Another Nvidia problem: gnome refuses to use the nvidia driver

Annoyingly my NVIDIA installation from RPM-Fusion is (again) not working.
I noticed (under GNOME) that my system is not using the NVIDIA driver at all.
I first noticed on Fedora 30 so I uninstalled NVIDIA, upgraded to Fedora 31, reinstalled NVIDIA - but still nothing.
My System is a XPS 15 9550 with the internal Intel GPU and an additional GTX 950M.

Output of some commands:
$ glxinfo | grep vendor
server glx vendor string: SGI
client glx vendor string: Mesa Project and SGI
OpenGL vendor string: Intel Open Source Technology Center

$ DRI_PRIME=1 glxinfo | grep vendor
libGL error: failed to create dri screen
libGL error: failed to load driver: nouveau
server glx vendor string: SGI
client glx vendor string: Mesa Project and SGI
OpenGL vendor string: Intel Open Source Technology Center

$ lsmod | grep nvidia 
nvidia_drm             57344  4
nvidia_modeset       1118208  3 nvidia_drm
nvidia_uvm           1081344  0
nvidia              19980288  135 nvidia_uvm,nvidia_modeset
ipmi_msghandler        73728  2 ipmi_devintf,nvidia
drm_kms_helper        212992  2 nvidia_drm,i915
drm                   512000  12 drm_kms_helper,nvidia_drm,i915

$ cat /proc/cmdline
BOOT_IMAGE=(hd0,gpt2)/vmlinuz-5.3.11-300.fc31.x86_64 root=/dev/mapper/fedora-root ro resume=/dev/mapper/fedora-swap rhgb quiet rd.driver.blacklist=nouveau modprobe.blacklist=nouveau nvidia-drm.modeset=1```

I might add that i just noticed this one day, before that GNOME used the NVIDIA driver as expected.

Nvidia drivers are loaded, but Intel gets used

Maybe there is someone that has a hunch what could cause gnome to refuse using the NVIDIA driver.

Maybe you did read this before but if it was not the case, just can help you.

There are some ways to fix nvidia GPU permanently in place of DynamicPwerManagement. (It was updated recently)


Ok, thank you.
Following the instructions under “NVIDIA PrimaryGPU Support” did the trick to get it working like it did before. (essentially disabling the Intel GPU)

Just for sake of completeness:
What is the intended way to use a Optimus notebook with the normal configuration?
(Optimus manager or Bumblebee)?

With the optimus active it would be use Intel GPU with low demand and Nvidia GPU with high load, and it would be done automatically to do sure a good balance between power saving and performance. About optimus manager or bumblebee I don’t know because i don’t use a nvida card and either a hybrid GPU but always read about problems about bumblebee there are users more expert about this theme and a exits good information in the net about it.

In this point if you are interested in a easy change between both GPU you should try install this package “mate-optimus” than looks like it should work also in gnome if we look the dependencies, look very simple to do the change look this article, But this is some thing than i don’t know at all.

[jorge@fedora31 ~]$ LANG=C sudo dnf install mate-optimus
Last metadata expiration check: 0:05:06 ago on Sat Nov 16 20:07:03 2019.
Dependencies resolved.
 Package                                        Architecture       Version                      Repository          Size
 mate-optimus                                   noarch             19.10.4-1.fc31               fedora              27 k
Installing dependencies:
 glew                                           x86_64             2.1.0-5.fc31                 fedora             116 k
 python3-setproctitle                           x86_64             1.1.10-14.fc31               fedora              22 k
Installing weak dependencies:
 gnome-shell-extension-appindicator             noarch             30-7.fc31                    fedora              29 k

Transaction Summary
Install  4 Packages

Total download size: 195 k
Installed size: 760 k
Is this ok [y/N]: 


1 Like