfig:device total max size: 34357641216, page_mem_size_total_thresold: 32639759155, uncacheable_size_threshold: 17179869184 [INFO] GE(5051,python3.7):2024-01-11-05:50:11.978.657 [graph_var_manager.cc:1424][EVENT]5051 SetAllMemoryMaxValue:The graph_mem_max_size is 27917287424 and the var_mem_max_size is 5368709120 [INFO] GE(5051,python3.7):2024-01-11-05:50:11.978.775 [graph_manager.cc:1248][EVENT]5051 PreRun:PreRun start: graph node size 3, session id 1, graph id 0, graph name online. [INFO] ATRACE(5051,python3.7):2024-01-11-05:50:11.979.759 [atrace_api.c:28](tid:5051) AtraceCreate start [INFO] ATRACE(5051,python3.7):2024-01-11-05:50:11.979.841 [trace_rb_log.c:84](tid:5051) [RUNTIME_ATRACE_DEV64_TS0] create ring buffer success, buffer size : 131152. [INFO] ATRACE(5051,python3.7):2024-01-11-05:50:11.979.855 [atrace_api.c:32](tid:5051) AtraceCreate end [INFO] TDT(5051,python3.7):2024-01-11-05:50:11.979.885 [client_manager.cpp:157][SetProfilingCallback][tid:5051] [TsdClient] set profiling callback success [INFO] GE(5051,python3.7):2024-01-11-05:50:11.980.976 [parallel_partitioner.cc:165][EVENT]5051 DoPipelinePartition:[GEPERFTRACE] The time cost of OptimizeSubgraph::PipelinePartition is [34] micro second. [INFO] GE(5051,python3.7):2024-01-11-05:50:11.981.026 [parallel_partitioner.cc:178][EVENT]5051 DoFlowGraphPartition:[GEPERFTRACE] The time cost of OptimizeSubgraph::FlowGraphPartition is [20] micro second. [INFO] GE(5051,python3.7):2024-01-11-05:50:11.981.090 [graph_prepare.cc:1378][EVENT]5051 Init:[GEPERFTRACE] The time cost of FileConstantUtils::ConvertFileConstToConst is [4] micro second. [INFO] GE(5051,python3.7):2024-01-11-05:50:11.981.747 [graph_manager.cc:1050][EVENT]5051 PreRunOptimizeOriginalGraph:[GEPERFTRACE] The time cost of GraphManager::stages.preparer.PrepareInit is [678] micro second. [INFO] GE(5051,python3.7):2024-01-11-05:50:11.981.779 [graph_manager.cc:1052][EVENT]5051 PreRunOptimizeOriginalGraph:[GEPERFTRACE] The time cost of GraphManager::stages.optimizer.HandleSummaryOp is [9] micro second. [INFO] GE(5051,python3.7):2024-01-11-05:50:11.981.930 [base_pass.cc:339][EVENT]5051 Run:[GEPERFTRACE] The time cost of ForToWhilePass is [2] micro second, call num is [3] [INFO] GE(5051,python3.7):2024-01-11-05:50:11.981.983 [pass_manager.cc:82][EVENT]5051 Run:[GEPERFTRACE] The time cost of ProcessNetOutput::SavePass is [4] micro second. [INFO] GE(5051,python3.7):2024-01-11-05:50:11.982.127 [pass_manager.cc:82][EVENT]5051 Run:[GEPERFTRACE] The time cost of ProcessNetOutput::NetOutputPass is [132] micro second. [INFO] GE(5051,python3.7):2024-01-11-05:50:11.982.142 [pass_manager.cc:82][EVENT]5051 Run:[GEPERFTRACE] The time cost of ProcessNetOutput::DataPass is [1] micro second. [INFO] GE(5051,python3.7):2024-01-11-05:50:11.982.254 [pass_manager.cc:82][EVENT]5051 Run:[GEPERFTRACE] The time cost of CreateSubGraphWithScopePass is [39] micro second. [INFO] GE(5051,python3.7):2024-01-11-05:50:11.982.273 [pass_manager.cc:82][EVENT]5051 Run:[GEPERFTRACE] The time cost of SubgraphMultiDimsClonePass is [6] micro second. [INFO] GE(5051,python3.7):2024-01-11-05:50:11.982.292 [pass_manager.cc:82][EVENT]5051 Run:[GEPERFTRACE] The time cost of MultiBatchClonePass is [8] micro second. [INFO] GE(5051,python3.7):2024-01-11-05:50:11.982.384 [base_pass.cc:339][EVENT]5051 Run:[GEPERFTRACE] The time cost of SplitVariableIntoSubgraphPass is [2] micro second, call num is [3] [INFO] GE(5051,python3.7):2024-01-11-05:50:11.982.407 [graph_manager.cc:1054][EVENT]5051 PreRunOptimizeOriginalGraph:[GEPERFTRACE] The time cost of GraphManager::stages.preparer.NormalizeGraph is [615] micro second. [INFO] GE(5051,python3.7):2024-01-11-05:50:11.989.723 [graph_manager.cc:1055][EVENT]5051 PreRunOptimizeOriginalGraph:[GEPERFTRACE] The time cost of GraphManager::stages.optimizer.OptimizeGraphInit is [7299] micro second. [INFO] GE(5051,python3.7):2024-01-11-05:50:11.991.150 [base_pass.cc:339][EVENT]5051 Run:[GEPERFTRACE] The time cost of SwitchDeadBranchElimination is [3] micro second, call num is [3] [INFO] GE(5051,python3.7):2024-01-11-05:50:11.991.178 [base_pass.cc:339][EVENT]5051 Run:[GEPERFTRACE] The time cost of MergePass is [14] micro second, call num is [3] [INFO] GE(5051,python3.7):2024-01-11-05:50:11.991.190 [base_pass.cc:339][EVENT]5051 Run:[GEPERFTRACE] The time cost of InferShapePass is [369] micro second, call num is [3] [INFO] GE(5051,python3.7):2024-01-11-05:50:11.991.200 [base_pass.cc:339][EVENT]5051 Run:[GEPERFTRACE] The time cost of ReplaceWithEmptyConstPass is [44] micro second, call num is [3] [INFO] GE(5051,python3.7):2024-01-11-05:50:11.991.210 [base_pass.cc:339][EVENT]5051 Run:[GEPERFTRACE] The time cost of SplitShapeNPass is [1] micro second, call num is [3] [INFO] GE(5051,python3.7):2024-01-11-05:50:11.991.218 [base_pass.cc:339][EVENT]5051 Run:[GEPERFTRACE] The time cost of DimensionComputePass is [134] micro second, call num is [3] [INFO] GE(5051,python3.7):2024-01-11-05:50:11.991.227 [base_pass.cc:339][EVENT]5051 Run:[GEPERFTRACE] The time cost of ConstantFoldingPass is [29] micro second, call num is [3] [INFO] GE(5051,python3.7):2024-01-11-05:50:11.991.235 [base_pass.cc:339][EVENT]5051 Run:[GEPERFTRACE] The time cost of InferValuePass is [16] micro second, call num is [3] [INFO] GE(5051,python3.7):2024-01-11-05:50:11.993.009 [graph_manager.cc:1056][EVENT]5051 PreRunOptimizeOriginalGraph:[GEPERFTRACE] The time cost of GraphManager::stages.optimizer.OptimizeOriginalGraphForQuantize is [3246] micro second. [INFO] GE(5051,python3.7):2024-01-11-05:50:11.993.095 [base_pass.cc:339][EVENT]5051 Run:[GEPERFTRACE] The time cost of CondRemovePass is [21] micro second, call num is [3] [INFO] GE(5051,python3.7):2024-01-11-05:50:11.993.165 [graph_prepare.cc:1982][EVENT]5051 PrepareDynShape:[GEPERFTRACE] The time cost of Prepare::ProcessBeforeInfershape is [73] micro second. [INFO] GE(5051,python3.7):2024-01-11-05:50:11.993.534 [base_pass.cc:339][EVENT]5051 Run:[GEPERFTRACE] The time cost of SwitchDeadBranchElimination is [2] micro second, call num is [3] [INFO] GE(5051,python3.7):2024-01-11-05:50:11.993.557 [base_pass.cc:339][EVENT]5051 Run:[GEPERFTRACE] The time cost of MergePass is [1] micro second, call num is [3] [INFO] GE(5051,python3.7):2024-01-11-05:50:11.993.568 [base_pass.cc:339][EVENT]5051 Run:[GEPERFTRACE] The time cost of InferShapePass is [175] micro second, call num is [3] [INFO] GE(5051,python3.7):2024-01-11-05:50:11.993.589 [base_pass.cc:339][EVENT]5051 Run:[GEPERFTRACE] The time cost of ReplaceWithEmptyConstPass is [7] micro second, call num is [3] [INFO] GE(5051,python3.7):2024-01-11-05:50:11.993.598 [base_pass.cc:339][EVENT]5051 Run:[GEPERFTRACE] The time cost of SplitShapeNPass is [0] micro second, call num is [3] [INFO] GE(5051,python3.7):2024-01-11-05:50:11.993.607 [base_pass.cc:339][EVENT]5051 Run:[GEPERFTRACE] The time cost of DimensionComputePass is [8] micro second, call num is [3] [INFO] GE(5051,python3.7):2024-01-11-05:50:11.993.616 [base_pass.cc:339][EVENT]5051 Run:[GEPERFTRACE] The time cost of ConstantFoldingPass is [9] micro second, call num is [3] [INFO] GE(5051,python3.7):2024-01-11-05:50:11.993.624 [base_pass.cc:339][EVENT]5051 Run:[GEPERFTRACE] The time cost of InferValuePass is [5] micro second, call num is [3] [INFO] GE(5051,python3.7):2024-01-11-05:50:11.993.650 [graph_prepare.cc:1983][EVENT]5051 PrepareDynShape:[GEPERFTRACE] The time cost of Prepare::FormatAndShapeProcess is [453] micro second. [INFO] GE(5051,python3.7):2024-01-11-05:50:11.993.674 [pass_manager.cc:82][EVENT]5051 Run:[GEPERFTRACE] The time cost of PreRun::MarkForceUnknownForCondPass is [5] micro second. [INFO] GE(5051,python3.7):2024-01-11-05:50:11.993.686 [graph_prepare.cc:1984][EVENT]5051 PrepareDynShape:[GEPERFTRACE] The time cost of Prepare::CtrlFlowPreProcess is [21] micro second. [INFO] GE(5051,python3.7):2024-01-11-05:50:11.993.700 [graph_prepare.cc:1985][EVENT]5051 PrepareDynShape:[GEPERFTRACE] The time cost of Prepare::multibatch::GetDynamicOutputShape is [3] micro second. [INFO] GE(5051,python3.7):2024-01-11-05:50:11.993.730 [graph_prepare.cc:1986][EVENT]5051 PrepareDynShape:[GEPERFTRACE] The time cost of Prepare::InsertAippOpUtil::Instance().UpdateDataNodeByAipp is [20] micro second. [INFO] GE(5051,python3.7):2024-01-11-05:50:11.993.742 [graph_prepare.cc:1987][EVENT]5051 PrepareDynShape:[GEPERFTRACE] The time cost of Prepare::SaveOriginalGraphToOmModel is [0] micro second. [INFO] GE(5051,python3.7):2024-01-11-05:50:11.993.777 [pass_manager.cc:82][EVENT]5051 Run:[GEPERFTRACE] The time cost of PrepareOptimize::ShapeOperateOpRemovePass is [21] micro second. [INFO] GE(5051,python3.7):2024-01-11-05:50:11.993.790 [pass_manager.cc:82][EVENT]5051 Run:[GEPERFTRACE] The time cost of PrepareOptimize::ReplaceTransShapePass is [1] micro second. [INFO] GE(5051,python3.7):2024-01-11-05:50:11.993.803 [pass_manager.cc:82][EVENT]5051 Run:[GEPERFTRACE] The time cost of PrepareOptimize::MarkAgnosticPass is [5] micro second. [INFO] GE(5051,python3.7):2024-01-11-05:50:11.993.881 [base_pass.cc:339][EVENT]5051 Run:[GEPERFTRACE] The time cost of EnterPass is [2] micro second, call num is [3] [INFO] GE(5051,python3.7):2024-01-11-05:50:11.993.894 [base_pass.cc:339][EVENT]5051 Run:[GEPERFTRACE] The time cost of CondPass is [2] micro