npu-smi info正常, acl.init()正常，但是acl.rt.set_device(0)返回507033

一、问题现象（附报错日志上下文）：
[EVENT] PROFILING(415452,python3):2024-08-23-10:27:44.983.225 [msprof_callback_impl.cpp:324] >>> (tid:415452) Started to register profiling ctrl callback.
[EVENT] PROFILING(415452,python3):2024-08-23-10:27:44.983.346 [msprof_callback_impl.cpp:331] >>> (tid:415452) Started to register profiling hash id callback.
[EVENT] PROFILING(415452,python3):2024-08-23-10:27:44.983.353 [prof_atls_plugin.cpp:83] >>> (tid:415452) RegisterProfileCallback, callback type is 7
[EVENT] PROFILING(415452,python3):2024-08-23-10:27:44.983.357 [msprof_callback_impl.cpp:338] >>> (tid:415452) Started to register profiling enable host freq callback.
[EVENT] PROFILING(415452,python3):2024-08-23-10:27:44.983.361 [prof_atls_plugin.cpp:83] >>> (tid:415452) RegisterProfileCallback, callback type is 8
[EVENT] PROFILING(415452,python3):2024-08-23-10:27:45.017.795 [prof_atls_plugin.cpp:160] >>> (tid:415452) Module[6] register callback of ctrl handle.
[EVENT] ASCENDCL(415452,python3):2024-08-23-10:27:45.021.948 [acl_resource_manager.cpp:90]415452 GetRuntimeV2Env: runtime v2 flag : model flag = 1, singleOp flag = 1
[EVENT] PROFILING(415452,python3):2024-08-23-10:27:45.046.349 [prof_atls_plugin.cpp:160] >>> (tid:415452) Module[48] register callback of ctrl handle.
[EVENT] PROFILING(415452,python3):2024-08-23-10:27:45.046.376 [prof_atls_plugin.cpp:160] >>> (tid:415452) Module[45] register callback of ctrl handle.
[EVENT] PROFILING(415452,python3):2024-08-23-10:27:45.538.905 [msprof_callback_impl.cpp:79] >>> (tid:415452) MsprofCtrlCallback called, type: 255
[EVENT] PROFILING(415452,python3):2024-08-23-10:27:45.539.071 [ai_drv_dev_api.cpp:333] >>> (tid:415452) Succeeded to DrvGetApiVersion version: 0x72313
[EVENT] RUNTIME(415452,python3):2024-08-23-10:27:45.539.151 [runtime.cc:4393] 415452 GetVisibleDevices: ASCEND_RT_VISIBLE_DEVICES param was not set
[EVENT] PROFILING(415452,python3):2024-08-23-10:27:45.540.281 [prof_atls_plugin.cpp:160] >>> (tid:415452) Module[7] register callback of ctrl handle.
[EVENT] TDT(415452,python3):2024-08-23-10:27:45.540.695 [process_mode_manager.cpp:107][OpenProcess][tid:415452] [ProcessModeManager] enter into open process deviceId[0] rankSize[0]
[EVENT] TDT(415452,python3):2024-08-23-10:27:45.541.045 [process_mode_manager.cpp:377][InitTsdClient][tid:415452] [TsdClient] deviceId[0] begin to init hdc client
[EVENT] TDT(415452,python3):2024-08-23-10:27:45.541.163 [version_verify.cpp:34][SetVersionInfo][tid:415452] VersionVerify: send client version to server
[EVENT] TDT(415452,python3):2024-08-23-10:27:45.541.174 [version_verify.cpp:51][SetVersionInfo][tid:415452] send feature_info:{msg_type:35, features:{check before send aicpu package,}}
[EVENT] TDT(415452,python3):2024-08-23-10:27:45.541.181 [version_verify.cpp:51][SetVersionInfo][tid:415452] send feature_info:{msg_type:37, features:{check before send open qs message,}}
[EVENT] TDT(415452,python3):2024-08-23-10:27:45.541.821 [version_verify.cpp:67][PeerVersionCheck][tid:415452] VersionVerify: Check client version info, server[1230], client[1230]
[EVENT] TDT(415452,python3):2024-08-23-10:27:45.541.828 [version_verify.cpp:88][ParseVersionInfo][tid:415452] VersionVerify: pass client version info success
[EVENT] TDT(415452,python3):2024-08-23-10:27:45.541.831 [hdc_client.cpp:276][CheckHdcConnection][tid:415452] Service[2] create hdc success
[EVENT] TDT(415452,python3):2024-08-23-10:27:45.541.838 [version_verify.cpp:121][SpecialFeatureCheck][tid:415452] VersionVerify: new type[35], supported
[EVENT] TDT(415452,python3):2024-08-23-10:27:45.541.880 [process_mode_manager.cpp:748][GetDeviceCheckCode][tid:415452] [TsdClient][deviceId=0] [sessionId=1]wait package info respond
