llama.cpp
llama.cpp copied to clipboard
cmake : fix ARM feature verification
Use check_cxx_source_compiles to prevent conflicts with the existing GGML_NATIVE detection code.