second, call num is [3] [INFO] GE(5051,python3.7):2024-01-11-05:50:11.993.904 [base_pass.cc:339][EVENT]5051 Run:[GEPERFTRACE] The time cost of PrintOpPass is [3] micro second, call num is [3] [INFO] GE(5051,python3.7):2024-01-11-05:50:11.993.912 [base_pass.cc:339][EVENT]5051 Run:[GEPERFTRACE] The time cost of NoUseReshapeRemovePass is [1] micro second, call num is [3] [INFO] GE(5051,python3.7):2024-01-11-05:50:11.993.921 [base_pass.cc:339][EVENT]5051 Run:[GEPERFTRACE] The time cost of TransposeRemovePass is [2] micro second, call num is [3] [INFO] GE(5051,python3.7):2024-01-11-05:50:11.993.929 [base_pass.cc:339][EVENT]5051 Run:[GEPERFTRACE] The time cost of UnusedConstPass is [1] micro second, call num is [3] [INFO] GE(5051,python3.7):2024-01-11-05:50:11.993.937 [base_pass.cc:339][EVENT]5051 Run:[GEPERFTRACE] The time cost of StopGradientPass is [1] micro second, call num is [3] [INFO] GE(5051,python3.7):2024-01-11-05:50:11.993.945 [base_pass.cc:339][EVENT]5051 Run:[GEPERFTRACE] The time cost of PreventGradientPass is [1] micro second, call num is [3] [INFO] GE(5051,python3.7):2024-01-11-05:50:11.993.953 [base_pass.cc:339][EVENT]5051 Run:[GEPERFTRACE] The time cost of PlaceholderWithDefaultPass is [1] micro second, call num is [3] [INFO] GE(5051,python3.7):2024-01-11-05:50:11.993.970 [base_pass.cc:339][EVENT]5051 Run:[GEPERFTRACE] The time cost of SnapshotPass is [1] micro second, call num is [3] [INFO] GE(5051,python3.7):2024-01-11-05:50:11.993.979 [base_pass.cc:339][EVENT]5051 Run:[GEPERFTRACE] The time cost of GuaranteeConstPass is [3] micro second, call num is [3] [INFO] GE(5051,python3.7):2024-01-11-05:50:11.993.987 [base_pass.cc:339][EVENT]5051 Run:[GEPERFTRACE] The time cost of VarIsInitializedOpPass is [4] micro second, call num is [3] [INFO] GE(5051,python3.7):2024-01-11-05:50:11.993.995 [base_pass.cc:339][EVENT]5051 Run:[GEPERFTRACE] The time cost of ParallelConcatStartOpPass is [1] micro second, call num is [3] [INFO] GE(5051,python3.7):2024-01-11-05:50:11.994.004 [base_pass.cc:339][EVENT]5051 Run:[GEPERFTRACE] The time cost of IdentityPass is [2] micro second, call num is [3] [INFO] GE(5051,python3.7):2024-01-11-05:50:11.994.027 [pass_manager.cc:82][EVENT]5051 Run:[GEPERFTRACE] The time cost of PrepareOptimize::PrunePass is [10] micro second. [INFO] GE(5051,python3.7):2024-01-11-05:50:11.994.042 [pass_manager.cc:82][EVENT]5051 Run:[GEPERFTRACE] The time cost of PrepareOptimize::HcclMemcpyPass is [2] micro second. [INFO] GE(5051,python3.7):2024-01-11-05:50:11.994.076 [graph_prepare.cc:1988][EVENT]5051 PrepareDynShape:[GEPERFTRACE] The time cost of Prepare::PrepareOptimize is [324] micro second. [INFO] GE(5051,python3.7):2024-01-11-05:50:11.994.089 [graph_manager.cc:1065][EVENT]5051 PreRunOptimizeOriginalGraph:[GEPERFTRACE] The time cost of GraphManager::stages.preparer.PrepareDynShape is [1046] micro second. [INFO] GE(5051,python3.7):2024-01-11-05:50:12.014.540 [graph_manager.cc:1077][EVENT]5051 PreRunOptimizeOriginalGraph:[GEPERFTRACE] The time cost of GraphManager::stages.optimizer.OptimizeOriginalGraph is [20430] micro second. [INFO] GE(5051,python3.7):2024-01-11-05:50:12.014.623 [pass_manager.cc:82][EVENT]5051 Run:[GEPERFTRACE] The time cost of PrepareRunningFormatRefiner::VariablePrepareOpPass is [7] micro second. [INFO] GE(5051,python3.7):2024-01-11-05:50:12.014.648 [graph_manager.cc:1080][EVENT]5051 PreRunOptimizeOriginalGraph:[GEPERFTRACE] The time cost of GraphManager::stages.preparer.PrepareRunningFormatRefiner is [64] micro second. [INFO] GE(5051,python3.7):2024-01-11-05:50:12.021.627 [graph_manager.cc:1081][EVENT]5051 PreRunOptimizeOriginalGraph:[GEPERFTRACE] The time cost of GraphManager::stages.optimizer.OptimizeOriginalGraphJudgeInsert is [6963] micro second. [INFO] GE(5051,python3.7):2024-01-11-05:50:12.021.671 [pass_manager.cc:82][EVENT]5051 Run:[GEPERFTRACE] The time cost of SubexpressionMigrationPass is [4] micro second. [INFO] GE(5051,python3.7):2024-01-11-05:50:12.021.685 [pass_manager.cc:82][EVENT]5051 Run:[GEPERFTRACE] The time cost of UnusedArgsCleanPass is [2] micro second. [INFO] GE(5051,python3.7):2024-01-11-05:50:12.021.697 [graph_manager.cc:1082][EVENT]5051 PreRunOptimizeOriginalGraph:[GEPERFTRACE] The time cost of GraphManager::SubexpressionMigration is [36] micro second. [INFO] GE(5051,python3.7):2024-01-11-05:50:12.021.729 [pass_manager.cc:82][EVENT]5051 Run:[GEPERFTRACE] The time cost of OptimizeStage1_1::MergeInputMemcpyPass is [5] micro second. [INFO] GE(5051,python3.7):2024-01-11-05:50:12.021.745 [pass_manager.cc:82][EVENT]5051 Run:[GEPERFTRACE] The time cost of OptimizeStage1_1::SwitchDataEdgesBypass is [4] micro second. [INFO] GE(5051,python3.7):2024-01-11-05:50:12.021.759 [pass_manager.cc:82][EVENT]5051 Run:[GEPERFTRACE] The time cost of OptimizeStage1_1::ConstantFuseSamePass is [5] micro second. [INFO] GE(5051,python3.7):2024-01-11-05:50:12.021.874 [pass_manager.cc:82][EVENT]5051 Run:[GEPERFTRACE] The time cost of OptimizeStage1_1::CSEBeforeFuseDataNodesWithCommonInputPass is [104] micro second. [INFO] GE(5051,python3.7):2024-01-11-05:50:12.021.893 [pass_manager.cc:82][EVENT]5051 Run:[GEPERFTRACE] The time cost of OptimizeStage1_1::FuseDataNodesWithCommonInputPass is [3] micro second. [INFO] GE(5051,python3.7):2024-01-11-05:50:12.021.958 [pass_manager.cc:82][EVENT]5051 Run:[GEPERFTRACE] The time cost of OptimizeStage1_1::CommonSubexpressionEliminationPass is [41] micro second. [INFO] GE(5051,python3.7):2024-01-11-05:50:12.021.975 [pass_manager.cc:82][EVENT]5051 Run:[GEPERFTRACE] The time cost of OptimizeStage1_1::PermutePass is [3] micro second. [INFO] GE(5051,python3.7):2024-01-11-05:50:12.022.021 [pass_manager.cc:82][EVENT]5051 Run:[GEPERFTRACE] The time cost of OptimizeStage1_1::SameTransdataBreadthFusionPass is [34] micro second. [INFO] GE(5051,python3.7):2024-01-11-05:50:12.022.040 [pass_manager.cc:82][EVENT]5051 Run:[GEPERFTRACE] The time cost of OptimizeStage1_1::VariableOpPass is [8] micro second. [INFO] GE(5051,python3.7):2024-01-11-05:50:12.022.058 [pass_manager.cc:82][EVENT]5051 Run:[GEPERFTRACE] The time cost of OptimizeStage1_1::TransOpWithoutReshapeFusionPass is [6] micro second. [INFO] GE(5051,python3.7):2024-01-11-05:50:12.022.175 [pass_manager.cc:82][EVENT]5051 Run:[GEPERFTRACE] The time cost of OptimizeStage1_1::TransOpBreadthFusionPass is [107] micro second. [INFO] GE(5051,python3.7):2024-01-11-05:50:12.022.213 [pass_manager.cc:82][EVENT]5051 Run:[GEPERFTRACE] The time cost of OptimizeStage1_1::DataFlowPreparePass is [25] micro second. [INFO] GE(5051,python3.7):2024-01-11-05:50:12.022.227 [pass_manager.cc:82][EVENT]5051 Run:[GEPERFTRACE] The time cost of OptimizeStage1_1::MergeUnknownShapeNPass is [2] micro second. [INFO] GE(5051,python3.7):2024-01-11-05:50:12.022.237 [graph_manager.cc:2700][EVENT]5051 OptimizeStage1:[GEPERFTRACE] The time cost of GraphManager::OptimizeStage1_1 is [513] micro second. [INFO] GE(5051,python3.7):2024-01-11-05:50:12.022.422 [base_pass.cc:339][EVENT]5051 Run:[GEPERFTRACE] The time cost of EnterPass is [2] micro second, call num is [3] [INFO] GE(5051,python3.7):2024-01-11-05:50:12.022.438 [base_pass.cc:339][EVENT]5051 Run:[GEPERFTRACE] The time cost of AddNPass is [1] micro second, call num is [3] [INFO] GE(5051,python3.7):2024-01-11-05:50:12.022.448 [base_pass.cc:339][EVENT]5051 Run:[GEPERFTRACE] The time cost of SwitchDeadBranchElimination is [2] micro second, call num is [3] [INFO] GE(5051,python3.7):2024-01-11-05:50:12.022.457 [base_pass.cc:339][EVENT]5051 Run:[GEPERFTRACE] The time cost of SwitchLogicRemovePass is [1] micro second, call num is [3] [INFO] GE(5051,python3.7):2024-01-11-05:50:12.022.465 [base_pass.cc:339][EVENT]5051 Run:[GEPERFTRACE] The time cost of MergePass is [0] micro second, call num is [3] [INFO] GE(5051,python3.7):2024-01-11-05:50:12.022.474 [base_pass.cc:339][EVENT]5051 Run:[GEPERFTRACE] The time cost of CastRemovePass is [20] micro second, call num is [3] [INFO] GE(5051,python3.7):2024-01-11-05:50:12.022.482 [base_pass.cc:339][EVENT]5051 Run:[GEPERFTRACE] The time cost of TransposeTransDataPass is [3] micro second, call num is [3] [INFO] GE(5051,python3.7):2024-01-11-05:50:12.022.491 [base_pass.cc:339][EVENT]5051 Run:[GEPERFTRACE] The time cost of ReshapeRemovePass is [27] micro second, call num is [3] [INFO] GE(5051,python3.7):2024-01-11-05:50:12.022.499 [base_pass.cc:339][EVENT]5051 Run:[GEPERFTRACE] The time cost of TransOpSymmetryEliminationPass is [25] micro second, call num is [3] [INFO] GE(5051,python3.7):2024-01-11-05:50:12.022.507 [base_pass.cc:339][EVENT]5051 Run:[GEPERFTRACE] The time cost of TransOpNearbyAllreduceFusionPass is [3] micro second, call num is [3] [INFO] GE(5051,python3.7):2024-01-11-05:50:12.022.515 [base_pass.cc:339][EVENT]5051 Run:[GEPERFTRACE] The time cost of ReplaceWithEmptyConstPass is [8] micro second, call num is [3] [INFO] GE(5051,python3.7):2024-01-11-05:50:12.022.523 [base_pass.cc:339][EVENT]5051 Run:[GEPERFTRACE] The time cost of DimensionComputePass is [6] micro second, call num is [3] [INFO] GE(5051,python3.7):2024-01-11-05:50:12.022.531 [base_pass.cc:339][EVENT]5051 Run:[GEPERFTRACE] The time cost of ConstantFoldingPass is [11] micro second, call num is [3] [INFO] GE(5051,python3.7):2024-01-11-05:50:12.022.539 [base_pass.cc:339][EVENT]5051 