This works fine. Although the ‘cudo’ version of this may be more appilcable for you - sorry I have no experience of dual grahics.
The differences are that I don’t have a secondary graphics like you. My ‘zypper se -si nvidia’ entry the same as yours except that I have the above driver installed instead of ‘nvidia-driver-G06-kmp-default 580.95.05_k6.17.0_2-41.3’ and don’t have any 32-bit or the MicroOS versions.
@raijar So you have nomodeset in your boot options? This looks like it’s disabling the Intel GPU? Have you tried switching to the Xe diver instead of i915?
I’m really unsure about how to boot my computer because I did a lot of “work” to get it working.
I removed the nomodeset parameter and the computer still booted properly.
splash=silent quiet security=selinux selinux=1
But I still don’t dare to move on from this situation (=working system) because I’m afraid my computer will get corrupted.
Can I save startup files etc. so I can safely install a new driver and remove the old ones? Or can I save the system files somehow?
There are no open-source kernel modules for the latest kernel version you (or any of us at the moment) are using (6.17.5). The latest seem to be 580.95.05_k6.17.3_1-3.2.
This will not work until the kernel version and nvidia-open-driver-G06-signed-kmp-default version match
This is a very bad system and forcing the use of the “legacy” modules if the user wants to keep Tumbleweed up to date.
Please do not spread false informations. This topic was already explained countless times. The kernel modules are built on the users machine. Only if KABI is incompatible between kernels, a new package is needed.
The open kernel module is working properly with the latest kernel:
:~> uname -r
6.17.6-1-default
:~> zypper se -si nvidia-open-driver-G06-signed-kmp-default
Repository-Daten werden geladen...
Installierte Pakete werden gelesen...
S | Name | Type | Version | Arch | Repository
---+-------------------------------------------+-------+-------------------------+--------+-----------
i+ | nvidia-open-driver-G06-signed-kmp-default | Paket | 580.95.05_k6.17.3_1-3.2 | x86_64 | repo-oss
:~> nvidia-smi
Sun Nov 2 15:58:25 2025
+-----------------------------------------------------------------------------------------+
| NVIDIA-SMI 580.95.05 Driver Version: 580.95.05 CUDA Version: 13.0 |
+-----------------------------------------+------------------------+----------------------+
| GPU Name Persistence-M | Bus-Id Disp.A | Volatile Uncorr. ECC |
| Fan Temp Perf Pwr:Usage/Cap | Memory-Usage | GPU-Util Compute M. |
| | | MIG M. |
|=========================================+========================+======================|
| 0 NVIDIA GeForce RTX 3080 Ti On | 00000000:01:00.0 On | N/A |
| 0% 32C P8 26W / 350W | 591MiB / 12288MiB | 1% Default |
| | | N/A |
+-----------------------------------------+------------------------+----------------------+
+-----------------------------------------------------------------------------------------+
| Processes: |
| GPU GI CI PID Type Process name GPU Memory |
| ID ID Usage |
|=========================================================================================|
| 0 N/A N/A 1420 G /usr/bin/Xorg.bin 21MiB |
| 0 N/A N/A 1667 G /usr/bin/ksecretd 3MiB |
| 0 N/A N/A 1705 G /usr/bin/kwin_wayland 37MiB |
| 0 N/A N/A 1783 G /usr/bin/Xwayland 4MiB |
| 0 N/A N/A 1800 G /usr/bin/ksmserver 3MiB |
| 0 N/A N/A 1802 G /usr/bin/kded6 3MiB |
| 0 N/A N/A 1912 G /usr/bin/plasmashell 165MiB |
| 0 N/A N/A 1927 G /usr/bin/kaccess 3MiB |
| 0 N/A N/A 1928 G ...it-kde-authentication-agent-1 3MiB |
| 0 N/A N/A 1930 G ...ibexec/xdg-desktop-portal-kde 3MiB |
| 0 N/A N/A 2050 G /usr/libexec/DiscoverNotifier 3MiB |
| 0 N/A N/A 2161 G /usr/bin/konsole 3MiB |
| 0 N/A N/A 2223 G /opt/google/chrome/chrome 3MiB |
| 0 N/A N/A 2270 G ...rack-uuid=3190708988185955192 95MiB |
| 0 N/A N/A 2277 G /usr/bin/kwalletd6 3MiB |
Indeed, 6.17.6 seems to work with the current open driver package.
Have had so much grief with the updates, was under the assumption there is a need for constant parity with kernel and module version. Glad to see this is not the case.
I’m sorry I can’t explain the problem without offending someone.
The situation is that I bought a computer. Maybe this:
ASUS RNUC14SRKU910002I Barebone Intel Core Ultra 9 185H NVIDIA GeForce RTX 4070 Kit L6 EU Cord
I thought it would be easy to install opensuse tumbleweed, but it didn’t work. I had a lot of problems, even though I installed my first linux in 1994.
And greetings to pullasuti
And you were absolutely correct for the open driver which is built (and signed) on the build system and provided as binary, not as source compiled on the user system.
It works not because it is compiled on the end user system, but because the kernel symbols required by the compiled binary module did not change between minor kernel updates.
There need to be parity between the symbols exported by kernel and the symbols used (and required) by the binary module.
So, yes - a kernel module compiled for some kernel version may work with a different kernel version. No, it happens not because this kernel module is rebuilt on the user system, but because kernel changes did not affect this module. I am not sure whether OBS detects that the new kernel becomes incompatible with the existing binary and rebuilds affected modules automatically.
But that is not specific to NVIDIA driver in any way, it affects any other out-of-tree kernel module provided as a separate package. Changes between minor kernel versions are hopefully minor, so we can expect a module built for 6.17.3 to work with 6.17.6. It is very unlikely that this module will work with the 6.18 (while closed source driver package will work as long as it still compiles).
I’m sorry, but I still don’t understand the root cause of this problem.
There are 3 ports on the back of my computer:
HDMI Port 1
DisplayPort Ports 2
I only use the HDMI port because I only have one monitor and I don’t have a DP-HDMI cable.
So, for example, is the HDMI port hardwired to a specific graphics driver? And if I had a DP-HDMI cable, would I be able to select the driver I want (i915/nvidia) from the DisplayPort Ports?
@raijar very likely the DP port will be connected to the Intel device, get a DP-HDMI cable or adapter and test. You can then use that and it’s features and Nvidia for Prime Render Offload, which is what I use here…
Thanks for the info!
I was already looking for a DP-HDMI cable or adapter. But I live in such a small town that the store didn’t have one.
I’ll let you know when I get the cable.
I finally got a PD-HDMI cable and connected it between my computer and the monitor.
I’m so stupid that I don’t understand which graphics driver is currently in use but it works fine.
@raijar follow the card and the connections on the Monitor output;
DP-4 mapped: DP-3 so it’s running off the Nvidia GPU, so try different ports to see what maps to what. Ideally you want the monitor running off the Intel GPU.
I’m sorry, but I don’t think I know how to do that. However, I know that the first printout is from when I connected the cable to the first DP port and the printout below is from when I connected the cable to the second DP port. And both work OK.
How can I print the actual information of the computer, i.e. not commercial information? Maybe that would help further.
On the other hand, I just need to get rid of the nvidia driver, if that’s even possible.
