Hello, guys! I’m worried about breaking my Fedora 42 setup while installing NVIDIA drivers on ASUS ROG Strix G513IC-EB73, need some advice.
I’m new to Fedora (just installed Workstation 42 on my ASUS ROG Strix G513IC-EB73 laptop) and running GNOME (ofc). The hardware is an AMD Ryzen CPU with integrated Radeon graphics and a discrete NVIDIA GeForce RTX 3050 GPU. The GPU is detected fine (lspci | grep -i nvidia shows it), but nvidia-smi isn’t available, it’s expected cuz I did not install the proprietary drivers yet.
I’d like to get the NVIDIA GPU working properly for offloading (gaming or CUDA tasks, like AI etc.) via PRIME, but I’m hesitant cuz I’ve read about potential issues like black screens, boot loops, or kernel conflicts on hybrid setups and many more…
Could this install possibly lead to unbootable states, (Secure Boot is off btw). What’s the worst scenario, and how recoverable is it?
I was about to run:
sudo dnf install kernel-devel kernel-headers gcc make dkms acpid libglvnd-glx libglvnd-opengl libglvnd-devel pkgconfig
sudo dnf install https://download1.rpmfusion.org/free/fedora/rpmfusion-free-release-$(rpm -E %fedora).noarch.rpm https://download1.rpmfusion.org/nonfree/fedora/rpmfusion-nonfree-release-$(rpm -E %fedora).noarch.rpm
sudo dnf makecache
sudo dnf install akmod-nvidia xorg-x11-drv-nvidia-cuda
If it goes wrong, is it possible to rollback safely? Also, once installed, how can I test if it’s working properly?
I’d be really grateful if someone has the time to answer some of my questions, (I’m a bit paranoid, already broke my shell once, when trying to develop an extension, lol, do not want to deal with grub, or other scary/tricky things, heh).
Thanks!