Run:[GEPERFTRACE] The time cost of DimensionAdjustPass is [5] micro second, call num is [3] [INFO] GE(5051,python3.7):2024-01-11-05:50:12.022.556 [base_pass.cc:339][EVENT]5051 Run:[GEPERFTRACE] The time cost of UselessControlOutRemovePass is [2] micro second, call num is [3] [INFO] GE(5051,python3.7):2024-01-11-05:50:12.022.567 [graph_manager.cc:2741][EVENT]5051 OptimizeStage1:[GEPERFTRACE] The time cost of GraphManager::OptimizeStage1_2 is [310] micro second. [INFO] GE(5051,python3.7):2024-01-11-05:50:12.022.576 [graph_manager.cc:2752][EVENT]5051 OptimizeStage1:[GEPERFTRACE] The time cost of extern constant folding is [0] micro second. [INFO] GE(5051,python3.7):2024-01-11-05:50:12.022.601 [pass_manager.cc:82][EVENT]5051 Run:[GEPERFTRACE] The time cost of OptimizeStage1_3::Migration is [4] micro second. [INFO] GE(5051,python3.7):2024-01-11-05:50:12.022.615 [pass_manager.cc:82][EVENT]5051 Run:[GEPERFTRACE] The time cost of OptimizeStage1_3::ArgsClean is [1] micro second. [INFO] GE(5051,python3.7):2024-01-11-05:50:12.022.633 [pass_manager.cc:82][EVENT]5051 Run:[GEPERFTRACE] The time cost of OptimizeStage1_3::PrunePass is [8] micro second. [INFO] GE(5051,python3.7):2024-01-11-05:50:12.022.649 [pass_manager.cc:82][EVENT]5051 Run:[GEPERFTRACE] The time cost of OptimizeStage1_3::NextIterationPass is [5] micro second. [INFO] GE(5051,python3.7):2024-01-11-05:50:12.022.661 [pass_manager.cc:82][EVENT]5051 Run:[GEPERFTRACE] The time cost of OptimizeStage1_3::ControlTriggerPass is [2] micro second. [INFO] GE(5051,python3.7):2024-01-11-05:50:12.022.674 [pass_manager.cc:82][EVENT]5051 Run:[GEPERFTRACE] The time cost of OptimizeStage1_3::MergeToStreamMergePass is [2] micro second. [INFO] GE(5051,python3.7):2024-01-11-05:50:12.022.701 [pass_manager.cc:82][EVENT]5051 Run:[GEPERFTRACE] The time cost of OptimizeStage1_3::SwitchToStreamSwitchPass is [17] micro second. [INFO] GE(5051,python3.7):2024-01-11-05:50:12.022.716 [pass_manager.cc:82][EVENT]5051 Run:[GEPERFTRACE] The time cost of OptimizeStage1_3::AttachStreamLabelPass is [4] micro second. [INFO] GE(5051,python3.7):2024-01-11-05:50:12.022.730 [pass_manager.cc:82][EVENT]5051 Run:[GEPERFTRACE] The time cost of OptimizeStage1_3::MultiBatchPass is [4] micro second. [INFO] GE(5051,python3.7):2024-01-11-05:50:12.022.740 [pass_manager.cc:82][EVENT]5051 Run:[GEPERFTRACE] The time cost of OptimizeStage1_3::SubgraphMultiDimsPass is [1] micro second. [INFO] GE(5051,python3.7):2024-01-11-05:50:12.022.752 [pass_manager.cc:82][EVENT]5051 Run:[GEPERFTRACE] The time cost of OptimizeStage1_3::IteratorOpPass is [4] micro second. [INFO] GE(5051,python3.7):2024-01-11-05:50:12.022.763 [pass_manager.cc:82][EVENT]5051 Run:[GEPERFTRACE] The time cost of OptimizeStage1_3::VariableRefUselessControlOutDeletePass is [2] micro second. [INFO] GE(5051,python3.7):2024-01-11-05:50:12.022.781 [pass_manager.cc:82][EVENT]5051 Run:[GEPERFTRACE] The time cost of OptimizeStage1_3::ReshapeRecoveryPass is [8] micro second. [INFO] GE(5051,python3.7):2024-01-11-05:50:12.022.793 [pass_manager.cc:82][EVENT]5051 Run:[GEPERFTRACE] The time cost of OptimizeStage1_3::RemoveSameConstPass is [3] micro second. [INFO] GE(5051,python3.7):2024-01-11-05:50:12.022.801 [graph_manager.cc:2810][EVENT]5051 OptimizeStage1:[GEPERFTRACE] The time cost of GraphManager::OptimizeStage1_3 is [205] micro second. [INFO] GE(5051,python3.7):2024-01-11-05:50:12.022.829 [base_pass.cc:339][EVENT]5051 Run:[GEPERFTRACE] The time cost of IdentityPass is [4] micro second, call num is [3] [INFO] GE(5051,python3.7):2024-01-11-05:50:12.022.842 [graph_manager.cc:2821][EVENT]5051 OptimizeStage1:[GEPERFTRACE] The time cost of GraphPrepare::node_pass is [32] micro second. [INFO] GE(5051,python3.7):2024-01-11-05:50:12.022.872 [graph_manager.cc:1087][EVENT]5051 PreRunOptimizeOriginalGraph:[GEPERFTRACE] The time cost of GraphManager::OptimizeStage1 is [1155] micro second. [INFO] GE(5051,python3.7):2024-01-11-05:50:12.023.006 [graph_manager.cc:1088][EVENT]5051 PreRunOptimizeOriginalGraph:[GEPERFTRACE] The time cost of GraphManager::stages.optimizer.OptimizeAfterStage1 is [120] micro second. [INFO] GE(5051,python3.7):2024-01-11-05:50:12.023.057 [graph_manager.cc:1089][EVENT]5051 PreRunOptimizeOriginalGraph:[GEPERFTRACE] The time cost of GraphManager::GraphUtilsEx::InferShapeInNeed is [21] micro second. [INFO] GE(5051,python3.7):2024-01-11-05:50:12.023.075 [pass_manager.cc:82][EVENT]5051 Run:[GEPERFTRACE] The time cost of PreRun::CtrlEdgeTransferPass is [2] micro second. [INFO] GE(5051,python3.7):2024-01-11-05:50:12.023.090 [graph_manager.cc:1097][EVENT]5051 PreRunOptimizeOriginalGraph:PreRun:PreRunOptimizeOriginalGraph success. [INFO] GE(5051,python3.7):2024-01-11-05:50:12.023.112 [graph_manager.cc:3325][EVENT]5051 OptimizeSubgraph:[GEPERFTRACE] The time cost of OptimizeSubgraph::StagePartition is [4] micro second. [INFO] GE(5051,python3.7):2024-01-11-05:50:12.026.133 [engine_place.cc:144][EVENT]5051 Run:The time cost of AIcoreEngine::CheckSupported is [2865] micro second. [INFO] GE(5051,python3.7):2024-01-11-05:50:12.026.163 [engine_place.cc:144][EVENT]5051 Run:The time cost of DNN_VM_GE_LOCAL_OP_STORE::CheckSupported is [6] micro second. [INFO] GE(5051,python3.7):2024-01-11-05:50:12.026.173 [engine_place.cc:144][EVENT]5051 Run:The time cost of DNN_VM_RTS_OP_STORE::CheckSupported is [8] micro second. [INFO] GE(5051,python3.7):2024-01-11-05:50:12.026.314 [graph_manager.cc:3351][EVENT]5051 OptimizeSubgraph:[GEPERFTRACE] The time cost of OptimizeSubgraph::GraphPartitionDynamicShape is [3189] micro second. [INFO] GE(5051,python3.7):2024-01-11-05:50:12.026.334 [graph_manager.cc:3364][EVENT]5051 OptimizeSubgraph:[GEPERFTRACE] The time cost of OptimizeSubgraph::SubgraphPartitionAndOptimization::CompositeEngine is [2] micro second. [INFO] GE(5051,python3.7):2024-01-11-05:50:12.026.400 [engine_partitioner.cc:1139][EVENT]5051 PartitionSubGraph:[GEPERFTRACE] The time cost of EnginePartitioner::PartitionInitialize is [20] micro second. [INFO] GE(5051,python3.7):2024-01-11-05:50:12.026.417 [engine_partitioner.cc:1142][EVENT]5051 PartitionSubGraph:[GEPERFTRACE] The time cost of EnginePartitioner::PartitionMarkClusters is [5] micro second. [INFO] GE(5051,python3.7):2024-01-11-05:50:12.026.558 [engine_partitioner.cc:1148][EVENT]5051 PartitionSubGraph:[GEPERFTRACE] The time cost of EnginePartitioner::PartitionSplitSubGraphs is [130] micro second. [INFO] GE(5051,python3.7):2024-01-11-05:50:12.026.601 [engine_partitioner.cc:1155][EVENT]5051 PartitionSubGraph:[GEPERFTRACE] The time cost of EnginePartitioner::PartitionSortSubGraphs is [31] micro second. [INFO] GE(5051,python3.7):2024-01-11-05:50:12.026.655 [engine_partitioner.cc:1164][EVENT]5051 PartitionSubGraph:[GEPERFTRACE] The time cost of EnginePartitioner::PartitionAddPartitionsToGraphNode is [42] micro second. [INFO] GE(5051,python3.7):2024-01-11-05:50:12.026.687 [graph_manager.cc:3405][EVENT]5051 SubgraphPartitionAndOptimization:[GEPERFTRACE] The time cost of OptimizeSubgraph::Partition1 is [339] micro second. [INFO] GE(5051,python3.7):2024-01-11-05:50:12.026.707 [graph_manager.cc:3412][EVENT]5051 SubgraphPartitionAndOptimization:[GEPERFTRACE] The time cost of OptimizeSubgraph::SetSubgraphPreProc is [7] micro second. [INFO] GE(5051,python3.7):2024-01-11-05:50:12.036.485 [graph_manager.cc:3422][EVENT]5051 SubgraphPartitionAndOptimization:[GEPERFTRACE] The time cost of OptimizeSubgraph::SetSubGraph is [9763] micro second. [INFO] GE(5051,python3.7):2024-01-11-05:50:12.036.523 [graph_manager.cc:3428][EVENT]5051 SubgraphPartitionAndOptimization:[GEPERFTRACE] The time cost of OptimizeSubgraph::SetSubgraphPostProc is [8] micro second. [INFO] GE(5051,python3.7):2024-01-11-05:50:12.036.653 [graph_manager.cc:3467][EVENT]5051 SubgraphPartitionAndOptimization:[GEPERFTRACE] The time cost of OptimizeSubgraph::MergeSubGraph is [110] micro second. [INFO] GE(5051,python3.7):2024-01-11-05:50:12.036.673 [graph_manager.cc:3377][EVENT]5051 OptimizeSubgraph:[GEPERFTRACE] The time cost of OptimizeSubgraph::SubgraphPartitionAndOptimization::AtomicEngine is [10327] micro second. [INFO] GE(5051,python3.7):2024-01-11-05:50:12.036.690 [graph_manager.cc:1106][EVENT]5051 PreRunOptimizeSubGraph:[GEPERFTRACE] The time cost of GraphManager::OptimizeSubgraph is [13585] micro second. [INFO] GE(5051,python3.7):2024-01-11-05:50:12.036.722 [graph_manager.cc:1115][EVENT]5051 PreRunOptimizeSubGraph:PreRun:PreRunOptimizeSubGraph success. [INFO] GE(5051,python3.7):2024-01-11-05:50:12.036.747 [graph_manager.cc:1130][EVENT]5051 PreRunAfterOptimizeSubGraph:[GEPERFTRACE] The time cost of GraphManager::stages.optimizer.ReplacePrecompiledNodeWithOmGraph is [5] micro second. [INFO] GE(5051,python3.7):2024-01-11-05:50:12.036.779 [graph_manager.cc:1131][EVENT]5051 PreRunAfterOptimizeSubGraph:[GEPERFTRACE] The time cost of GraphManager::stages.optimizer.OptimizeWholeGraph is [19] micro second. [INFO] GE(5051,python3.7):2024-01-11-05:50:12.036.836 [pass_manager.cc:82][EVENT]5051 Run:[GEPERFTRACE] The time cost of OptimizeStage2::AfterMergePasses::LinkGenMaskNodesPass is [37] micro second. [INFO] GE(5051,python3.7):2024-01-11-05:50:12.036.853 [pass_manager.cc:82][EVENT]5051 Run:[GEPERFTRACE] The time cost of OptimizeStage2::HcclContinuousMemcpyPass is [6] micro