Has anyone tried running LLMs on B-series GPUs?
I’ve just received my B50 and after some setbacks managed to pass it into a NixOS 25.11 VM from a NixOS 25.11 Incus host ( both have kernel 6.12.70 ) on ALTRAD8UD. But then I hit a brick wall as every related NixOS package (intel-compute-runtime in particular ) refuses to install, throwing the unsupported architecture error. Is it like that in every distro or it’s possible to make it work at least in some way ( with or without Ollama )?
P.S. @q66 @ara does it make sense to try Chimera Linux for LLMs?