llama.cpp

c49daff5 - ggml-cpu: Don't pass -mpowerpc64 when -mcpu already implies it (#17308)