second. [INFO] GE(5051,python3.7):2024-01-11-05:50:12.036.864 [graph_manager.cc:2837][EVENT]5051 OptimizeStage2:[GEPERFTRACE] The time cost of OptimizeStage2::AfterMergePasses is [67] micro second. [INFO] GE(5051,python3.7):2024-01-11-05:50:12.036.946 [base_pass.cc:339][EVENT]5051 Run:[GEPERFTRACE] The time cost of ConstantFoldingPass is [13] micro second, call num is [3] [INFO] GE(5051,python3.7):2024-01-11-05:50:12.036.959 [base_pass.cc:339][EVENT]5051 Run:[GEPERFTRACE] The time cost of ReshapeRemovePass is [1] micro second, call num is [3] [INFO] GE(5051,python3.7):2024-01-11-05:50:12.036.968 [base_pass.cc:339][EVENT]5051 Run:[GEPERFTRACE] The time cost of CondRemovePass is [4] micro second, call num is [3] [INFO] GE(5051,python3.7):2024-01-11-05:50:12.036.977 [base_pass.cc:339][EVENT]5051 Run:[GEPERFTRACE] The time cost of BitcastPass is [1] micro second, call num is [3] [INFO] GE(5051,python3.7):2024-01-11-05:50:12.036.985 [base_pass.cc:339][EVENT]5051 Run:[GEPERFTRACE] The time cost of AssignRemovePass is [4] micro second, call num is [3] [INFO] GE(5051,python3.7):2024-01-11-05:50:12.036.993 [base_pass.cc:339][EVENT]5051 Run:[GEPERFTRACE] The time cost of DimensionAdjustPass is [8] micro second, call num is [3] [INFO] GE(5051,python3.7):2024-01-11-05:50:12.037.003 [graph_manager.cc:2864][EVENT]5051 OptimizeStage2:[GEPERFTRACE] The time cost of OptimizeStage2::MergedGraphNameToPasses is [123] micro second. [INFO] GE(5051,python3.7):2024-01-11-05:50:12.037.015 [graph_manager.cc:2872][EVENT]5051 OptimizeStage2:[GEPERFTRACE] The time cost of OptimizeStage2::RemoveIsolatedConst is [3] micro second. [INFO] GE(5051,python3.7):2024-01-11-05:50:12.037.034 [pass_manager.cc:82][EVENT]5051 Run:[GEPERFTRACE] The time cost of OptimizeStage2::ControlAttrOptimize::FlowCtrlPass is [4] micro second. [INFO] GE(5051,python3.7):2024-01-11-05:50:12.037.048 [pass_manager.cc:82][EVENT]5051 Run:[GEPERFTRACE] The time cost of OptimizeStage2::ControlAttrOptimize::MultiBatchPass is [2] micro second. [INFO] GE(5051,python3.7):2024-01-11-05:50:12.037.062 [pass_manager.cc:82][EVENT]5051 Run:[GEPERFTRACE] The time cost of OptimizeStage2::AfterMergePasses::RefIdentityDeleteOpPass is [4] micro second. [INFO] GE(5051,python3.7):2024-01-11-05:50:12.037.078 [pass_manager.cc:82][EVENT]5051 Run:[GEPERFTRACE] The time cost of OptimizeStage2::AfterMergePasses::VariableRefDeleteOpPass is [5] micro second. [INFO] GE(5051,python3.7):2024-01-11-05:50:12.037.092 [compile_nodes_pass.cc:88][EVENT]5051 Run:[GEPERFTRACE] The time cost of OptimizeStage2::ControlAttrOptimize::CompileNodesPass is [3] micro second. [INFO] GE(5051,python3.7):2024-01-11-05:50:12.037.102 [pass_manager.cc:82][EVENT]5051 Run:[GEPERFTRACE] The time cost of OptimizeStage2::ControlAttrOptimize::CompileNodesPass is [14] micro second. [INFO] GE(5051,python3.7):2024-01-11-05:50:12.037.112 [pass_manager.cc:82][EVENT]5051 Run:[GEPERFTRACE] The time cost of OptimizeStage2::AfterMergePasses::SwapSpacePass is [2] micro second. [INFO] GE(5051,python3.7):2024-01-11-05:50:12.037.249 [pass_manager.cc:82][EVENT]5051 Run:[GEPERFTRACE] The time cost of OptimizeStage2::AfterMergePasses::InputOutputConnectionIdentifyPass is [83] micro second. [INFO] GE(5051,python3.7):2024-01-11-05:50:12.037.304 [pass_manager.cc:82][EVENT]5051 Run:[GEPERFTRACE] The time cost of OptimizeStage2::AfterMergePasses::AtomicAddrCleanPass is [33] micro second. [INFO] GE(5051,python3.7):2024-01-11-05:50:12.037.317 [pass_manager.cc:82][EVENT]5051 Run:[GEPERFTRACE] The time cost of OptimizeStage2::AfterMergePasses::EndOfSequenceAddControlPass is [3] micro second. [INFO] GE(5051,python3.7):2024-01-11-05:50:12.037.331 [pass_manager.cc:82][EVENT]5051 Run:[GEPERFTRACE] The time cost of OptimizeStage2::ControlAttrOptimize::SubgraphPass is [3] micro second. [INFO] GE(5051,python3.7):2024-01-11-05:50:12.037.345 [pass_manager.cc:82][EVENT]5051 Run:[GEPERFTRACE] The time cost of OptimizeStage2::ControlAttrOptimize::AttachStreamLabelPass is [3] micro second. [INFO] GE(5051,python3.7):2024-01-11-05:50:12.037.355 [graph_manager.cc:2927][EVENT]5051 OptimizeStage2:[GEPERFTRACE] The time cost of OptimizeStage2::ControlAttrOptimize is [325] micro second. [INFO] GE(5051,python3.7):2024-01-11-05:50:12.037.384 [graph_manager.cc:2937][EVENT]5051 OptimizeStage2:[GEPERFTRACE] The time cost of ModelBuilder::AssignFunctionalLabels is [21] micro second. [INFO] GE(5051,python3.7):2024-01-11-05:50:12.037.425 [graph_manager.cc:2943][EVENT]5051 OptimizeStage2:[GEPERFTRACE] The time cost of MemcpyAddrAsyncPass::Run. is [30] micro second. [INFO] GE(5051,python3.7):2024-01-11-05:50:12.037.437 [graph_manager.cc:2950][EVENT]5051 OptimizeStage2:[GEPERFTRACE] The time cost of BufferPoolMemoryPass::Run. is [2] micro second. [INFO] GE(5051,python3.7):2024-01-11-05:50:12.047.249 [graph_manager.cc:2958][EVENT]5051 OptimizeStage2:[GEPERFTRACE] The time cost of ParallelGroupPass::Run. is [47] micro second. [INFO] GE(5051,python3.7):2024-01-11-05:50:12.047.306 [graph_manager.cc:1132][EVENT]5051 PreRunAfterOptimizeSubGraph:[GEPERFTRACE] The time cost of GraphManager::OptimizeStage2 is [10512] micro second. [INFO] GE(5051,python3.7):2024-01-11-05:50:12.047.388 [graph_manager.cc:1135][EVENT]5051 PreRunAfterOptimizeSubGraph:[GEPERFTRACE] The time cost of GraphManager::GetCompilerStages(graph_node->GetGraphId()).optimizer.OptimizeGraphBeforeBuild is [65] micro second. [INFO] GE(5051,python3.7):2024-01-11-05:50:12.047.440 [graph_manager.cc:2975][EVENT]5051 MemConflictProc:[GEPERFTRACE] The time cost of HandleMemoryRWConflict is [34] micro second. [INFO] GE(5051,python3.7):2024-01-11-05:50:12.047.523 [graph_manager.cc:2981][EVENT]5051 MemConflictProc:[GEPERFTRACE] The time cost of MemLayoutConflictOptimizer::Run. is [69] micro second. [INFO] GE(5051,python3.7):2024-01-11-05:50:12.047.540 [pass_manager.cc:82][EVENT]5051 Run:[GEPERFTRACE] The time cost of OptimizeStage2::SetFftsPlusAttrPass is [1] micro second. [INFO] GE(5051,python3.7):2024-01-11-05:50:12.047.550 [graph_manager.cc:2986][EVENT]5051 MemConflictProc:[GEPERFTRACE] The time cost of SetFftsPlusAttrPass::last_passes.Run is [16] micro second. [INFO] GE(5051,python3.7):2024-01-11-05:50:12.047.560 [graph_manager.cc:1136][EVENT]5051 PreRunAfterOptimizeSubGraph:[GEPERFTRACE] The time cost of GraphManager::MemConflictProc is [154] micro second. [INFO] GE(5051,python3.7):2024-01-11-05:50:12.047.713 [graph_manager.cc:3555][EVENT]5051 Build:[GEPERFTRACE] The time cost of GraphManager::RecoverIrDefinitionAndModifyAippData is [117] micro second. [INFO] GE(5051,python3.7):2024-01-11-05:50:12.047.859 [engine_partitioner.cc:1139][EVENT]5051 PartitionSubGraph:[GEPERFTRACE] The time cost of EnginePartitioner::PartitionInitialize is [18] micro second. [INFO] GE(5051,python3.7):2024-01-11-05:50:12.047.876 [engine_partitioner.cc:1142][EVENT]5051 PartitionSubGraph:[GEPERFTRACE] The time cost of EnginePartitioner::PartitionMarkClusters is [2] micro second. [INFO] GE(5051,python3.7):2024-01-11-05:50:12.047.984 [engine_partitioner.cc:1148][EVENT]5051 PartitionSubGraph:[GEPERFTRACE] The time cost of EnginePartitioner::PartitionSplitSubGraphs is [98] micro second. [INFO] GE(5051,python3.7):2024-01-11-05:50:12.048.029 [engine_partitioner.cc:1155][EVENT]5051 PartitionSubGraph:[GEPERFTRACE] The time cost of EnginePartitioner::PartitionSortSubGraphs is [22] micro second. [INFO] GE(5051,python3.7):2024-01-11-05:50:12.048.070 [engine_partitioner.cc:1164][EVENT]5051 PartitionSubGraph:[GEPERFTRACE] The time cost of EnginePartitioner::PartitionAddPartitionsToGraphNode is [30] micro second. [INFO] GE(5051,python3.7):2024-01-11-05:50:12.048.094 [graph_builder.cc:865][EVENT]5051 SecondPartition:[GEPERFTRACE] The time cost of EnginePartitioner::Partition2 is [278] micro second. [INFO] RUNTIME(5051,python3.7):2024-01-11-05:50:12.048.618 [logger.cc:1071] 5051 ModelBindStream: model_id=320, stream_id=577, flag=0. [INFO] GE(5051,python3.7):2024-01-11-05:50:12.048.686 [task_generator.cc:804][EVENT]5051 GenerateTask:[GEPERFTRACE] The time cost of TaskGenerator::SetStreamCtx is [213] micro second. [INFO] GE(5051,python3.7):2024-01-11-05:50:12.048.785 [task_generator.cc:805][EVENT]5051 GenerateTask:[GEPERFTRACE] The time cost of TaskGenerator::PrepareForGenerateTask is [82] micro second. [INFO] GE(5051,python3.7):2024-01-11-05:50:12.049.734 [task_generator.cc:814][EVENT]5051 GenerateTask:[GEPERFTRACE] The time cost of TaskGenerator::DoGenerateTask is [922] micro second. [INFO] GE(5051,python3.7):2024-01-11-05:50:12.049.754 [task_generator.cc:954][EVENT]5051 GetTaskInfo:[GEPERFTRACE] The time cost of TaskGenerator::GenerateTask is [1283] micro second. [INFO] GE(5051,python3.7):2024-01-11-05:50:12.049.824 [task_generator.cc:967][EVENT]5051 GetTaskInfo:[GEPERFTRACE] The time cost of TaskGenerator::AddModelTaskToModel is [43] micro second. [INFO] RUNTIME(5051,python3.7):2024-01-11-05:50:12.049.844 [logger.cc:1084] 5051 ModelUnbindStream: model_id=320, stream_id=577, [INFO] GE(5051,python3.7):2024-01-11-05:50:12.050.053 [graph_manager.cc:1152][EVENT]5051 PreRunAfterOptimizeSubGraph:[GEPERFTRACE] The time cost of GraphManager::Build is [2465] micro second. [INFO] GE(5051,python3.7):2024-01-11-05:50:12.050.075 [graph_manager.cc:1164][EVENT]5051 