Open adriantorrie opened 3 weeks ago
Hi @adriantorrie,
I'm not very familiar with WSL2, but I see that you are using a NVIDIA GeForce GTX 1080 Ti which is a Pascal generation GPU.
Unfortunately, Pascal support was dropped starting with RAPIDS 24.02 (https://docs.rapids.ai/notices/rsn0034/). You could try a 23.12
image such as rapidsai/base:23.12-cuda12.0-py3.10
, but it's not officially supported anymore.
Cool, I'll give it a go.
Describe the bug
nvidia-smi
can see the device 2.1. in Windows 2.2. in WSL 2.3. from within the running RapidsAI container (in WSL), via theipython
prompt 2.4. from within the running RapidsAI container (in WSL), via the run command overrideFor step 1 above (from WSL):
Steps/Code to reproduce bug
Note: For step 5 above, the Release Selector provided me with the following docker command:
Expected behavior
The expectation was that the following line of python would execute, and not raise an exception with "No NVIDIA GPU detected":
Environment details (please complete the following information):
Ubuntu-22.02
andUbuntu-24.04
Additional context
Ubuntu-22.04
Ubuntu-24.04
Ubuntu-22.04
andUbuntu-24.04
):