the full set of llama.cpp binaries builds in under 5 minutes with an unmodified build workflow straight from their github page on a literally ten year old dual xeon.