PreRunAfterOptimizeSubGraph:PreRun:PreRunAfterOptimizeSubGraph success. [INFO] GE(5051,python3.7):2024-01-11-05:50:12.050.117 [graph_manager.cc:1271][EVENT]5051 PreRun:[GEPERFTRACE] The time cost of FlowModelBuild is [69357] micro second. [INFO] GE(5051,python3.7):2024-01-11-05:50:12.050.128 [graph_manager.cc:1272][EVENT]5051 PreRun:[GEPERFTRACE] GE PreRun End [INFO] ATRACE(5051,python3.7):2024-01-11-05:50:12.050.452 [atrace_api.c:93](tid:5051) AtraceDestroy start [INFO] ATRACE(5051,python3.7):2024-01-11-05:50:12.050.475 [atrace_api.c:95](tid:5051) AtraceDestroy end [INFO] GE(5051,python3.7):2024-01-11-05:50:12.055.865 [graph_converter.cc:838][EVENT]5051 ConvertComputeGraphToExecuteGraph:[GEPERFTRACE] The time cost of ConvertComputeGraphToExecuteGraph::CreateMainNode is [1640] micro second. [INFO] GE(5051,python3.7):2024-01-11-05:50:12.056.054 [base_optimizer.cc:70][EVENT]5051 Run:[GEPERFTRACE] The time cost of ZeroCopy is [144] micro second. [INFO] GE(5051,python3.7):2024-01-11-05:50:12.056.567 [base_optimizer.cc:70][EVENT]5051 Run:[GEPERFTRACE] The time cost of CEM is [490] micro second. [INFO] GE(5051,python3.7):2024-01-11-05:50:12.056.786 [copy_flow_launch_fuse.cc:395][EVENT]5051 Run:[GEPERFTRACE] The time cost of Pass::CopyFlowLaunchFuse is [194] micro second. [INFO] GE(5051,python3.7):2024-01-11-05:50:12.056.807 [base_optimizer.cc:70][EVENT]5051 Run:[GEPERFTRACE] The time cost of CopyFlowLaunch is [217] micro second. [INFO] GE(5051,python3.7):2024-01-11-05:50:12.057.063 [base_optimizer.cc:70][EVENT]5051 Run:[GEPERFTRACE] The time cost of TrustOutTensor is [243] micro second. [INFO] GE(5051,python3.7):2024-01-11-05:50:12.057.105 [base_optimizer.cc:70][EVENT]5051 Run:[GEPERFTRACE] The time cost of AicpuFuseHostInputs is [22] micro second. [INFO] GE(5051,python3.7):2024-01-11-05:50:12.057.169 [base_optimizer.cc:70][EVENT]5051 Run:[GEPERFTRACE] The time cost of ZeroCopy is [27] micro second. [INFO] GE(5051,python3.7):2024-01-11-05:50:12.057.367 [base_optimizer.cc:70][EVENT]5051 Run:[GEPERFTRACE] The time cost of CEM is [182] micro second. [INFO] GE(5051,python3.7):2024-01-11-05:50:12.057.464 [copy_flow_launch_fuse.cc:395][EVENT]5051 Run:[GEPERFTRACE] The time cost of Pass::CopyFlowLaunchFuse is [70] micro second. [INFO] GE(5051,python3.7):2024-01-11-05:50:12.057.478 [base_optimizer.cc:70][EVENT]5051 Run:[GEPERFTRACE] The time cost of CopyFlowLaunch is [85] micro second. [INFO] GE(5051,python3.7):2024-01-11-05:50:12.057.506 [base_optimizer.cc:70][EVENT]5051 Run:[GEPERFTRACE] The time cost of TrustOutTensor is [19] micro second. [INFO] GE(5051,python3.7):2024-01-11-05:50:12.057.516 [base_optimizer.cc:70][EVENT]5051 Run:[GEPERFTRACE] The time cost of AicpuFuseHostInputs is [0] micro second. [INFO] GE(5051,python3.7):2024-01-11-05:50:12.057.541 [base_optimizer.cc:70][EVENT]5051 Run:[GEPERFTRACE] The time cost of ZeroCopy is [16] micro second. [INFO] GE(5051,python3.7):2024-01-11-05:50:12.057.614 [base_optimizer.cc:70][EVENT]5051 Run:[GEPERFTRACE] The time cost of CEM is [64] micro second. [INFO] GE(5051,python3.7):2024-01-11-05:50:12.057.679 [copy_flow_launch_fuse.cc:395][EVENT]5051 Run:[GEPERFTRACE] The time cost of Pass::CopyFlowLaunchFuse is [53] micro second. [INFO] GE(5051,python3.7):2024-01-11-05:50:12.057.691 [base_optimizer.cc:70][EVENT]5051 Run:[GEPERFTRACE] The time cost of CopyFlowLaunch is [66] micro second. [INFO] GE(5051,python3.7):2024-01-11-05:50:12.057.716 [base_optimizer.cc:70][EVENT]5051 Run:[GEPERFTRACE] The time cost of TrustOutTensor is [17] micro second. [INFO] GE(5051,python3.7):2024-01-11-05:50:12.057.726 [base_optimizer.cc:70][EVENT]5051 Run:[GEPERFTRACE] The time cost of AicpuFuseHostInputs is [1] micro second. [INFO] GE(5051,python3.7):2024-01-11-05:50:12.057.739 [graph_converter.cc:849][EVENT]5051 ConvertComputeGraphToExecuteGraph:[GEPERFTRACE] The time cost of ConvertComputeGraphToExecuteGraph::RunAllPass is [1836] micro second. [INFO] GE(5051,python3.7):2024-01-11-05:50:12.057.947 [graph_converter.cc:853][EVENT]5051 ConvertComputeGraphToExecuteGraph:[GEPERFTRACE] The time cost of ConvertComputeGraphToExecuteGraph::TopologicalSorting is [198] micro second. [INFO] GE(5051,python3.7):2024-01-11-05:50:12.058.722 [graph_converter.cc:857][EVENT]5051 ConvertComputeGraphToExecuteGraph:[GEPERFTRACE] The time cost of ConvertComputeGraphToExecuteGraph::AppendGraphLevelData is [760] micro second. [INFO] GE(5051,python3.7):2024-01-11-05:50:12.058.882 [graph_converter.cc:862][EVENT]5051 ConvertComputeGraphToExecuteGraph:[GEPERFTRACE] The time cost of ConvertComputeGraphToExecuteGraph::CalculatePriority is [134] micro second. TotalTime = 0.0218135, [20] [parse]: 0.0106139 [symbol_resolve]: 0.00080032, [1] [Cycle 1]: 0.00076834, [1] [resolve]: 0.00075168 [combine_like_graphs]: 1.04e-06 [graph_reusing]: 2.66e-06 [meta_unpack_prepare]: 3.56e-05 [pre_cconv]: 1.37999e-06 [abstract_specialize]: 0.00163862 [pack_expand]: 8.58e-06 [auto_monad]: 3.276e-05 [inline]: 1.23e-06 [pre_auto_parallel]: 1.494e-05 [pipeline_split]: 2.55e-06 [optimize]: 0.00788049, [35] [py_interpret_to_execute]: 8.11e-06 [rewriter_before_opt_a]: 3.518e-05 [opt_a]: 0.00715173, [1] [Cycle 1]: 0.00058751, [30] [expand_dump_flag]: 2.95e-06 [switch_simplify]: 7.39e-06 [a_1]: 1.604e-05 [recompute_prepare]: 1.84e-06 [updatestate_depend_eliminate]: 5.88e-06 [updatestate_assign_eliminate]: 3.32999e-06 [updatestate_loads_eliminate]: 2.49e-06 [parameter_eliminate]: 2.71999e-06 [a_2]: 2.619e-05 [accelerated_algorithm]: 2.69e-06 [pynative_shard]: 1.68999e-06 [auto_parallel]: 4.19e-06 [parallel]: 2.007e-05 [merge_comm]: 1.132e-05 [allreduce_fusion]: 1.77e-06 [virtual_dataset]: 2.37e-06 [get_grad_eliminate_]: 1.8e-06 [virtual_output]: 1.58e-06 [merge_forward]: 3.99e-06 [cell_reuse_recompute_pass]: 8.2e-07 [cell_reuse_handle_not_recompute_node_pass]: 7.02e-06 [meta_fg_expand]: 2.49e-06 [after_resolve]: 1.86e-06 [a_after_grad]: 2.14e-06 [renormalize]: 0.0002658 [real_op_eliminate]: 4.31e-06 [auto_monad_grad]: 3.42001e-06 [auto_monad_eliminator]: 8.35001e-06 [cse]: 1.612e-05 [a_3]: 1.359e-05 [py_interpret_to_execute_after_opt_a]: 6.75e-06 [slice_cell_reuse_recomputed_activation]: 2.89e-06 [rewriter_after_opt_a]: 0.00019223 [convert_after_rewriter]: 5.87e-06 [order_py_execute_after_rewriter]: 3.91e-06 [opt_b]: 9.078e-05, [1] [Cycle 1]: 8.544e-05, [7] [b_1]: 3.65e-05 [b_2]: 3.02e-06 [updatestate_depend_eliminate]: 2.36e-06 [updatestate_assign_eliminate]: 1.95e-06 [updatestate_loads_eliminate]: 1.73e-06 [renormalize]: 3.39998e-07 [cse]: 8.39e-06 [cconv]: 1.498e-05 [opt_after_cconv]: 4.888e-05, [1] [Cycle 1]: 4.478e-05, [7] [c_1]: 4.26e-06 [parameter_eliminate]: 6.20006e-07 [updatestate_depend_eliminate]: 1.81e-06 [updatestate_assign_eliminate]: 1.43e-06 [updatestate_loads_eliminate]: 1.41e-06 [cse]: 5.64e-06 [renormalize]: 2.70004e-07 [remove_dup_value]: 1.054e-05 [tuple_transform]: 3.571e-05, [1] [Cycle 1]: 3.167e-05, [3] [d_1]: 1.271e-05 [d_2]: 5.86e-06 [renormalize]: 1.69995e-07 [add_cache_embedding]: 7.75e-06 [add_recomputation]: 2.396e-05 [cse_after_recomputation]: 1.534e-05, [1] [Cycle 1]: 1.117e-05, [1] [cse]: 6.24e-06 [environ_conv]: 2.064e-05 [label_micro_interleaved_index]: 2.67e-06 [label_fine_grained_interleaved_index]: 2.12e-06 [assign_add_opt]: 1.60999e-06 [slice_recompute_activation]: 2.08e-06 [micro_interleaved_order_control]: 1.84e-06 [full_micro_interleaved_order_control]: 1.73e-06 [comp_comm_scheduling]: 2.1e-06 [reorder_send_recv_between_fp_bp]: 2.06e-06 [comm_op_add_attrs]: 1.04e-06 [add_comm_op_reuse_tag]: 8.89995e-07 [overlap_opt_shard_in_pipeline]: 1.06e-06 [grouped_pairwise_exchange_alltoall]: 1.34e-06 [overlap_recompute_and_grad_model_parallel]: 1.74e-06 [overlap_grad_matmul_and_grad_allreduce]: 8.29998e-07 [split_matmul_comm_elemetwise]: 2.68e-06 [split_layernorm_comm]: 1.87e-06 [process_send_recv_for_ge]: 2.14999e-06 [handle_group_info]: 9.5e-07 [auto_monad_reorder]: 1.68e-05 [get_jit_bprop_graph]: 3.69997e-07 [eliminate_special_op_node]: 0.00050844 [validate]: 5.013e-05 [distribtued_split]: 1.05e-06 [task_emit]: 9.20001e-07 [execute]: 1.08e-06 Sums parse : 0.010614s : 72.76% symbol_resolve.resolve : 0.000752s : 5.15% combine_like_graphs : 0.000001s : 0.01% graph_reusing : 0.000003s : 0.02% meta_unpack_prepare : 0.000036s : 0.24% pre_cconv : 0.000001s : 0.01% abstract_specialize : 0.001639s : 11.23% pack_expand : 0.000009s : 0.06% auto_monad : 0.000033s : 0.22% inline : 0.000001s : 0.01% pre_auto_parallel : 0.000015s : 0.10% pipeline_split : 0.000003s : 0.02% optimize.py_interpret_to_execute : 0.000008s : 0.06% optimize.rewriter_before_opt_a : 0.000035s : 0.24% optimize.opt_a.expand_dump_flag : 0.000003s : 0.02% optimize.opt_a.switch_simplify : 0.000007s : 0.05% optimize.opt_a.a_1 : 0.000016s : 0.11% optimize.opt_a.recompute_prepare : 0.000002s : 0.01% optimize.opt_a.updatestate_depend_eliminate : 0.000006s : 0.04% optimize.opt_a.updatestate_assign_eliminate : 0.000003s : 0.02% optimize.opt_a.updatestate_loads_eliminate : 0.000002s : 0.02% optimize.opt_a.parameter_eliminate : 0.000003s : 0.02% optimize.opt_a.a_2 : 0.000026s : 0.18% optimize.opt_a.accelerated_algorithm : 0.000003s : 0.02% optimize.opt_a.pynative_shard : 0.000002s : 0.01% optimize.opt_a.auto_parallel : 0.000004s : 0.03% optimize.opt_a.parallel : 0.000020s : 