解决Ubuntu系统更新后N卡掉驱动的问题
1 卸载已安装驱动
首先需要卸载所有和n卡有关的系统安装
1 2 3
| sudo apt --purge remove "*cuda*" "*cublas*" "*cufft*" "*cufile*" "*curand*" "*cusolver*" "*cusparse*" "*gds-tools*" "*npp*" "*nvjpeg*" "nsight*" "*nvvm*" sudo apt purge "*nvidia*" "libxnvctrl*" sudo apt autoremove --purge -V
|
2 停用相关程序
首先检查一下内核的情况:
1
| sudo lsmod | grep nvidia
|
可能会看到类似这样的内容:
1 2 3 4 5 6
| nvidia_uvm 2162688 0 nvidia_drm 131072 0 nvidia_modeset 1716224 3 nvidia_drm nvidia 11628544 11 nvidia_uvm,nvidia_modeset video 77824 1 nvidia_modeset ecc 45056 1 nvidia
|
此时可以具体查看一下那些程序仍然在使用N卡:
比如我这里就有一些和N卡持久化的进程在占用:
1 2 3 4 5 6 7 8 9 10 11 12
| COMMAND PID USER FD TYPE DEVICE SIZE/OFF NODE NAME nvidia-pe 113901 nvidia-persistenced 2u CHR 195,255 0t0 1209 /dev/nvidiactl nvidia-pe 113901 nvidia-persistenced 3u CHR 195,0 0t0 1210 /dev/nvidia0 nvidia-pe 113901 nvidia-persistenced 5u CHR 195,0 0t0 1210 /dev/nvidia0 nvidia-pe 113901 nvidia-persistenced 6u CHR 195,0 0t0 1210 /dev/nvidia0 nvidia-pe 113901 nvidia-persistenced 7u CHR 195,254 0t0 1214 /dev/nvidia-modeset nvidia-pe 113901 nvidia-persistenced 8u CHR 195,0 0t0 1210 /dev/nvidia0 nvidia-pe 113901 nvidia-persistenced 9u CHR 195,1 0t0 1211 /dev/nvidia1 nvidia-pe 113901 nvidia-persistenced 10u CHR 195,1 0t0 1211 /dev/nvidia1 nvidia-pe 113901 nvidia-persistenced 11u CHR 195,1 0t0 1211 /dev/nvidia1 nvidia-pe 113901 nvidia-persistenced 12u CHR 195,254 0t0 1214 /dev/nvidia-modeset nvidia-pe 113901 nvidia-persistenced 13u CHR 195,1 0t0 1211 /dev/nvidia1
|
根据不同情况,停掉相应的进程(比如用systemctl或者其它的方式),然后卸载内核:
1 2
| sudo modprobe -r nvidia_modeset sudo modprobe -r nvidia
|
然后就可以重新安装了