Unable to use nvidia card

Question

I have dual graphics Intel 520 integrated and nVidia 930MX dedicated. I'm running 20.04.2 and installed nvidia-driver-460 through Additional Drivers, and intel-media-va-driver-non-free through terminal. Currently I'm only getting display (and successful boot) on intel GPU. After I run sudo prime-select nvidia and reboot, I'm no longer able to boot, and have to revert to intel through TTY. Running nvidia-smi gives following error:

NVIDIA-SMI has failed because it couldn't communicate with the NVIDIA driver. Make sure that the latest NVIDIA driver is installed and running.

The nvidia-xconfig generates a xorg.conf in which the 'Files' section is empty.

nvidia-settings gives error:

ERROR: NVIDIA driver is not loaded
ERROR: Unable to load info from any available system
(nvidia-settings:10549): GLib-GObject-CRITICAL **: 18:16:11.739: g_object_unref: assertion 
'G_IS_OBJECT (object)' failed
** Message: 18:16:11.749: PRIME: Requires offloading
** Message: 18:16:11.749: PRIME: is it supported? yes
** Message: 18:16:11.857: PRIME: Usage: /usr/bin/prime-select nvidia|intel|on-demand|query
** Message: 18:16:11.857: PRIME: on-demand mode: "1"
** Message: 18:16:11.857: PRIME: is "on-demand" mode supported? yes

Running whereis nvidia gives:

nvidia: /usr/lib/x86_64-linux-gnu/nvidia /usr/lib/nvidia /usr/share/nvidia /usr/src/nvidia-460.73.01/nvidia

So driver is installed but not loading. How to get the driver working? I'd like to retain intel as well.

EDIT: running sudo prime-select nvidia, reboot with nomodeset, then running nvidia-smi from TTY (because there is no GUI) gives following output:

Tue May 25 18:33:17 2021       
+-----------------------------------------------------------------------------+
| NVIDIA-SMI 460.73.01    Driver Version: 460.73.01    CUDA Version: 11.2     |
|-------------------------------+----------------------+----------------------+
| 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  GeForce 930MX       Off  | 00000000:01:00.0 Off |                  N/A |
| N/A   39C    P8    N/A /  N/A |     10MiB /  2004MiB |      0%      Default |
|                               |                      |                  N/A |
+-------------------------------+----------------------+----------------------+
+-----------------------------------------------------------------------------+
| Processes:                                                                  |
|  GPU   GI   CI        PID   Type   Process name                  GPU Memory |
|        ID   ID                                                   Usage      |
|=============================================================================|
|    0   N/A  N/A      1063      G   /usr/lib/xorg/Xorg                  6MiB |
|    0   N/A  N/A      1292      G   /usr/bin/gnome-shell                2MiB |
+-----------------------------------------------------------------------------+

That shows nvidia driver is loading and working, but why the GUI isn't working is beyond me.

EDIT 2: running nvidia-settings and selecting any graphics from PRIME profiles results in Segmentation fault.

Adnan · Accepted Answer · 2021-05-26T15:24:35.400

The answer posted by Celebre Asm worked for me.

First remove all the occurrences of xorg.conf and xorg.conf.old from /etc/X11/.

Then run inxi -Fxzc0 and note down the BusID of your GPUs. In my case intel was 00:02.0 = 0:2:0 and nvidia was 01:00.0 = 1:0:0.

Then create a xorg.conf.d folder, make 20-intel.conf file inside it with following content:

Section "Device"
    Identifier  "Intel Graphics"
    Driver "intel"
    Option "AccelMethod" "sna"
EndSection

And 90-nvidia.conf file with following content:

Section "ServerLayout"
    Identifier "layout"
    Screen 0 "iGPU"
    Option "AllowNVIDIAGPUScreens"
EndSection
Section "Device"
    Identifier "iGPU"
    Driver "modesetting"
    BusID "PCI:0:2:0"

EndSection
Section "Screen"
    Identifier "iGPU"
    Device "iGPU"
EndSection
Section "Device"
    Identifier "dGPU"
    Driver "nvidia"
    BusID "PCI:1:0:0"

EndSection

Additionally I removed xserver-xorg-video-intel because somehow I was getting rectangles diagonally drawn on desktop and unreadable fonts, and removing it solved that.

Reboot.

**Footnote: This mess happened because I tried to boot my laptop using nvidia to use VDPAU. As of now, GM108 series of nvidia GPU doesn't have decoding support, and offloads decoding to intel counterpart. Everything was smooth as long as I didn't change "on-demand" option which was default since I first installed Ubuntu.

score -1 · Answer 2 · answered May 25 '21 at 22:00

hesitate to ask this but... you probably have two different slots on the back of your computer if it is a desktop, one for the intel graphics on your motherboard, and another where your nvicia card is slotted in. You cannot run the nvidia one if you are only plugged into the one on your motherboard. Does this make sense to you?

Unable to use nvidia card

2 Answers2