0.14% optimize.opt_a.merge_comm : 0.000011s : 0.08% optimize.opt_a.allreduce_fusion : 0.000002s : 0.01% optimize.opt_a.virtual_dataset : 0.000002s : 0.02% optimize.opt_a.get_grad_eliminate_ : 0.000002s : 0.01% optimize.opt_a.virtual_output : 0.000002s : 0.01% optimize.opt_a.merge_forward : 0.000004s : 0.03% optimize.opt_a.cell_reuse_recompute_pass : 0.000001s : 0.01% optimize.opt_a.cell_reuse_handle_not_recompute_node_pass : 0.000007s : 0.05% optimize.opt_a.meta_fg_expand : 0.000002s : 0.02% optimize.opt_a.after_resolve : 0.000002s : 0.01% optimize.opt_a.a_after_grad : 0.000002s : 0.01% optimize.opt_a.renormalize : 0.000266s : 1.82% optimize.opt_a.real_op_eliminate : 0.000004s : 0.03% optimize.opt_a.auto_monad_grad : 0.000003s : 0.02% optimize.opt_a.auto_monad_eliminator : 0.000008s : 0.06% optimize.opt_a.cse : 0.000016s : 0.11% optimize.opt_a.a_3 : 0.000014s : 0.09% optimize.py_interpret_to_execute_after_opt_a : 0.000007s : 0.05% optimize.slice_cell_reuse_recomputed_activation : 0.000003s : 0.02% optimize.rewriter_after_opt_a : 0.000192s : 1.32% optimize.convert_after_rewriter : 0.000006s : 0.04% optimize.order_py_execute_after_rewriter : 0.000004s : 0.03% optimize.opt_b.b_1 : 0.000037s : 0.25% optimize.opt_b.b_2 : 0.000003s : 0.02% optimize.opt_b.updatestate_depend_eliminate : 0.000002s : 0.02% optimize.opt_b.updatestate_assign_eliminate : 0.000002s : 0.01% optimize.opt_b.updatestate_loads_eliminate : 0.000002s : 0.01% optimize.opt_b.renormalize : 0.000000s : 0.00% optimize.opt_b.cse : 0.000008s : 0.06% optimize.cconv : 0.000015s : 0.10% optimize.opt_after_cconv.c_1 : 0.000004s : 0.03% optimize.opt_after_cconv.parameter_eliminate : 0.000001s : 0.00% optimize.opt_after_cconv.updatestate_depend_eliminate : 0.000002s : 0.01% optimize.opt_after_cconv.updatestate_assign_eliminate : 0.000001s : 0.01% optimize.opt_after_cconv.updatestate_loads_eliminate : 0.000001s : 0.01% optimize.opt_after_cconv.cse : 0.000006s : 0.04% optimize.opt_after_cconv.renormalize : 0.000000s : 0.00% optimize.remove_dup_value : 0.000011s : 0.07% optimize.tuple_transform.d_1 : 0.000013s : 0.09% optimize.tuple_transform.d_2 : 0.000006s : 0.04% optimize.tuple_transform.renormalize : 0.000000s : 0.00% optimize.add_cache_embedding : 0.000008s : 0.05% optimize.add_recomputation : 0.000024s : 0.16% optimize.cse_after_recomputation.cse : 0.000006s : 0.04% optimize.environ_conv : 0.000021s : 0.14% optimize.label_micro_interleaved_index : 0.000003s : 0.02% optimize.label_fine_grained_interleaved_index : 0.000002s : 0.01% optimize.assign_add_opt : 0.000002s : 0.01% optimize.slice_recompute_activation : 0.000002s : 0.01% optimize.micro_interleaved_order_control : 0.000002s : 0.01% optimize.full_micro_interleaved_order_control : 0.000002s : 0.01% optimize.comp_comm_scheduling : 0.000002s : 0.01% optimize.reorder_send_recv_between_fp_bp : 0.000002s : 0.01% optimize.comm_op_add_attrs : 0.000001s : 0.01% optimize.add_comm_op_reuse_tag : 0.000001s : 0.01% optimize.overlap_opt_shard_in_pipeline : 0.000001s : 0.01% optimize.grouped_pairwise_exchange_alltoall : 0.000001s : 0.01% optimize.overlap_recompute_and_grad_model_parallel : 0.000002s : 0.01% optimize.overlap_grad_matmul_and_grad_allreduce : 0.000001s : 0.01% optimize.split_matmul_comm_elemetwise : 0.000003s : 0.02% optimize.split_layernorm_comm : 0.000002s : 0.01% optimize.process_send_recv_for_ge : 0.000002s : 0.01% optimize.handle_group_info : 0.000001s : 0.01% auto_monad_reorder : 0.000017s : 0.12% get_jit_bprop_graph : 0.000000s : 0.00% eliminate_special_op_node : 0.000508s : 3.49% validate : 0.000050s : 0.34% distribtued_split : 0.000001s : 0.01% task_emit : 0.000001s : 0.01% execute : 0.000001s : 0.01% Time group info: ------[substitution.] 0.000629 21 97.96% : 0.000616s : 11: substitution.getattr_setattr_resolve 0.78% : 0.000005s : 2: substitution.graph_param_transform 0.84% : 0.000005s : 5: substitution.meta_unpack_prepare 0.29% : 0.000002s : 2: substitution.partial_unused_args_eliminate 0.14% : 0.000001s : 1: substitution.remove_not_recompute_node ------[renormalize.] 0.000261 2 55.21% : 0.000144s : 1: renormalize.infer 44.79% : 0.000117s : 1: renormalize.specialize ------[replace.] 0.000038 2 100.00% : 0.000038s : 2: replace.getattr_setattr_resolve ------[match.] 0.000375 2 100.00% : 0.000375s : 2: match.getattr_setattr_resolve ------[func_graph_cloner_run.] 0.000114 3 8.97% : 0.000010s : 1: func_graph_cloner_run.FuncGraphClonerGraph 91.03% : 0.000104s : 2: func_graph_cloner_run.FuncGraphSpecializer ------[meta_graph.] 0.000000 0 ------[manager.] 0.000000 0 ------[pynative] 0.000000 0 ------[others.] 0.000850 69 4.05% : 0.000034s : 26: opt.transform.opt_a 3.06% : 0.000026s : 23: opt.transform.opt_b 83.74% : 0.000712s : 2: opt.transform.opt_resolve 1.85% : 0.000016s : 1: opt.transforms.meta_unpack_prepare 3.86% : 0.000033s : 10: opt.transforms.opt_a 0.35% : 0.000003s : 1: opt.transforms.opt_after_cconv 0.23% : 0.000002s : 1: opt.transforms.opt_b 1.98% : 0.000017s : 2: opt.transforms.opt_trans_graph 0.88% : 0.000008s : 3: opt.transforms.special_op_eliminate TotalTime = 0.0525068, [20] [parse]: 0.00220468 [symbol_resolve]: 0.00053353, [1] [Cycle 1]: 0.00050738, [1] [resolve]: 0.00049416 [combine_like_graphs]: 5.9e-07 [graph_reusing]: 2.50999e-06 [meta_unpack_prepare]: 5.831e-05 [pre_cconv]: 2.80001e-07 [abstract_specialize]: 0.0452183 [pack_expand]: 8.58e-06 [auto_monad]: 1.884e-05 [inline]: 1.54e-06 [pre_auto_parallel]: 1.133e-05 [pipeline_split]: 2.53e-06 [optimize]: 0.00375826, [35] [py_interpret_to_execute]: 7.64e-06 [rewriter_before_opt_a]: 2.104e-05 [opt_a]: 0.00308859, [1] [Cycle 1]: 0.00069824, [30] [expand_dump_flag]: 2.32e-06 [switch_simplify]: 8.2e-06 [a_1]: 1.705e-05 [recompute_prepare]: 1.95e-06 [updatestate_depend_eliminate]: 5.89999e-06 [updatestate_assign_eliminate]: 2.3e-06 [updatestate_loads_eliminate]: 1.93e-06 [parameter_eliminate]: 2.43e-06 [a_2]: 2.736e-05 [accelerated_algorithm]: 2.54e-06 [pynative_shard]: 2.01001e-06 [auto_parallel]: 2.91e-06 [parallel]: 6.69e-06 [merge_comm]: 2.45e-06 [allreduce_fusion]: 1.41e-06 [virtual_dataset]: 2.15e-06 [get_grad_eliminate_]: 1.74e-06 [virtual_output]: 1.6e-06 [merge_forward]: 3.15e-06 [cell_reuse_recompute_pass]: 1.1e-06 [cell_reuse_handle_not_recompute_node_pass]: 7.16e-06 [meta_fg_expand]: 2.38e-06 [after_resolve]: 1.7e-06 [a_after_grad]: 2.15e-06 [renormalize]: 0.00039739 [real_op_eliminate]: 5.24e-06 [auto_monad_grad]: 3.14e-06 [auto_monad_eliminator]: 7.46e-06 [cse]: 1.546e-05 [a_3]: 1.429e-05 [py_interpret_to_execute_after_opt_a]: 6.12999e-06 [slice_cell_reuse_recomputed_activation]: 2.03e-06 [rewriter_after_opt_a]: 0.00023256 [convert_after_rewriter]: 5.07e-06 [order_py_execute_after_rewriter]: 3.59e-06 [opt_b]: 9.067e-05, [1] [Cycle 1]: 8.557e-05, [7] [b_1]: 3.641e-05 [b_2]: 3.12e-06 [updatestate_depend_eliminate]: 2.3e-06 [updatestate_assign_eliminate]: 1.72e-06 [updatestate_loads_eliminate]: 2e-06 [renormalize]: 3.20004e-07 [cse]: 8.61e-06 [cconv]: 1.468e-05 [opt_after_cconv]: 4.948e-05, [1] [Cycle 1]: 4.53e-05, [7] [c_1]: 4.63e-06 [parameter_eliminate]: 7.7e-07 [updatestate_depend_eliminate]: 1.87e-06 [updatestate_assign_eliminate]: 1.45e-06 [updatestate_loads_eliminate]: 1.33e-06 [cse]: 5.93e-06 [renormalize]: 2.3e-07 [remove_dup_value]: 8.95e-06 [tuple_transform]: 3.391e-05, [1] [Cycle 1]: 3.046e-05, [3] [d_1]: 1.226e-05 [d_2]: 5.51e-06 [renormalize]: 1.50001e-07 [add_cache_embedding]: 7.45001e-06 [add_recomputation]: 1.946e-05 [cse_after_recomputation]: 1.609e-05, [1] [Cycle 1]: 1.16e-05, [1] [cse]: 6.65e-06 [environ_conv]: 7.71e-06 [label_micro_interleaved_index]: 2.3e-06 [label_fine_grained_interleaved_index]: 2.31e-06 [assign_add_opt]: 1.22e-06 [slice_recompute_activation]: 1.95e-06 [micro_interleaved_order_control]: 1.48e-06 [full_micro_interleaved_order_control]: 1.89e-06 [comp_comm_scheduling]: 1.97e-06 [reorder_send_recv_between_fp_bp]: 1.63e-06 [comm_op_add_attrs]: 1.01e-06 [add_comm_op_reuse_tag]: 8.49999e-07 [overlap_opt_shard_in_pipeline]: 8.2e-07 [grouped_pairwise_exchange_alltoall]: 6.69999e-07 [overlap_recompute_and_grad_model_parallel]: 1.59e-06 [overlap_grad_matmul_and_grad_allreduce]: 5.30003e-07 [split_matmul_comm_elemetwise]: 1.99999e-06 [split_layernorm_comm]: 2.04999e-06 [process_send_recv_for_ge]: 7.90002e-07 [handle_group_info]: 9.60004e-07 [auto_monad_reorder]: 1.249e-05 [get_jit_bprop_graph]: 4.39999e-07 [eliminate_special_op_node]: 0.00047489 [validate]: 3.04e-05 [distribtued_split]: 1.08001e-06 [task_emit]: 9.79999e-07 [execute]: 7.50006e-07 Sums parse : 0.002205s : 4.45% symbol_resolve.resolve : 0.000494s : 1.00% combine_like_graphs : 0.000001s : 0.00% graph_reusing : 0.000003s : 0.01% meta_unpack_prepare : 0.000058s : 0.12% pre_cconv : 0.000000s : 0.00% abstract_specialize : 0.045218s : 91.25% pack_expand : 0.000009s : 0.02% auto_monad : 0.000019s : 0.04% inline : 0.000002s : 0.00% pre_auto_parallel : 0.000011s : 0.02% pipeline_split : 0.000003s : 0.01% optimize.py_interpret_to_execute : 0.000008s : 0.02% optimize.rewriter_before_opt_a : 0.000021s : 0.04% optimize.opt_a.expand_dump_flag : 0.000002s : 0.00% optimize.opt_a.switch_simplify : 0.000008s : 0.02% optimize.opt_a.a_1 : 0.000017s : 0.03% optimize.opt_a.recompute_prepare : 0.000002s : 0.00% optimize.opt_a.updatestate_depend_eliminate : 0.000006s : 0.01% optimize.opt_a.updatestate_assign_eliminate : 0.000002s : 0.00% optimize.opt_a.updatestate_loads_eliminate : 0.000002s : 0.00% optimize.opt_a.parameter_eliminate : 0.000002s : 0.00% optimize.opt_a.a_2 : 0.000027s : 0.06% optimize.opt_a.accelerated_algorithm : 0.000003s : 0.01% optimize.opt_a.pynative_shard : 0.000002s : 0.00% optimize.opt_a.auto_parallel : 0.000003s : 0.01% optimize.opt_a.parallel : 0.000007s : 0.01% optimize.opt_a.merge_comm : 0.000002s : 0.00% optimize.opt_a.allreduce_fusion : 0.000001s : 0.00% optimize.opt_a.virtual_dataset : 0.000002s : 0.00% optimize.opt_a.get_grad_eliminate_ : 0.000002s : 0.00% optimize.opt_a.virtual_output : 0.000002s : 0.00% optimize.opt_a.merge_forward : 0.000003s : 0.01% optimize.opt_a.cell_reuse_recompute_pass : 0.000001s : 0.00% optimize.opt_a.cell_reuse_handle_not_recompute_node_pass : 0.000007s : 0.01% optimize.opt_a.meta_fg_expand : 0.000002s : 0.00% optimize.opt_a.after_resolve : 0.000002s : 0.00% optimize.opt_a.a_after_grad : 0.000002s : 0.00% optimize.opt_a.renormalize : 0.000397s : 0.80% optimize.opt_a.real_op_eliminate : 0.000005s : 0.01% optimize.opt_a.auto_monad_grad : 0.000003s : 0.01% optimize.opt_a.auto_monad_eliminator : 0.000007s : 0.02% optimize.opt_a.cse : 0.000015s : 0.03% optimize.opt_a.a_3 : 0.000014s : 0.03% optimize.py_interpret_to_execute_after_opt_a : 0.000006s : 0.01% optimize.slice_cell_reuse_recomputed_activation : 0.000002s : 0.00% optimize.rewriter_after_opt_a : 0.000233s : 0.47% optimize.convert_after_rewriter : 0.000005s : 0.01% optimize.order_py_execute_after_rewriter : 0.000004s : 0.01% optimize.opt_b.b_1 : 0.000036s : 0.07% optimize.opt_b.b_2 : 0.000003s : 0.01% optimize.opt_b.updatestate_depend_eliminate : 0.000002s : 0.00% optimize.opt_b.updatestate_assign_eliminate : 0.000002s : 0.00% optimize.opt_b.updatestate_loads_eliminate : 0.000002s : 0.00% optimize.opt_b.renormalize : 0.000000s : 0.00% optimize.opt_b.cse : 0.000009s : 0.02% optimize.cconv : 0.000015s : 0.03% optimize.opt_after_cconv.c_1 : 0.000005s : 0.01% optimize.opt_after_cconv.parameter_eliminate : 0.000001s : 0.00% optimize.opt_after_cconv.updatestate_depend_eliminate : 0.000002s : 0.00% optimize.opt_after_cconv.updatestate_assign_eliminate : 0.000001s : 0.00% optimize.opt_after_cconv.updatestate_loads_eliminate : 0.000001s : 0.00% optimize.opt_after_cconv.cse : 0.000006s : 0.01% optimize.opt_after_cconv.renormalize : 0.000000s : 0.00% optimize.remove_dup_value : 0.000009s : 0.02% optimize.tuple_transform.d_1 : 0.000012s : 0.02% optimize.tuple_transform.d_2 : 0.000006s : 0.01% optimize.tuple_transform.renormalize : 0.000000s : 0.00% optimize.add_cache_embedding : 0.000007s : 0.02% optimize.add_recomputation : 0.000019s : 0.04% optimize.cse_after_recomputation.cse : 0.000007s : 0.01% optimize.environ_conv : 0.000008s : 0.02% optimize.label_micro_interleaved_index : 0.000002s : 0.00% optimize.label_fine_grained_interleaved_index : 0.000002s : 0.00% optimize.assign_add_opt : 0.000001s : 0.00% optimize.slice_recompute_activation : 0.000002s : 0.00% optimize.micro_interleaved_order_control : 0.000001s : 0.00% optimize.full_micro_interleaved_order_control : 0.000002s : 0.00% optimize.comp_comm_scheduling : 0.000002s : 0.00% optimize.reorder_send_recv_between_fp_bp : 0.000002s : 0.00% optimize.comm_op_add_attrs : 0.000001s : 0.00% optimize.add_comm_op_reuse_tag : 0.000001s : 0.00% optimize.overlap_opt_shard_in_pipeline : 0.000001s : 0.00% optimize.grouped_pairwise_exchange_alltoall : 0.000001s : 0.00% optimize.overlap_recompute_and_grad_model_parallel : 0.000002s : 0.00% optimize.overlap_grad_matmul_and_grad_allreduce : 0.000001s : 0.00% optimize.split_matmul_comm_elemetwise : 0.000002s : 0.00% optimize.split_layernorm_comm : 0.000002s : 0.00% optimize.process_send_recv_for_ge : 0.000001s : 0.00% optimize.handle_group_info : 0.000001s : 0.00% auto_monad_reorder : 0.000012s : 0.03% get_jit_bprop_graph : 0.000000s : 0.00% eliminate_special_op_node : 0.000475s : 0.96% validate : 0.000030s : 0.06% distribtued_split : 0.000001s : 0.00% task_emit : 0.000001s : 0.00% execute : 0.000001s : 0.00% Time group info: ------[substitution.] 0.000453 15 90.97% : 0.000412s : 7: substitution.getattr_setattr_resolve 0.90% : 0.000004s : 2: substitution.graph_param_transform 7.78% : 0.000035s : 3: substitution.meta_unpack_prepare 0.25% : 0.000001s : 2: substitution.partial_unused_args_eliminate 0.10% : 0.000000s : 1: substitution.remove_not_recompute_node ------[renormalize.] 0.000393 2 72.08% : 0.000283s : 1: renormalize.infer 27.92% : 0.000110s : 1: renormalize.specialize ------[replace.] 0.000039 3 100.00% : 0.000039s : 3: replace.getattr_setattr_resolve ------[match.] 0.000241 3 100.00% : 0.000241s : 3: match.getattr_setattr_resolve ------[func_graph_cloner_run.] 0.000395 7 77.04% : 0.000304s : 5: func_graph_cloner_run.FuncGraphClonerGraph 22.96% : 0.000091s : 2: func_graph_cloner_run.FuncGraphSpecializer ------[meta_graph.] 0.000000 0 ------[manager.] 0.000000 0 ------[pynative] 0.000000 0 ------[others.] 0.000658 69 5.45% : 0.000036s : 26: opt.transform.opt_a 4.01% : 0.000026s : 23: opt.transform.opt_b 74.47% : 0.000490s : 2: opt.transform.opt_resolve 6.55% : 0.000043s : 1: opt.transforms.meta_unpack_prepare 5.24% : 0.000034s : 10: opt.transforms.opt_a 0.48% : 0.000003s : 1: opt.transforms.opt_after_cconv 0.31% : 0.000002s : 1: opt.transforms.opt_b 2.43% : 0.000016s : 2: opt.transforms.opt_trans_graph 1.07% : 0.000007s : 3: opt.transforms.special_op_eliminate TotalTime = 0.00949841, [20] [parse]: 0.00176332 [symbol_resolve]: 0.00026776, [1] [Cycle 1]: 0.00024874, [1] [resolve]: 0.00023703 [combine_like_graphs]: 6.00005e-07 [graph_reusing]: 2.66e-06 [meta_unpack_prepare]: 2.301e-05 [pre_cconv]: 2.70004e-07 [abstract_specialize]: 0.00347031 [pack_expand]: 5.3e-06 [auto_monad]: 1.126e-05 [inline]: 9.89996e-07 [pre_auto_parallel]: 6.18e-06 [pipeline_split]: 1.21e-06 [optimize]: 0.0033063, [35] [py_interpret_to_execute]: 5.7e-06 [rewriter_before_opt_a]: 1.562e-05 [opt_a]: 0.00275777, [1] [Cycle 1]: 0.00051124, [30] [expand_dump_flag]: 1.41e-06 [switch_simplify]: 7.08e-06 [a_1]: 1.483e-05 [recompute_prepare]: 1.95e-06 [updatestate_depend_eliminate]: 3.85e-06 [updatestate_assign_eliminate]: 2.32e-06 [updatestate_loads_eliminate]: 1.76e-06 [parameter_eliminate]: 1.57e-06 [a_2]: 2.579e-05 [accelerated_algorithm]: 2.18e-06 [pynative_shard]: 1.18e-06 [auto_parallel]: 2.92e-06 [parallel]: 5.62e-06 [merge_comm]: 2.09e-06 [allreduce_fusion]: 1.4e-06 [virtual_dataset]: 2.17999e-06 [get_grad_eliminate_]: 1.79e-06 [virtual_output]: 1.54e-06 [merge_forward]: 2.63e-06 [cell_reuse_recompute_pass]: 4.30002e-07 [cell_reuse_handle_not_recompute_node_pass]: 6.7e-06 [meta_fg_expand]: 1.79e-06 [after_resolve]: 1.64e-06 [a_after_grad]: 2.12e-06 [renormalize]: 0.000238 [real_op_eliminate]: 4.39e-06 [auto_monad_grad]: 2.67e-06 [auto_monad_eliminator]: 5.23e-06 [cse]: 1.112e-05 [a_3]: 1.358e-05 [py_interpret_to_execute_after_opt_a]: 6.03e-06 [slice_cell_reuse_recomputed_activation]: 1.46e-06 [rewriter_after_opt_a]: 0.000144 [convert_after_rewriter]: 3.9e-06 [order_py_execute_after_rewriter]: 2.47001e-06 [opt_b]: 9.316e-05, [1] [Cycle 1]: 8.819e-05, [7] [b_1]: 3.813e-05 [b_2]: 3.16e-06 [updatestate_depend_eliminate]: 2.53e-06 [updatestate_assign_eliminate]: 2.04e-06 [updatestate_loads_eliminate]: 1.72e-06 [renormalize]: 4.1e-07 [cse]: 8.34e-06 [cconv]: 1.099e-05 [opt_after_cconv]: 5.026e-05, [1] [Cycle 1]: 4.619e-05, [7] [c_1]: 4.49e-06 [parameter_eliminate]: 6.90001e-07 [updatestate_depend_eliminate]: 1.97e-06 [updatestate_assign_eliminate]: 1.59e-06 [updatestate_loads_eliminate]: 1.38e-06 [cse]: 5.73e-06 [renormalize]: 2.09999e-07 [remove_dup_value]: 5.58e-06 [tuple_transform]: 3.334e-05, [1] [Cycle 1]: 2.973e-05, [3] [d_1]: 1.096e-05 [d_2]: 5.08e-06 [renormalize]: 1.70003e-07 [add_cache_embedding]: 5.79e-06 [add_recomputation]: 1.649e-05 [cse_after_recomputation]: 1.521e-05, [1] [Cycle 1]: 1.111e-05, [1] [cse]: 6.3e-06 [environ_conv]: 7.15e-06 [label_micro_interleaved_index]: 1.29e-06 [label_fine_grained_interleaved_index]: 1.43e-06 [assign_add_opt]: 7.7e-07 [slice_recompute_activation]: 9.89996e-07 [micro_interleaved_order_control]: 9.70002e-07 [full_micro_interleaved_order_control]: 9.89996e-07 [comp_comm_scheduling]: 9.79999e-07 [reorder_send_recv_between_fp_bp]: 1.12e-06 [comm_op_add_attrs]: 4.60001e-07 [add_comm_op_reuse_tag]: 6.99998e-07 [overlap_opt_shard_in_pipeline]: 4.89999e-07 [grouped_pairwise_exchange_alltoall]: 5.4e-07 [overlap_recompute_and_grad_model_parallel]: 8.10003e-07 [overlap_grad_matmul_and_grad_allreduce]: 3.6e-07 [split_matmul_comm_elemetwise]: 1.32e-06 [split_layernorm_comm]: 1.07e-06 [process_send_recv_for_ge]: 4.50003e-07 [handle_group_info]: 4.50003e-07 [auto_monad_reorder]: 7.75e-06 [get_jit_bprop_graph]: 1.90004e-07 [eliminate_special_op_node]: 0.00045481 [validate]: 2.583e-05 [distribtued_split]: 7.2e-07 [task_emit]: 6.10002e-07 [execute]: 8.79998e-07 Sums parse : 0.001763s : 26.24% symbol_resolve.resolve : 0.000237s : 3.53% combine_like_graphs : 0.000001s : 0.01% graph_reusing : 0.000003s : 0.04% meta_unpack_prepare : 0.000023s : 0.34% pre_cconv : 0.000000s : 0.00% abstract_specialize : 0.003470s : 51.64% pack_expand : 0.000005s : 0.08% auto_monad : 