You are not logged in.
Hey,
I am running a fresh install of Arch Linux on my laptop (ASUS ROG G501VW series). I have KDE plasma up and running. I am unable to get the NVIDIA GeForce GTX 960M working.
I have executed the install instructions on the NVIDIA page in the wiki here:
https://wiki.archlinux.org/title/NVIDIA
Below is the output of lspci
[prajwal@Panini ~]$ lspci -k | grep -A 2 -E "(VGA|3D)"
00:02.0 VGA compatible controller: Intel Corporation HD Graphics 530 (rev 06)
Subsystem: ASUSTeK Computer Inc. Device 1090
Kernel driver in use: i915
--
01:00.0 3D controller: NVIDIA Corporation GM107M [GeForce GTX 960M] (rev a2)
Subsystem: ASUSTeK Computer Inc. Device 1090
Kernel driver in use: nvidia
And below is the output of nvidia-smi:
[prajwal@Panini ~]$ nvidia-smi
Tue Jul 20 12:05:24 2021
+-----------------------------------------------------------------------------+
| NVIDIA-SMI 470.57.02 Driver Version: 470.57.02 CUDA Version: 11.4 |
|-------------------------------+----------------------+----------------------+
| 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 55C P0 N/A / N/A | 0MiB / 4046MiB | 0% Default |
| | | N/A |
+-------------------------------+----------------------+----------------------+
+-----------------------------------------------------------------------------+
| Processes: |
| GPU GI CI PID Type Process name GPU Memory |
| ID ID Usage |
|=============================================================================|
| No running processes found |
+-----------------------------------------------------------------------------+
I have tried the nvidia, nvidia-lts and nvidia-dkms packages available in the official repos, as well as the nvidia-beta packages in the AUR repos (installed using yay).
I have also tried downloading and running the driver installation from the official NVIDIA download page here:
https://www.nvidia.com/Download/driverR … 7145/en-us
Any suggestions on how I can get this working?
Thanks in advance for any help.
Last edited by prajS (2021-07-21 06:27:12)
Offline
960m souds like Optimus. Look up Optimus in the Wiki.
Online
Thanks for the response. I already tried the steps in the Optimus Wiki.
The methods suggested under NVIDIA graphics only section did not lead to any change.
I also tried to install and run the bumblebee service. That just led to the system freezing and becoming non-responsive (unable to even shift to the TTY consoles). Although I could hear the graphics card fans running at full speed when I attempted to run the bumblebee service.
Anything else I could try? I am not sure where to even look at the moment.
If it is any help, below is the output from dmesg | grep nvidia:
[prajwal@Panini ~]$ sudo dmesg | grep nvidia
[sudo] password for prajwal:
[ 18.672946] nvidia: module license 'NVIDIA' taints kernel.
[ 18.698785] nvidia-nvlink: Nvlink Core is being initialized, major device number 234
Offline
1. define "not working" - the GPU is controlled by the nvidia driver and responds to nvidia-smi. Everything you posted says that it does work
=> What do you expect, what do you get? (Be exact, don 't use the word "work" in your explanation)
2. Post your xorg log
Online
Hi
1. define "not working" - the GPU is controlled by the nvidia driver and responds to nvidia-smi. Everything you posted says that it does work
=> What do you expect, what do you get? (Be exact, don 't use the word "work" in your explanation)2. Post your xorg log
Thanks Seth for pointing it out. The graphics card and the installed drivers are indeed "working".
I dug some more and figured that I was thrown off by 2 things.
1) The output of the nvidia-smi shows "off" next to the name of the card. I assumed this was the status of the card itself. It turns out that this is the status of the persistence mode (which is a user settable parameter).
I found the clarification on this thread:
https://forums.developer.nvidia.com/t/w … -smi/37509
2) The 2nd thing that threw me off was that I was checking the "working" of the graphics card by rendering a sample Blender scene. Outputs of both nvtop and nvidia-smi showed no processes utilising the Graphics Card. Instead all 8 cores of the CPU were fully utilised (checked using htop) and the system was grinding to a halt. Based on this, I assumed the card was not functional.
It turns out that I had the OpenCL option in Blender selected, and that the GTX960M is not compatible with OpenCL. Hence Blender was rendering using the CPU only. When I deselected the OpenCL option, Blender rendered using GPU compute on the GTX960M, thus indicating that the graphics card was indeed working as desired.
Thanks again for pointing out that the card was already responsive to the installed driver.
Marking this thread as Solved.
Offline