[EVENT] TDT(415452,python3):2024-08-23-10:27:45.542.043 [process_mode_manager.cpp:377][InitTsdClient][tid:415452] [TsdClient] deviceId[0] begin to init hdc client
[EVENT] TDT(415452,python3):2024-08-23-10:27:45.542.109 [version_verify.cpp:34][SetVersionInfo][tid:415452] VersionVerify: send client version to server
[EVENT] TDT(415452,python3):2024-08-23-10:27:45.542.116 [version_verify.cpp:51][SetVersionInfo][tid:415452] send feature_info:{msg_type:35, features:{check before send aicpu package,}}
[EVENT] TDT(415452,python3):2024-08-23-10:27:45.542.125 [version_verify.cpp:51][SetVersionInfo][tid:415452] send feature_info:{msg_type:37, features:{check before send open qs message,}}
[EVENT] TDT(415452,python3):2024-08-23-10:27:45.542.624 [version_verify.cpp:67][PeerVersionCheck][tid:415452] VersionVerify: Check client version info, server[1230], client[1230]
[EVENT] TDT(415452,python3):2024-08-23-10:27:45.542.629 [version_verify.cpp:88][ParseVersionInfo][tid:415452] VersionVerify: pass client version info success
[EVENT] TDT(415452,python3):2024-08-23-10:27:45.542.633 [hdc_client.cpp:276][CheckHdcConnection][tid:415452] Service[2] create hdc success
[EVENT] TDT(415452,python3):2024-08-23-10:27:45.542.638 [process_mode_manager.cpp:424][ConstructOpenMsg][tid:415452] [TsdClient] tsd get process sign successfully, procpid[415452] signSize[48]
[EVENT] TDT(415452,python3):2024-08-23-10:27:45.542.643 [version_verify.cpp:113][SpecialFeatureCheck][tid:415452] VersionVerify: previous type[6], supported
[EVENT] TDT(415452,python3):2024-08-23-10:27:45.542.653 [process_mode_manager.cpp:124][OpenProcess][tid:415452] [ProcessModeManager] deviceId[0] sessionId[1] rankSize[0], wait sub process start respond
[EVENT] TDT(415452,python3):2024-08-23-10:27:45.899.115 [stub_process_mode_nowin.cpp:61][ProcessQueueForMdc][tid:415452] [TsdClient] it is unnecessary of current mode[0] chiptype[2] to grant queue auth to aicpusd
[EVENT] TDT(415452,python3):2024-08-23-10:27:45.899.143 [stub_process_mode_nowin.cpp:99][OpenInHost][tid:415452] enter into OpenInHost deviceid[0]
[EVENT] TDT(415452,python3):2024-08-23-10:27:45.899.149 [stub_process_mode_nowin.cpp:103][OpenInHost][tid:415452] host cpu not support
[EVENT] TDT(415452,python3):2024-08-23-10:27:45.899.152 [process_mode_manager.cpp:154][OpenProcess][tid:415452] [TsdClient][deviceId=0] [sessionId=1] start hccp and computer process success
[ERROR] DRV(415452,python3):2024-08-23-10:27:45.899.259 [ascend][curpid: 415452, 415452][drv][tsdrv][drvDeviceOpen 361]open devdrv_device(0) failed, fd(-1), errno(14), error(Bad address).
[ERROR] DRV(415452,python3):2024-08-23-10:27:45.899.268 [ascend][curpid: 415452, 415452][drv][tsdrv][share_log_read 569]invalid para, vdevid(0), devid(0), fid(0) invalid mode(1) or ctx is null.
[ERROR] RUNTIME(415452,python3):2024-08-23-10:27:45.899.329 [npu_driver.cc:2266]415452 DeviceOpen:[drv api] drvDeviceOpen failed: deviceId=0, drvRetCode=4!
[ERROR] RUNTIME(415452,python3):2024-08-23-10:27:45.899.337 [device.cc:268]415452 InitRawDriver:report error module_type=0, module_name=EE9999
[ERROR] RUNTIME(415452,python3):2024-08-23-10:27:45.899.341 [device.cc:268]415452 InitRawDriver:Failed to open device, retCode=0x7020014, deviceId=0.
[ERROR] RUNTIME(415452,python3):2024-08-23-10:27:45.899.360 [device.cc:312]415452 Init:report error module_type=0, module_name=EE9999
[ERROR] RUNTIME(415452,python3):2024-08-23-10:27:45.899.365 [device.cc:312]415452 Init:Failed to init RawDriver, device_id=0, retCode=0x7020014
[ERROR] RUNTIME(415452,python3):2024-08-23-10:27:45.899.376 [runtime.cc:3019]415452 DeviceRetain:Failed to init device.
[EVENT] TDT(415452,python3):2024-08-23-10:27:45.899.383 [process_mode_manager.cpp:182][Close][tid:415452] [TsdClient] Close [deviceId=0][sessionId=1] hccp and computer enter
[EVENT] TDT(415452,python3):2024-08-23-10:27:45.899.389 [version_verify.cpp:113][SpecialFeatureCheck][tid:415452] VersionVerify: previous type[7], supported
