nvidia-smi突然报错

查看GPU占用情况遇到错误

吐槽一下，这个GPU真的是一波三折，太麻烦了。

nvidia-smi

NVIDIA-SMI has failed because it couldn't communicate with the NVIDIA driver. Make sure that the latest NVIDIA driver is installed and running.

原因

应该是系统的内核升级，新内核不支持cuda

解决，重新编译动态内核

root@z600:~# apt-get install dkms

root@z600:~# dkms install -m nvidia -v 430.40

430.40 是显卡驱动的版本号可以在/usr/src/中查看对应文件夹名称

root@z600:~# ls /usr/src

linux-headers-4.4.0-169 linux-headers-4.4.0-21-generic nvidia-430.40
linux-headers-4.4.0-169-generic linux-source-4.4.0
linux-headers-4.4.0-21 linux-source-4.4.0.tar.bz2