How to Install nVidia drivers in Fedora

[dos2unix]

Enable the rpmfusion repo's. For fedora that looks like this.

dnf install https://mirrors.rpmfusion.org/free/fedora/rpmfusion-free-release-$(rpm -E %fedora).noarch.rpm https://mirrors.rpmfusion.org/nonfree/fedora/rpmfusion-nonfree-release-$(rpm -E %fedora).noarch.rpm

Install the following

dnf install akmod-nvidia kmod-nvidia libva-nvidia-driver nvidia-gpu-firmware nvidia-settings xorg-x11-drv-nvidia

reboot. You're done. This will auto-update the kernel modules for every new kernel release.

 

[dos2unix]

This works for most RPM based distro's.

( Redhat, SuSE, CentOS, AlmaLinux, RockyLinux, OracleLinux, AmazonLinux, and ScientificLinux )

The main difference, is that you also need to enable the epel repos.

Extra Packages for Enterprise Linux (EPEL)

Learn more about Fedora Linux, the Fedora Project & the Fedora Community.

docs.fedoraproject.org

 

[f33dm3bits]

Howto/NVIDIA - RPM Fusion

On my Rocky Linux 9 system I just use the Nvidia Cuda repo.

Streamlining NVIDIA Driver Deployment on RHEL 8 with Modularity Streams | NVIDIA Technical Blog

NVIDIA GPUs have become mainstream for accelerating a variety of workloads from machine learning, high-performance computing (HPC), content creation workflows, and data center applications.

developer.nvidia.com

Index of /compute/cuda/repos

 

[dos2unix]

On my Rocky Linux 9 system I just use the Nvidia Cuda repo.

The link you gave was for rhel8, but it looks like now they support rhel9 (as od Feb 10, 2025)

NVIDIA Open GPU Datacenter Drivers for RHEL9 Signed by Red Hat | NVIDIA Technical Blog

NVIDIA and Red Hat have partnered to bring continued improvements to the precompiled NVIDIA Driver introduced in 2020. Last month, NVIDIA announced that the open GPU driver modules will become the…

developer.nvidia.com

 

[dos2unix]

For Fedora 41, there are a few more options.

Go ahead and add the rpmfusion repo.

sudo dnf install https://mirrors.rpmfusion.org/free/fedora/rpmfusion-free-release-$(rpm -E %fedora).noarch.rpm https://mirrors.rpmfusion.org/nonfree/fedora/rpmfusion-nonfree-release-$(rpm -E %fedora).noarch.rpm -y

sudo dnf config-manager setopt fedora-cisco-openh264.enabled=1

Then we can install these packages. This is more or less everything. This includes the automated kernel updates.

dnf install -y akmod-nvidia kmod-nvidia libva-nvidia-driver nvidia-gpu-firmware nvidia-modprobe nvidia-settings nvidia-xconfig xorg-x11-drv-nvidia xorg-x11-drv-nvidia-cuda xorg-x11-drv-nvidia-kmodsrc Cg nv-codec-headers gwe nvidia-texture-tools

If you go this route, the kernel modules get updated every time the kernel gets updated. If you use the nvidia drivers on the nvidia website (which are the same versions) then you'll manually have to to re-install every time you get a kernel update.

 

[f33dm3bits]

The link you gave was for rhel8, but it looks like now they support rhel9 (as od Feb 10, 2025)

Yeah it was, I had been using the RHEL9 one for a while, if you look at the dates in the repo you can see it's been active for a while.

Index of /compute/cuda/preview/repos/rhel9/x86_64

I recently upgraded that system that was using that repo from CentOS Stream 9 to CentOS Stream 10, so I am currently just using the run file with dkms to use the Nvidia drivers which works fine as well.