Hello!
I am using a computer with 2 graphic cards to do deep learning (CUDA version 12.2). I’ve encountered a problem, that my device ordinals in nvidia-smi and Nvidia Control Panel mismatch. Furthermore, I attach screenshots to demonstrate.
nvidia-smi says, that my 3090ti has number 1 and 3060 has number 0, whereas control panel shows 3090ti as a device 0 and 3060 as a device 1. When using pytorch and sending, let’s say some model to a device ‘cuda:1’, it sends it to 3060 and if I send it to device ‘cuda:0’ it sends it to a 3090ti. So the working device ordinal coincides with the control panel.
Driver version: 535.86.05
OS: Ubuntu 22.04 LTS
Cuda version: 12.2
I would like to fix this issue, thanks for your support!
±--------------------------------------------------------------------------------------+
| NVIDIA-SMI 535.86.05 Driver Version: 535.86.05 CUDA Version: 12.2 |
|-----------------------------------------±---------------------±---------------------+
| 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 NVIDIA GeForce RTX 3060 Off | 00000000:08:00.0 Off | N/A |
| 0% 38C P8 20W / 170W | 312MiB / 12288MiB | 0% Default |
| | | N/A |
±----------------------------------------±---------------------±---------------------+
| 1 NVIDIA GeForce RTX 3090 Ti Off | 00000000:42:00.0 Off | Off |
| 35% 35C P8 24W / 450W | 2263MiB / 24564MiB | 0% Default |
| | | N/A |
±----------------------------------------±---------------------±---------------------+