So I’ve been trying to install the proprietary Nvidia drivers on my homelab so I can get my fine ass art generated using Automatic1111 & Stable diffusion. I installed the Nvidia 510 server drivers, everything seems fine, then when I reboot, nothing. WTF Nvidia, why you gotta break X? Why is x even needed on a server driver. What’s your problem Nvidia!
I’m hoping the recent explosion of AI/ML stuff will create more incentives for them to have proper support for desktop Linux, but I’m not counting on it.
Those are different drivers, or rather different parts of the driver.
CUDA has been a staple in HPC for years now and the situation didn’t exactly improve.
I mean the number of people using beefy Linux workstations with desktop environments is likely to increase because of it, not referring to the datacenter market they’re already entrenched in.
They dont see a reason to their biggest buyers are in enterprise level and will pay for the extra support.
Common ML/AI stuff already work on AMD albeit not optimally (Tensorflow, Pytorch) and even some projects already work on AMD (e.g Stable Diffusion). Users are far better off creating a more generic branch for projects that would support CPU based acceleration (via both Intels and AMDs inclusion of AI acceleators in their products) then to hope Nvidia of all companies mess with their bottom line to give linux proper support.