I didn’t want to install all cards, instead I wanted to have only the driver for my card. Next comes the step of identifying the right driver for my specific card, and installing it. Regenerate the kernel initramfs sudo update-initramfs -u Step 2: Install the right driver for the card. sudo nano nfĪnd comment the line with the nvidiafb #blacklist nvidiafb While in the modprobe.d directory it is beneficial to edit the framebuffer blacklisting of nvidiafb. cd /etc/modprobe.d/Īdd the following two lines to the file blacklist nouveau The procedure is to blacklist them so they don’t get loaded with modprobe. I am not sure if this step is really necessary for just having the system up and running but as I want to use CUDA in my Deep Learning Adventures and the Nouveau drivers don’t support CUDA I make this preparation. Next, I prepare the system by blacklisting the Nouveau drivers (the open source reverse engineered drivers for NVIDIA hardware). sudo add-apt-repository ppa:graphics-drivers/ppa I wanted the proprietary drivers with bleeding edge repository available to have the latest drivers from NVIDIA, I have never had problems with an unstable system so far (fingers crossed). Step 1: Make the NVIDIA repository available Anyway, hope this helps someone else that’s in need of minimal instructions. There are many guides available but I want to keep my dedicated information at an arm-lengths distance. Now I have it available if I would make dumb mistakes again.
So I needed to fix the disks and make a reinstall of Ubuntu and dig up the information that I hadn’t saved about the steps and necessary libraries.
I got a disk corruption due to accidentally pressing the reset button, when I was searching in blind for another button on the computer chassis.
This post documents what I did to install the drivers for my new NVIDIA graphics hardware ASUS RTX 2080 Ti Dual 11GB that I have in my personal computation server that I have prepared for my new adventures into “Deep Learning Land”.
0 kommentar(er)
