llama.cpp b9690 Update Improves Metal RoPE Operator for macOS/iOS

Core summary

One sentence to understand this update

llama.cpp's b9690 update refactors the Metal `rope_back` operator to reuse existing kernels, enhancing efficiency for macOS/iOS.

Impact & opportunity

What this could mean

macOS and iOS developers utilizing llama.cpp for local LLM inference will benefit from more optimized and efficient model operations.