That’s a warning and not an error - the kernel module exists. Did you perhaps use the nvidia website script before trying to install the drivers from RPMFusion?
The title suggests you’re trying to run on the discrete GPU exclusively instead of the integrated. Sometimes there is a BIOS setting for this and that’s the approach I would recommend. Otherwise, I know there’s a method @computersavvy has shared a number of times here for doing it with an Xorg.conf for X11. I’m otherwise not sure how to do this with Wayland if you don’t have a BIOS option.
By default, Optimus lets you run on the integrated GPU for regular workloads and use the Discrete for things like gaming. On Gnome, when you launch something, right click and you should see an option for “Run on Discrete Graphics” and then you’ll use the nvidia card for that. You can also set DRI_PRIME=1 in the Desktop file or launching things from terminal.
Probably yes, I’m a mess I don’t remember exactly what I did when after installing Fedora, only now I realized that the GPU is not active, in particular I needed it to be able to use a second monitor
I think the best practice would be to uninstall the current driver and dependencies and reinstall them with RPMFusion. So give me a moment and I’ll provide some commands for you.
Note:
Do you still have the .run file from the previous install ?
To completely remove the files installed with that .run file the command would be something like sudo /path/to/NVIDIA-Linux-x86-x64_64-555.58.run --uninstall.
Once that completes then install the drivers from rpmfusion.
Verify that you have the rpmfusion-nonfree-nvidia-driver repo enabled with the command dnf repolist.
If it is not enabled then it can be enabled by opening the gnome software app and using the ‘hamburger’ menu in the upper right corner then selecting the repos option and enable that repo.
install the drivers with sudo dnf install akmod-nvidia xorg-x11-drv-nvidia-cuda
Wait at least 5 minutes then reboot and the newer drivers from rpmfusion should now be active.
Note that the drivers installed with the nvidia .run file are compiled with dkms, so using akmods will not work to update the drivers and that is why the error in the OP occurred.
Well, everything seems ok, the second monitor is now up and running.
In the system menu I find the following situation:
Citazione #Rapporto sui dettagli del sistema
Modello hardware:** ASUSTeK COMPUTER INC. ASUS TUF Dash F15 FX517ZM_FX517ZM
Memoria: 32,0 GiB
Processore: 12th Gen Intel® Core™ i7-12650H × 16
Scheda grafica: Intel® Graphics (ADL GT2)
Scheda grafica 1: NVIDIA GeForce RTX™ 3060 Laptop GPU
Capacità del disco: 1,5 TB
Informazioni sul software:
Versione del firmware: FX517ZM.318
Nome del sistema operativo Fedora Linux 40 (Workstation Edition)
Build del sistema operativo (null)
Tipo di sistema operativo 64-bit
Versione di GNOME: 46
Gestore grafico: Wayland
Versione del kernel: Linux 6.9.10-200.fc40.x86_64
It’s correct?
Because before the graphics manager was x11, now it is Wayland.
Wayland is definitely preferred over X11 unless you have some specific legacy or 3rd party need for it. If you’re not sure, I would stick with wayland.
Here is a wild and off topic question. What’s the viability of using XWayland to launch some of these legacy/3rd party apps. If you know we can split the conversation here and continue anew.
The short answer is very good for most things. The long answer is that there are numerous topics about this here. RHEL10 is planning on dropping X11 entirely for DEs and falling back to XWayland. There are a bunch of caveats to that, many of which are DE and specialty specific. For most users, this is the way to go.