Looking at the above guide, I installed my nvidia drivers using YasT. Everything looks fine; all of the commands that are suggested in the the guide show my drivers as being installed (lsmod, hwinfo, inxi, etc.) However, these commands also show my onboard Intel graphics.
So how do I know that my nvidia drivers are actually the default and they are being used instead of the intel drivers?
When I go to âAbout this Systemâ from KDE it just says âMesa Intel Graphicsâ which makes me suspect that the nvidia drivers are not actually being used as the default.
All I do right now is watch videos using VLC; Iâm not playing games or anything.
You have an Optimus system in which you can choose between the onboard CPU graphics card driver (Intel) and the Nvidia GPU driver. In Windows, the choice between the two drivers is made automatically. If there is an application that places greater demands on the graphics performance (e.g. games), Windows will automatically switch to the Nvidia driver, which is not possible under Linux for licensing reasons. You therefore have to switch manually. To do this, install Bumblebee and suse-prime. You can then switch manually from the Intel driver to the Nvidia driver.
@wiking can you show the output from inxi -Gxxz not there is no need for either bumblebee (deprecated from openSUSE) or in fact suse-prime, you can use Prime Render Offload to play vidieos on the dGPU, but show the above output first to confirm.
If you donât play games, but just want to watch videos, I donât see any reason why you would need the Nvidia driver at all. I watch videos and TV here on an eight-year-old laptop with the Mesa Intel driver without any problems.
At some point I do plan to go on Steam and play some 3D games briefly, just to see what the graphics system is capable of. I will probably get hooked, so thereâs that.
Iâm confused because these two people seem to be saying conflicting statements about bumblebee and suse-prime:
@wiking I guess you could say bumblebee was the first generation of tools for Optimus hardware (dual graphics in laptops), then came suse-prime, which still works (for I think the bulk of hardware). Now with more modern RTX cards and AMD is Prime Render Offload which can be controlled via dbus and the switcherooctl service, basically on demandâŠ
Can you check via running the command as your user suse-prime or zypper se suse-prime it does have offload capabilities AFAIK, but can be funky setting up, and limited to Xorg and relies on configuration files⊠It should work, but sometimes it doesnât especially with newer hardware.
I use switcherooctl (Prime Render Offload) here with AMD/AMD Laptop and Intel/Nvidia, Nvidia/Nvidia desktops as itâs integrated in the GNOME DE, works with Xorg and Wayland and needs no configuration https://en.opensuse.org/SDB:NVIDIA_Switcheroo_Control
Looks like I have both suse-prime and switcherooctl,
Any idea how to open the âPlasma 5 appletâ that is mentioned below? I canât find it by searching for it.
wiking@localhost:~> prime-select
Absolute path to âprime-selectâ is â/usr/sbin/prime-selectâ, so running it may require superuser privileges (eg. root).
wiking@localhost:~> zypper se suse-prime
Repository : Time since last refresh
FreeOffice : 3 D 20 h
repo-non-free : 19 h 50 min
repo-non-oss : 19 h 50 min
repo-openh264 : 19 h 50 min
repo-oss : 19 h 50 min
update-slowroll : 19 h 50 min
packman : 19 h 50 min
Note: Running with user privileges. From time to time run 'zypper refresh' as root to make sure
the repository metadata are complete and up-to-date.
S | Name | Summary | Type
â±--------------------------±---------------------------------------------------------------------±-------
i | plasma5-applet-suse-prime | Plasma 5 applet for controlling SUSE Prime | package
i | suse-prime | GPU (nvidia/intel) selection for NVIDIA optimus laptops with bbswi-> | package
@wiking So what does the command (as your user) show for switcherooctl --list if there is no output switch to root user and run systemctl enable --now switcheroo-control.service and then as your user run the switcherooctl --list command again, do you see both devices?
wiking@localhost:~> switcherooctl --list
Traceback (most recent call last):
File â/usr/bin/switcherooctlâ, line 187, in
launch(args, gpu)
File â/usr/bin/switcherooctlâ, line 66, in launch
os.execvp(args[0], args)
File ââ, line 574, in execvp
File ââ, line 616, in _execvpe
File ââ, line 607, in _execvpe
FileNotFoundError: [Errno 2] No such file or directory
wiking@localhost:~> su
Password:
localhost:/home/wiking # systemctl enable --now switcheroo-control.service
Created symlink /etc/systemd/system/graphical.target.wants/switcheroo-control.service â /usr/lib/systemd/system/switcheroo-control.service.
localhost:/home/wiking # exit
exit
wiking@localhost:~> switcherooctl --list
Traceback (most recent call last):
File â/usr/bin/switcherooctlâ, line 187, in
launch(args, gpu)
File â/usr/bin/switcherooctlâ, line 66, in launch
os.execvp(args[0], args)
File ââ, line 574, in execvp
File ââ, line 616, in _execvpe
File ââ, line 607, in _execvpe
FileNotFoundError: [Errno 2] No such file or directory
wiking@localhost:~> sudo switcherooctl --list
[sudo] password for root:
Traceback (most recent call last):
File â/usr/bin/switcherooctlâ, line 187, in
launch(args, gpu)
File â/usr/bin/switcherooctlâ, line 66, in launch
os.execvp(args[0], args)
File ââ, line 574, in execvp
File ââ, line 616, in _execvpe
File ââ, line 607, in _execvpe
FileNotFoundError: [Errno 2] No such file or directory
wiking@localhost:~>
> switcherooctl --list
Traceback (most recent call last):
File "/usr/bin/switcherooctl", line 187, in <module>
launch(args, gpu)
File "/usr/bin/switcherooctl", line 66, in launch
os.execvp(args[0], args)
File "<frozen os>", line 574, in execvp
File "<frozen os>", line 616, in _execvpe
File "<frozen os>", line 607, in _execvpe
FileNotFoundError: [Errno 2] No such file or directory
>
@wiking So you can use the command switcherooctl launch /usr/bin/vlc-qt or what ever is in the desktop file for vlc and you will be using the Nvidia driver. You can check via nvidia-smi or nvtop and will see the process running.
AFAIK Plasma is meant to be getting switcherooctl integration, so like on GNOME can just right-click the desktop item and select there to launch with the dGPU, from memory Steam will see it present and launch the dGPUâŠ
What sort of guidelines should I use to determine when to use switcherooctl launch? What kinds of video and graphics are more efficient on Nvidia graphics vs intel Iris? What is the reason we donât use Nvidia graphics 100% of the time?
System temperature plus noise (cooling fans) might be another criteria.
The battery of my laptop will last one hour with 100% NVIDIA usage but 3 hours with 100% Intel graphics usage. And while the NVIDIA card is in use the fans will run all the time and the system becomes quite hot.
