General AICORE Troubleshooting

Procedure

  1. Collect the /var/log/ascend/ascend-dmi.log file, search for logs containing ret =, and record their process IDs and return codes.

  2. Collect plogs for CANN 8.0.RC2 and later versions and message logs for CANN earlier than 8.0.RC2. For details, see Plogs and Message Logs.

    Search for the log with a specified process ID and view the AML error information. case name indicates the specific operator test case, and error code indicates the corresponding error identifier.

  3. Rectify the fault based on the error code.

    Error Code

    Description

    100000

    Incorrect input parameter

    100001

    Invalid device ID

    100002

    AML postprocessing failure

    100003

    AML preprocessing failure

    100004

    Concurrent operations not supported

    100005

    No stress test case found

    100006

    Stress test interrupted abnormally

    100007

    ACL library or interface call failure

    100008

    Operator case preprocessing failure

    100009

    Operator case postprocessing failure

    100010

    Restoration interface call failure

    100011

    ASCEND_RT_VISIBLE_DEVICES not supported

    110000

    Case execution failure

    110001

    Case execution flow synchronization timeout

    120000

    Local resource error

    120001

    Mismatched driver version

    120002

    Chip type not supported

    120003

    Insufficient memory for case execution

    120004

    Case execution failure; no operator stress test case found

    130000

    Case execution failure

    130001

    Execution timeout for operator stress testing

    140000

    Global resource error

    190001

    Stress test environment setting failure

    190002

    Stress test environment restoration failure