llama.cpp

b7feacf7 - ggml: new backend for Virglrenderer API Remoting acceleration (v2) (#18718)