1

I have followed the instructions mentioned here (same instructions are also mentioned here). Unfortunately, these instructions are catered to Ubuntu 20.04 and I am currently using Ubuntu 22.04. Here are the results:

  1. Instead of X11 (which was being used before this modification), Wayland is being used as the Windowing System.
  2. In the output of nvidia-smi, one can see that a /usr/bin/gnome-shell process is being run on the GPU. (Note that the "NVIDIA On-Demand" PRIME Profile is in use)

Here is the nvidia-smi output:

+-----------------------------------------------------------------------------------------+
| NVIDIA-SMI 550.120                Driver Version: 550.120        CUDA Version: 12.4     |
|-----------------------------------------+------------------------+----------------------+
| 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 RTX 4060 ...    Off |   00000000:01:00.0 Off |                  N/A |
| N/A   30C    P4            310W /   30W |       5MiB /   8188MiB |      0%      Default |
|                                         |                        |                  N/A |
+-----------------------------------------+------------------------+----------------------+

+-----------------------------------------------------------------------------------------+ | Processes: | | GPU GI CI PID Type Process name GPU Memory | | ID ID Usage | |=========================================================================================| | 0 N/A N/A 1816 G /usr/bin/gnome-shell 2MiB | +-----------------------------------------------------------------------------------------+

Goal:

  1. Unless other programs that really need the dGPU (e.g. games or simulators utilizing Vulkan API, Deep Learning script that uses PyTorch) are being run, I want the dGPU to have no processes running on it.
  2. nvidia-smi still works, the dGPU is still available for Deep Learning, heavy games, CUDA Debugging etc. (which isn't the case for rebooting after sudo prime-select intel)

What further changes do I need to make? (willing to do a fresh install)

0 Answers0