Oh yes llama.cpp's trick is it supports any hardware setup! It might be a bit slower, but it should function well!