llama.cpp b9999 adds SME2 f32 kernel and dynamic scheduling

The llama.cpp project released build b9999, which introduces a new SME2 f32 kernel and enables dynamic scheduling for it. This update is part of the ongoing development of the library's core inference capabilities.

• Added kleidiai SME2 f32 kernel with dynamic scheduling support.
• macOS Apple Silicon KleidiAI integration remains disabled.
• Binaries are available for macOS (Apple Silicon, Intel), Linux (CPU, Vulkan, ROCm 7.2, OpenVINO, SYCL FP32/FP16), Android, Windows (CPU, OpenCL Adreno, CUDA 12/13, Vulkan, OpenVINO, SYCL, HIP), and openEuler.

The release provides updated binaries across multiple platforms and hardware accelerators for users requiring the latest build.