Hi all,
Can someone point me out on how to properly install drivers for the GeForce RTX 3050 Ti on Ubuntu 22.04.1 LTS ?
Till now i try all the methods like “apt” using drivers GUI app, using the .run from the NVIDIA page. The installation went ok but somehow the NVIDIA is not being used. Last i install them using apt but should i add some different repo PPA ?
To be honest i dont quit catch it. As I understand you, you are telling me that there is some scenario where the Graphic card is being used when something is triggering and not the whole time and i need to change this ? Where should i add those env variables ?
I found some configuration from NVIDIA app and i change it to be in Performance Mode (From Nvidia Settings > PRIME PRofiles > NVIDIA (Performance Mode)).
So i change the also from the drivers to be used NVIDIA.
Now i have this as result:
➜ ~ nvidia-smi
Thu Oct 13 10:46:56 2022
+-----------------------------------------------------------------------------+
| NVIDIA-SMI 515.65.01 Driver Version: 515.65.01 CUDA Version: 11.7 |
|-------------------------------+----------------------+----------------------+
| 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 ... Off | 00000000:01:00.0 Off | N/A |
| N/A 48C P8 6W / N/A | 1217MiB / 4096MiB | 18% Default |
| | | N/A |
+-------------------------------+----------------------+----------------------+
+-----------------------------------------------------------------------------+
| Processes: |
| GPU GI CI PID Type Process name GPU Memory |
| ID ID Usage |
|=============================================================================|
| 0 N/A N/A 2150 G /usr/lib/xorg/Xorg 592MiB |
| 0 N/A N/A 2789 G /usr/bin/gnome-shell 99MiB |
| 0 N/A N/A 3552 G ...RendererForSitePerProcess 55MiB |
| 0 N/A N/A 5057 G ...793255411633242967,131072 220MiB |
| 0 N/A N/A 5648 G gnome-control-center 1MiB |
| 0 N/A N/A 5692 G /opt/viber/Viber 59MiB |
| 0 N/A N/A 5810 G ...veSuggestionsOnlyOnDemand 132MiB |
| 0 N/A N/A 14461 G /usr/bin/nvidia-settings 0MiB |
+-----------------------------------------------------------------------------+
➜ ~ prime-select query
nvidia
➜ ~
I see some errors from the bug report but im not understanding them. nvidia-bug-report.log.gz (318.2 KB)
Is there some other way to check that the Graphic card is being used and work properly?
This is also what I told you, to use prime-select to change modes (also available in the ubuntu patched version of nvidia-settings). You now switched to nvidia primary aka performance mode, everything now runs on the nvidia gpu at the cost of higher power consumption.
There are no errors in the logs.
You should have done your homework and use google to find out what “offload mode” and “prime-select” and not ask the same question again.
The reason I agreed with johndouganhtx criticism is that you are practicing a type of elitism, even if you aren’t aware. I’ve been going in circles trying to solve similar problems. I’m trying to exhaustively search for solutions offered, and not making headway. Yet, I feel discouraged to ask. Likely I will be criticized for NOT practicing due diligence, despite my best effort.
The implication is that if we aren’t expert enough to find the answer or understand your explanations right away, then we have no business trying.
And what you don’t know is that the OP user hammered me with PMs in parallel where I explained everything, nevertheless he ignored everything not doing anything. What you’re asking from me is becoming a voiceless slave.
Addendum: this also was already the second thread with the same question from OP. There’s no sense in being nice to abusive people.
I can appreciate that. Still it would have appeared more professional if you had saved your frustration for the private messages instead of public.
As it happens, my issue I am trying to resolve it’s not quite the same. If there is not a quick simple answer, just let me know, and I will try to work up a formal request.
My Mint Linux with NVDIA RTX 3060 reports that it can’t find libGL.so and libGLU.so any time I use make command on projects requiring OpenGL. This CUDA is a 12.3 version installed from the UBUNTU 22.04 local .deb. I know the system has those files. I found other users have encountered this. But am finding no help in the documentation or online help requests. Can I change configuration file settings to correct the location, symlink to correct, or include an option when calling make?
Thanks in advance, even if your answer is to submit a more formal request.
Those libs belong to the package libglvnd0, unrelated to the nvidia driver but an essential part of the Linux graphics stack. Is this on a system running headless or a normal desktop-like system?
Being “professional” by definition requires being paid for it. This is a user2user forum.