An error occurs after the nvida-smi command is run，“Unable to determine the device handle for GPU 0000:01:00.0: Not Found”.
This error has occurred twice, both times after the machine was restarted. The first error was fixed by reinstalling the cuda driver. There it is again today. What is the cause of this failure and is there a permanent solution?
Machine environment is :
Graphics card: 2 ✖ Geforce RTX 3090 24GB
OS: Ubuntu 22.04.1 LTS (Jammy Jellyfish)，kernel：5.15.0-53-generic
NVRM version: NVIDIA UNIX Open Kernel Module for x86_64 520.61.05 Release Build (dvs-builder@U16-I2-C02-14-2) Thu Sep 29 05:39:52 UTC 2022
bug report log:
nvidia-bug-report.log.gz (192.3 KB)