I’m currently using Nvidia drivers 580.119 and 580.126 recently appeared. It’s been my experience that until nvidia-open-driver-G06-signed-kmp-default
and nvidia-open-driver-G06-signed-kmp-meta get a 580.126 version available, the update will cause my system to boot to non-GUI. Using YaST to drop all relevant packages that have 580.126, back to 580.119 corrects the system. So, would a Rule of Thumb be to wait til all versioned Nvidia packages get the new version?
Yes you are right, wait for the updated nvidia-open-driver-GO6 to be available.
I know no one wants to recommend this.
In my case I always use the .run driver from nvidia and turn on the DKMS to start at boot so every time there is a kernel upgrade I don’t worry about reinstalling the .run nvidia driver. DKMS rebuilds the nvidia driver when machine reboots. This way like what many experiencing now, I am able to avoid the dependency issues that is happening with the nvidia drivers. Just my opinion I am not encouraging anyone to follow my path.
@conram I also use the run files driver and cuda on Tumbleweed… I prefer booting to the new kernel and running a script to add all the options I want as well as use the open or proprietary version depending on what I’m testing…
On Leap 16.0 I’m using the proprietary licensed rpms.
1 Like
Today was the day to move to 580.126, as the last two packages were updated!
I used to update the “hard way”, run files, too, but want to see how far I can update the “easy way”. Before, graphics mode would break so often that it was easier to revert with the “hard way”. No breakage in the recent drivers for awhile now.