llama.cpp b11267 adds ROCm 10 and Snapdragon support

llama.cpp has released build b11267, expanding hardware support to include ROCm 10.0 for AMD GPUs on Linux and Windows, as well as native Snapdragon support for Linux devices. The update also provides CUDA 13 binaries (13.4 libraries) alongside existing CUDA 12 builds across multiple platforms. Notably, KleidiAI optimizations for Apple Silicon and openEuler builds have been disabled in this release. This version reinforces llama.cpp's position as a cross-platform standard for local AI inference.