llama.cpp b9295 fixes Vulkan SPIRV-Headers issue on Windows
The popular LLM inference engine now supports Vulkan on Windows with proper SPIRV-Headers.
The latest release of llama.cpp (b9295) is a maintenance update that addresses a critical Vulkan build issue on Windows. Specifically, the CMake find_package for SPIRV-Headers was failing due to platform-specific path handling. This fix ensures that Windows users can compile and run the Vulkan backend reliably, enabling GPU-accelerated LLM inference on a broader set of hardware.
The release also includes pre-built binaries for a wide range of platforms: macOS (Apple Silicon and Intel with KleidiAI), Linux (x64, arm64, s390x) with various backends (Vulkan, ROCm 7.2, OpenVINO, SYCL FP32/FP16), Windows (x64 and arm64 CPU, CUDA 12/13, Vulkan, SYCL, HIP), Android arm64, iOS XCFramework, and openEuler (with ACL Graph support). With 112k stars and 18.6k forks, llama.cpp remains the gold standard for running LLMs locally on consumer hardware.
- Fixed Windows CMake find_package for SPIRV-Headers in Vulkan backend
- Pre-built binaries for 20+ platform/backend combinations including ROCm, CUDA, and SYCL
- GitHub release with 112k stars and 18.6k forks, signed with verified GPG key
Why It Matters
Windows users can now reliably use Vulkan for local LLM inference, expanding GPU acceleration options beyond CUDA.