Hello, I am running a desktop PC with an NVIDIA RTX 3060 and an AMD Ryzen 5 7600 and I wanted to try out hybrid graphics, since it sounds like it could be helpful e.g. if I want to do machine learning, blender etc. with my dGPU and use my iGPU for the rest of my applications. I understand that hybrid graphics were developed with laptops in mind, so I haven’t really found documentation dedicated to desktop systems on this topic. I have found that for ujust, there exists enable-supergfxctl and configure-nvidia-optimus, but I am kind of lost what even the difference is (except that supergfxctl seems to be especially dedicated to laptops, but I have no idea if that’s of any relevance). Could someone just give me a rundown on how to best approach this on bazzite for my desktop PC?
You don’t need Optimus, etc.
I have an Intel Arc A380, and an Nvidia RTX A4500. My monitors are connected to the Intel card. I use Bluefin-dx-nvidia, which, as it says, has the nvidia drivers. I use Tensorflow and CUDA to do machine learning. In a venv, I install tensorflow[and-cuda] , which installs all the Python dependencies.
If you run nvidia-smi in a terminal, and see your GPU listed, you’re good to go.
E.g.
> nvidia-smi
Sat May 17 22:30:19 2025
+-----------------------------------------------------------------------------------------+
| NVIDIA-SMI 570.144 Driver Version: 570.144 CUDA Version: 12.8 |
|-----------------------------------------+------------------------+----------------------+
| 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 RTX A4500 Off | 00000000:01:00.0 Off | Off |
| 30% 33C P8 16W / 200W | 13MiB / 20470MiB | 0% Default |
| | | N/A |
+-----------------------------------------+------------------------+----------------------+
+-----------------------------------------------------------------------------------------+
| Processes: |
| GPU GI CI PID Type Process name GPU Memory |
| ID ID Usage |
|=========================================================================================|
| 0 N/A N/A 5510 G /usr/bin/gnome-shell 3MiB |
+-----------------------------------------------------------------------------------------+
Hey John, thanks a lot for the reply! Do you know if GPU offloading (I hope that’s the correct term) also works automatically like that for blender?
Edit: Nvm, I found a thread regarding my question so I will just roll with that one: Making sure you're not a bot!
Not near my computer, but I think you can right-click the icon and select Run on GPU, or words to that effect.
Edit:
This topic was automatically closed 24 hours after the last reply. New replies are no longer allowed.