llama.cpp

eddcb523 - ggml : add and use ggml_cpu_has_llamafile() (#8664)