[EVENT] TDT(415452,python3):2024-08-23-10:27:45.899.411 [process_mode_manager.cpp:192][Close][tid:415452] [TsdClient][deviceId=0] [sessionId=1] wait hccp and computer process close respond
[EVENT] TDT(415452,python3):2024-08-23-10:27:45.913.403 [process_mode_manager.cpp:197][Close][tid:415452] [TsdClient][logicDeviceId_=0]has recv close hccp and computer process respond
[EVENT] TDT(415452,python3):2024-08-23-10:27:45.913.411 [stub_process_mode_nowin.cpp:149][CloseInHost][tid:415452] enter into CloseInHost deviceid[0]
[EVENT] TDT(415452,python3):2024-08-23-10:27:45.913.415 [stub_process_mode_nowin.cpp:152][CloseInHost][tid:415452] host cpu not support
[EVENT] TDT(415452,python3):2024-08-23-10:27:45.913.492 [process_mode_manager.cpp:210][Close][tid:415452] [TsdClient][deviceId=0] [sessionId=1] close hccp and computer process success
[ERROR] RUNTIME(415452,python3):2024-08-23-10:27:45.913.764 [runtime.cc:3041]415452 DeviceRetain:report error module_type=0, module_name=EE9999
[ERROR] RUNTIME(415452,python3):2024-08-23-10:27:45.913.770 [runtime.cc:3041]415452 DeviceRetain:Check param failed, dev can not be NULL!
[ERROR] RUNTIME(415452,python3):2024-08-23-10:27:45.913.783 [runtime.cc:2806]415452 PrimaryContextRetain:report error module_type=0, module_name=EE9999
[ERROR] RUNTIME(415452,python3):2024-08-23-10:27:45.913.788 [runtime.cc:2806]415452 PrimaryContextRetain:Check param failed, dev can not be NULL!
[ERROR] RUNTIME(415452,python3):2024-08-23-10:27:45.913.801 [runtime.cc:2833]415452 PrimaryContextRetain:report error module_type=0, module_name=EE9999
[ERROR] RUNTIME(415452,python3):2024-08-23-10:27:45.913.806 [runtime.cc:2833]415452 PrimaryContextRetain:Check param failed, ctx can not be NULL!
[ERROR] RUNTIME(415452,python3):2024-08-23-10:27:45.913.816 [api_impl.cc:2070]415452 NewDevice:report error module_type=0, module_name=EE9999
[ERROR] RUNTIME(415452,python3):2024-08-23-10:27:45.913.820 [api_impl.cc:2070]415452 NewDevice:Check param failed, context can not be null.
[ERROR] RUNTIME(415452,python3):2024-08-23-10:27:45.913.832 [api_impl.cc:2093]415452 SetDevice:report error module_type=0, module_name=EE9999
[ERROR] RUNTIME(415452,python3):2024-08-23-10:27:45.913.836 [api_impl.cc:2093]415452 SetDevice:New device failed, retCode=0x7010006
[ERROR] RUNTIME(415452,python3):2024-08-23-10:27:45.913.846 [logger.cc:836]415452 SetDevice:Set device failed, device_id=0, deviceMode=0.
[ERROR] RUNTIME(415452,python3):2024-08-23-10:27:45.913.871 [api_c.cc:1763]415452 rtSetDevice:ErrCode=507033, desc=[device retain error], InnerCode=0x7010006
[ERROR] RUNTIME(415452,python3):2024-08-23-10:27:45.913.877 [error_message_manage.cc:50]415452 FuncErrorReason:report error module_type=3, module_name=EE8888
[ERROR] RUNTIME(415452,python3):2024-08-23-10:27:45.913.886 [error_message_manage.cc:50]415452 FuncErrorReason:rtSetDevice execute failed, reason=[device retain error]
[ERROR] ASCENDCL(415452,python3):2024-08-23-10:27:45.913.905 [device.cpp:151]415452 aclrtSetDevice: open device 0 failed, runtime result = 507033.
set_device:507033
[EVENT] RUNTIME(415452,python3):2024-08-23-10:27:45.984.329 [runtime.cc:1699] 415452 ~Runtime: deconstruct runtime.

二、软件版本:
安装的固件，驱动，CANN版本如下：
Ascend-hdk-310p-npu-driver_23.0.1_linux-aarch64.run
Ascend-hdk-310p-npu-firmware_7.1.0.4.220.run
Ascend-cann-toolkit_7.0.0_linux-aarch64.run

--操作系统版本
Linux tdhnodesn13260 4.19.90-17.ky10.aarch64 #1 SMP Sun Jun 28 14:27:40 CST 2020 aarch64 aarch64 aarch64 GNU/Linux

npu-smi info正常
![输入图片说明](https://foruda.gitee.com/images/1724380833884747387/bd972b2d_11762594.png "2024-08-23 10-39-36 的屏幕截图.png")

import acl
acl.init()正常返回0
但是物理机上执行acl.rt.set_device(0)返回507033错误码．

Ascend/samples

内容风险标识

评论 (4)

Ascend/samples .gitee-modal { width: 500px !important; }

内容风险标识