There's a fundamental tension between "conservative desktop-origin mass appeal Linux distribution" and "extreme performance hardware accelerated numerics coprocessor". In the places where a mass market need is obvious (local LLM) solutions are emerging with vibrant and diverve back end options (gguf).
I think its OK to install stuff to get extreme scientific compute performance.
>There's a fundamental tension between "conservative desktop-origin mass appeal Linux distribution" and "extreme performance hardware accelerated numerics coprocessor". In the places where a mass market need is obvious (local LLM) solutions are emerging with vibrant and diverve back end options (gguf).
I was being a little facetious (the sibling said something about Arch, obligatory "I use Arch BTW" reference").
But in the context of GPU drivers, NixOS has all of the same advantages that it does for any other application: once you've got it working it stays working until you change something (and often it stays working even then). For stuff like GPU/GPGPU this is really fucking nice because that is not at all the status quo.
It's also nice that once a piece of hardware is supported well by anyone, that person can share the configuration as a module and it will work for other people, so in a sense it makes Apple-grade "perfect out of the box" experiences realistic on other hardware.
I think its OK to install stuff to get extreme scientific compute performance.
I use NixOS BTW.