I am trying to Deploy a VM on Azure for a CUDA based tool. I am struggling to manually install the proper drivers and get them working.
I have tried the Linux Extension which fails during deployment as it appears to use RHEL 7.x, I have manually installed drivers using Microsoft’s support page:
Azure N-series GPU driver setup for Linux - Azure Virtual Machines | Microsoft Learn
I did manage to get some older drivers to connect on a Tesla T4 Azure Size, but it was inconsistent with any documented approach and that is not working while I try to get an Azure NC A10 deployment setup. I just can’t get the latest drivers to load and connect with the card. I get an error connecting with the ‘nvidia-smi’ command and my application complains it can’t find the CUDA compatible device.
I am running RHEL 9.x as the operating system. I have disabled Secure boot which helps generally, but would rather not use that in production as we deploy as a Managed Application.
If anyone has any good pointers or tips, I would greatly appreciated it. Ultimately, I need to get a Bicep script where I know the drivers are installed and will connect to the Azure VM without any issues.