Verify the CUDA runfile with md5sum cuda_10.0.130_410.48_n (optional).

Cd to the directory where you have the CUDA runfile downloaded, run cd Downloads/.

After the above, run sudo update-initramfs -u.

To do so, run sudo cp /home/avin/nf /etc/modprobe.d Change avin with your user name.

Now we have to copy the blacklist file we created at our home directory to the /etc/modprobe.d directory.

Go to super user mode by running sudo -i.

We need to disable the X-server by running sudo service lightdm stop in this black screen terminal.

(Don’t worry, CTRL+ALT+ F7 is your savior to log back to the GUI) It will be a black screen.Įnter your user name and password to go to the terminal.

Now logout from the GUI with CTRL+ALT+ F1. Make sure to remove all Nvidia-CUDA installations you have in the machine using sudo apt-get purge nvidia-cuda* if you have previously tried to install CUDA. Important: select No when it asks to automatically update your X configuration file. You could choose Yes to install Nvidia 32-bit compatible libraries. However, registering the kernel module sources with DKMS is optional and you may select No here. Select Continue installation in the GUI and follow the steps. However, if it is, it’s better to disable the secure boot from boot options. If this is the case, the installation GUI will indicate this. Note: Sometimes, Nvidia drivers require Linux’s UEFI secure boot to be disabled. So you would want to go ahead and remove this driver by sudo nvidia-uninstall and reboot the machine, then follow the installation commands again to install the correct driver. Nvidia GUI will be opened once you run the final command.Īt this step, Ubuntu 18.04 users might get an installation failure error because, in this case, the machine is using some form of Nvidia driver. But at this point, we won’t bother doing that. Note that, if you run the final command without -no-x-check flag, it will throw you an error saying X-Server needs to be disabled before installing the drivers.

Uninstall nvidia drivers ubuntu cuda