1. 官网下载4090：驱动程序
2. 关闭图形界面
sudo systemctl set-default multi-user.target
sudo reboot对应的打开图形界面命令为：
sudo systemctl set-default graphical.target
sudo reboot3. tty登录之后
安装新版驱动：
# 切换到驱动目录
cd /home/user/download/
# 权限
sudo chmod a+x NVIDIA-Linux-x86_64-520.56.06.run
# 安装
sudo ./NVIDIA-Linux-x86_64-520.56.06.run如果报错，信息如下：
ERROR: An NVIDIA kernel module 'nvidia-drm' appears to already be loaded in your kernel. This may be because it is in use (for example, by an X server, a CUDA program, or the NVIDIA
Persistence Daemon), but this may also happen if your kernel was configured without support for module unloading. Please be sure to exit any programs that may be using the GPU(s)
before attempting to upgrade your driver. If no GPU-based programs are running, you know that your kernel supports module unloading, and you still receive this message, then an
error may have occurred that has corrupted an NVIDIA kernel module's usage count, for which the simplest remedy is to reboot your computer.
则需要卸载老版 nvidia 驱动：
（1）如果有老版驱动安装包，则运行如下命令：
./NVIDIA-Linux-x86-310.19.run --uninstall
reboot now（2）没有老版驱动安装包，可以使用如下命令删除 niidia 驱动，并使用 autoremove 清除所有不必要的相关套件：
apt-get purge nvidia*
apt-get autoremove
reboot或者关闭所有装置，并停止载入 nvidia 驱动：
systemctl isolate multi-user.target
modprobe -r nvidia-drm成功卸载掉老款驱动后，重新运行下载的驱动安装包，期间可能提示你安装 gcc、make 命令后再安装。
4. 打开图形界面并重启
sudo systemctl set-default graphical.target
sudo reboot安装后第一次开机时间较长，耐心等待
经验证，更换4090显卡后，基于cuda10.2编译的pytorch已不受支持；
更换为cuda11.6编译的pytorch可正常使用。
conda install pytorch torchvision torchaudio cudatoolkit=11.6 -c pytorch -c conda-forge5. 安装 cuda + cudnn，参考: