llama.cpp b9978 release with macOS KleidiAI disabled

The llama.cpp project has released version b9978, providing binaries for macOS, iOS, Linux, Android, Windows, and openEuler platforms.

• macOS Apple Silicon builds are available, but the KleidiAI optimization feature is explicitly disabled in this release.
• Linux support includes CPU, Vulkan, ROCm 7.2, OpenVINO, and SYCL variants for x64, arm64, and s390x architectures.
• Windows binaries cover CPU, CUDA 12/13, Vulkan, OpenCL Adreno, OpenVINO, SYCL, and HIP backends.
• Android arm64 (CPU) and iOS XCFramework builds are included alongside the standard UI package.