llama.cpp b9846 release with Vulkan matmul optimization for Asahi Linux

The llama.cpp project has released version b9846, which includes a Vulkan backend optimization for Asahi Linux. This update rolls back the block size loop in matrix multiplication to improve compatibility and performance on Apple Silicon hardware running Linux.

• Vulkan: Rolls back the BK loop in matmul for Asahi Linux via direct SPIR-V editing.
• Vulkan: Reverts previous BK-loop unroll changes and fixes inline comments.
• macOS/iOS: Provides binaries for Apple Silicon (arm64) and Intel (x64), with KleidiAI disabled.
• Linux: Offers builds for Ubuntu x64, arm64, and s390x across CPU, Vulkan, ROCm 7.2, OpenVINO, and SYCL backends.
• Windows: Includes CPU, OpenCL Adreno, CUDA 12/13, Vulkan, OpenVINO, SYCL, and HIP builds for x64 and arm64.
• Android: Releases an arm64 CPU binary.

This release ensures stable Vulkan support on Asahi Linux while providing comprehensive pre-built binaries for a wide range of hardware architectures and acceleration libraries.