llama.cpp

ggml-cpu: Optimized x86 and generic cpu q1_0 dot (follow up)

#21636

Merged