问题原因:系统自动更新驱动版本or库版本,导致不匹配
输出:
mzh@ubuntu-System-Product-Name:~$ nvidia-smi -V
Failed to initialize NVML: Driver/library version mismatch
NVML library version: 535.171mzh@ubuntu-System-Product-Name:~$ cat /proc/driver/nvidia/version
NVRM version: NVIDIA UNIX x86_64 Kernel Module 535.154.05 Thu Dec 28 15:37:48 UTC 2023
GCC version: gcc version 12.3.0 (Ubuntu 12.3.0-1ubuntu1~22.04)
尝试:重启/sudo apt upgrade
结果:还是报错
解决方法:重新安装nvidia driver
sudo cp /etc/apt/apt.conf.d/50unattended-upgrades /etc/apt/apt.conf.d/50unattended-upgrades.bak
# make backup
sudo sed -i "/Unattended-Upgrade::Package-Blacklist {/,/}/ s/}/ \"nvidia-\";\n \"libnvidia-\";\n}/" /etc/apt/apt.conf.d/50unattended-upgrades