0.000011s : 0.17% inline : 0.000001s : 0.01% pre_auto_parallel : 0.000006s : 0.09% pipeline_split : 0.000001s : 0.02% optimize.py_interpret_to_execute : 0.000006s : 0.08% optimize.rewriter_before_opt_a : 0.000016s : 0.23% optimize.opt_a.expand_dump_flag : 0.000001s : 0.02% optimize.opt_a.switch_simplify : 0.000007s : 0.11% optimize.opt_a.a_1 : 0.000015s : 0.22% optimize.opt_a.recompute_prepare : 0.000002s : 0.03% optimize.opt_a.updatestate_depend_eliminate : 0.000004s : 0.06% optimize.opt_a.updatestate_assign_eliminate : 0.000002s : 0.03% optimize.opt_a.updatestate_loads_eliminate : 0.000002s : 0.03% optimize.opt_a.parameter_eliminate : 0.000002s : 0.02% optimize.opt_a.a_2 : 0.000026s : 0.38% optimize.opt_a.accelerated_algorithm : 0.000002s : 0.03% optimize.opt_a.pynative_shard : 0.000001s : 0.02% optimize.opt_a.auto_parallel : 0.000003s : 0.04% optimize.opt_a.parallel : 0.000006s : 0.08% optimize.opt_a.merge_comm : 0.000002s : 0.03% optimize.opt_a.allreduce_fusion : 0.000001s : 0.02% optimize.opt_a.virtual_dataset : 0.000002s : 0.03% optimize.opt_a.get_grad_eliminate_ : 0.000002s : 0.03% optimize.opt_a.virtual_output : 0.000002s : 0.02% optimize.opt_a.merge_forward : 0.000003s : 0.04% optimize.opt_a.cell_reuse_recompute_pass : 0.000000s : 0.01% optimize.opt_a.cell_reuse_handle_not_recompute_node_pass : 0.000007s : 0.10% optimize.opt_a.meta_fg_expand : 0.000002s : 0.03% optimize.opt_a.after_resolve : 0.000002s : 0.02% optimize.opt_a.a_after_grad : 0.000002s : 0.03% optimize.opt_a.renormalize : 0.000238s : 3.54% optimize.opt_a.real_op_eliminate : 0.000004s : 0.07% optimize.opt_a.auto_monad_grad : 0.000003s : 0.04% optimize.opt_a.auto_monad_eliminator : 0.000005s : 0.08% optimize.opt_a.cse : 0.000011s : 0.17% optimize.opt_a.a_3 : 0.000014s : 0.20% optimize.py_interpret_to_execute_after_opt_a : 0.000006s : 0.09% optimize.slice_cell_reuse_recomputed_activation : 0.000001s : 0.02% optimize.rewriter_after_opt_a : 0.000144s : 2.14% optimize.convert_after_rewriter : 0.000004s : 0.06% optimize.order_py_execute_after_rewriter : 0.000002s : 0.04% optimize.opt_b.b_1 : 0.000038s : 0.57% optimize.opt_b.b_2 : 0.000003s : 0.05% optimize.opt_b.updatestate_depend_eliminate : 0.000003s : 0.04% optimize.opt_b.updatestate_assign_eliminate : 0.000002s : 0.03% optimize.opt_b.updatestate_loads_eliminate : 0.000002s : 0.03% optimize.opt_b.renormalize : 0.000000s : 0.01% optimize.opt_b.cse : 0.000008s : 0.12% optimize.cconv : 0.000011s : 0.16% optimize.opt_after_cconv.c_1 : 0.000004s : 0.07% optimize.opt_after_cconv.parameter_eliminate : 0.000001s : 0.01% optimize.opt_after_cconv.updatestate_depend_eliminate : 0.000002s : 0.03% optimize.opt_after_cconv.updatestate_assign_eliminate : 0.000002s : 0.02% optimize.opt_after_cconv.updatestate_loads_eliminate : 0.000001s : 0.02% optimize.opt_after_cconv.cse : 0.000006s : 0.09% optimize.opt_after_cconv.renormalize : 0.000000s : 0.00% optimize.remove_dup_value : 0.000006s : 0.08% optimize.tuple_transform.d_1 : 0.000011s : 0.16% optimize.tuple_transform.d_2 : 0.000005s : 0.08% optimize.tuple_transform.renormalize : 0.000000s : 0.00% optimize.add_cache_embedding : 0.000006s : 0.09% optimize.add_recomputation : 0.000016s : 0.25% optimize.cse_after_recomputation.cse : 0.000006s : 0.09% optimize.environ_conv : 0.000007s : 0.11% optimize.label_micro_interleaved_index : 0.000001s : 0.02% optimize.label_fine_grained_interleaved_index : 0.000001s : 0.02% optimize.assign_add_opt : 0.000001s : 0.01% optimize.slice_recompute_activation : 0.000001s : 0.01% optimize.micro_interleaved_order_control : 0.000001s : 0.01% optimize.full_micro_interleaved_order_control : 0.000001s : 0.01% optimize.comp_comm_scheduling : 0.000001s : 0.01% optimize.reorder_send_recv_between_fp_bp : 0.000001s : 0.02% optimize.comm_op_add_attrs : 0.000000s : 0.01% optimize.add_comm_op_reuse_tag : 0.000001s : 0.01% optimize.overlap_opt_shard_in_pipeline : 0.000000s : 0.01% optimize.grouped_pairwise_exchange_alltoall : 0.000001s : 0.01% optimize.overlap_recompute_and_grad_model_parallel : 0.000001s : 0.01% optimize.overlap_grad_matmul_and_grad_allreduce : 0.000000s : 0.01% optimize.split_matmul_comm_elemetwise : 0.000001s : 0.02% optimize.split_layernorm_comm : 0.000001s : 0.02% optimize.process_send_recv_for_ge : 0.000000s : 0.01% optimize.handle_group_info : 0.000000s : 0.01% auto_monad_reorder : 0.000008s : 0.12% get_jit_bprop_graph : 0.000000s : 0.00% eliminate_special_op_node : 0.000455s : 6.77% validate : 0.000026s : 0.38% distribtued_split : 0.000001s : 0.01% task_emit : 0.000001s : 0.01% execute : 0.000001s : 0.01% Time group info: ------[substitution.] 0.000198 12 95.92% : 0.000190s : 4: substitution.getattr_setattr_resolve 1.57% : 0.000003s : 2: substitution.graph_param_transform 2.03% : 0.000004s : 3: substitution.meta_unpack_prepare 0.28% : 0.000001s : 2: substitution.partial_unused_args_eliminate 0.19% : 0.000000s : 1: substitution.remove_not_recompute_node ------[renormalize.] 0.000235 2 62.54% : 0.000147s : 1: renormalize.infer 37.46% : 0.000088s : 1: renormalize.specialize ------[replace.] 0.000017 1 100.00% : 0.000017s : 1: replace.getattr_setattr_resolve ------[match.] 0.000077 1 100.00% : 0.000077s : 1: match.getattr_setattr_resolve ------[func_graph_cloner_run.] 0.000062 3 10.54% : 0.000006s : 1: func_graph_cloner_run.FuncGraphClonerGraph 89.46% : 0.000055s : 2: func_graph_cloner_run.FuncGraphSpecializer ------[meta_graph.] 0.000000 0 ------[manager.] 0.000000 0 ------[pynative] 0.000000 0 ------[others.] 0.000364 69 9.38% : 0.000034s : 26: opt.transform.opt_a 7.67% : 0.000028s : 23: opt.transform.opt_b 64.19% : 0.000233s : 2: opt.transform.opt_resolve 2.89% : 0.000011s : 1: opt.transforms.meta_unpack_prepare 8.54% : 0.000031s : 10: opt.transforms.opt_a 0.89% : 0.000003s : 1: opt.transforms.opt_after_cconv 0.57% : 0.000002s : 1: opt.transforms.opt_b 3.91% : 0.000014s : 2: opt.transforms.opt_trans_graph 1.96% : 0.000007s : 3: opt.transforms.special_op_eliminate . ============================== 1 passed in 20.71s ============================== [TRACE] GE(5051,python3.7):2024-01-11-05:50:13.947.430 [status:INIT] [ge_api.cc:463]5051 ~Session:Start to destruct session. [TRACE] GE(5051,python3.7):2024-01-11-05:50:13.947.479 [status:RUNNING] [ge_api.cc:475]5051 ~Session:Session id is 0 [TRACE] GE(5051,python3.7):2024-01-11-05:50:13.947.491 [status:RUNNING] [ge_api.cc:476]5051 ~Session:Destroying session [TRACE] GE(5051,python3.7):2024-01-11-05:50:13.948.440 [status:STOP] [ge_api.cc:491]5051 ~Session:Session Destructor finished [TRACE] GE(5051,python3.7):2024-01-11-05:50:13.948.470 [status:INIT] [ge_api.cc:301]5051 GEFinalize:GEFinalize start [INFO] GE(5051,python3.7):2024-01-11-05:50:13.948.561 [execution_runtime.cc:80][EVENT]5051 FinalizeExecutionRuntime:Execution runtime finalize begin. [INFO] GE(5051,python3.7):2024-01-11-05:50:13.948.580 [execution_runtime.cc:92][EVENT]5051 FinalizeExecutionRuntime:Execution runtime finalized. [TRACE] GE(5051,python3.7):2024-01-11-05:50:13.948.612 [status:RUNNING] [ge_api.cc:313]5051 GEFinalize:Finalizing environment [INFO] TUNE(5051,python3.7):2024-01-11-05:50:14.328.370 [cann_kb_pyfunc_mgr.cpp:127][CANNKB][Tid:5051]"CannKbPyfuncMgr: enter PyObjectDeinit function, reference_[1]" [INFO] TUNE(5051,python3.7):2024-01-11-05:50:14.328.434 [cann_kb_pyfunc_mgr.cpp:138][CANNKB][Tid:5051]"CannKbPyfuncMgr: PyObjectDeinit function end successfully!" [INFO] GE(5051,python3.7):2024-01-11-05:50:14.330.056 [gelib.cc:324][EVENT]5051 SystemFinalize:Online infer finalize GELib success. [TRACE] GE(5051,python3.7):2024-01-11-05:50:14.721.561 [status:STOP] [ge_api.cc:341]5051 GEFinalize:GEFinalize finished [INFO] TDT(5051,python3.7):2024-01-11-05:50:14.751.706 [process_mode_manager.cpp:184][Close][tid:5051] [TsdClient] Close [deviceId=0][sessionId=1] hccp and computer enter [INFO] TDT(5051,python3.7):2024-01-11-05:50:14.751.746 [version_verify.cpp:112][SpecialFeatureCheck][tid:5051] VersionVerify: previous type[7], supported [INFO] TDT(5051,python3.7):2024-01-11-05:50:14.751.793 [process_mode_manager.cpp:192][Close][tid:5051] [TsdClient][deviceId=0] [sessionId=1] wait hccp and computer process close respond [INFO] TDT(5051,python3.7):2024-01-11-05:50:14.782.753 [process_mode_manager.cpp:197][Close][tid:5051] [TsdClient][logicDeviceId_=0]has recv close hccp and computer process respond [INFO] TDT(5051,python3.7):2024-01-11-05:50:14.782.769 [stub_process_mode_nowin.cpp:151][CloseInHost][tid:5051] enter into CloseInHost deviceid[0] [INFO] TDT(5051,python3.7):2024-01-11-05:50:14.782.780 [stub_process_mode_nowin.cpp:154][CloseInHost][tid:5051] host cpu not support [INFO] TDT(5051,python3.7):2024-01-11-05:50:14.782.822 [process_mode_manager.cpp:208][Close][tid:5051] [TsdClient][deviceId=0] [sessionId=1] close hccp and computer process success [INFO] ATRACE(5051,python3.7):2024-01-11-05:50:14.782.835 [atrace_api.c:93](tid:5051) AtraceDestroy start [INFO] ATRACE(5051,python3.7):2024-01-11-05:50:14.782.853 [atrace_api.c:95](tid:5051) AtraceDestroy end [INFO] PROFILING(5051,python3.7):2024-01-11-05:50:14.782.876 [msprofiler_impl.cpp:156] >>> (tid:5051) ProfNotifySetDevice called, is open: 0, devId: 0 [INFO] RUNTIME(5051,python3.7):2024-01-11-05:50:16.873.611 [runtime.cc:1737] 5051 ~Runtime: deconstruct runtime.