I have installed CUDA through a .deb network installation using the download linked from the cuda installation documentation.
I have reasonable output from nvidia-smi:
+-----------------------------------------------------------------------------+
| NVIDIA-SMI 510.47.03 Driver Version: 510.47.03 CUDA Version: 11.6 |
|-------------------------------+----------------------+----------------------+
| 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 ... On | 00000000:01:00.0 Off | N/A |
| 0% 33C P8 6W / 200W | 313MiB / 8192MiB | 0% Default |
| | | N/A |
+-------------------------------+----------------------+----------------------+
+-----------------------------------------------------------------------------+
| Processes: |
| GPU GI CI PID Type Process name GPU Memory |
| ID ID Usage |
|=============================================================================|
| 0 N/A N/A 1022 G /usr/lib/xorg/Xorg 102MiB |
| 0 N/A N/A 1568 G /usr/lib/xorg/Xorg 187MiB |
| 0 N/A N/A 1697 G /usr/bin/gnome-shell 11MiB |
+-----------------------------------------------------------------------------+
and from cat /proc/driver/nvidia/version:
NVRM version: NVIDIA UNIX x86_64 Kernel Module 510.47.03 Mon Jan 24 22:58:54 UTC 2022
GCC version:
I’m also able to compile the deviceQuery sample. However, when I run the sample I get the following output:
$ ./deviceQuery
./deviceQuery Starting...
CUDA Device Query (Runtime API) version (CUDART static linking)
cudaGetDeviceCount returned 999
-> unknown error
Result = FAIL
Looking around I saw that it could because of a mismatch between CUDA toolkit and drivers, but mine seem to match. My system also has secureboot enabled, but I went through the signing process during install.
Does anyone have an idea of something I may have missed?