The llama.cpp project has released version b10018, introducing a SYCL implementation of the xielu operator via pull request #25550. This update provides pre-built binaries for macOS, Linux, Android, Windows, and openEuler across various hardware backends.

  • macOS builds include Apple Silicon (arm64) and Intel (x64) versions, alongside an iOS XCFramework; KleidiAI support on Apple Silicon is disabled.
  • Linux offerings cover CPU (x64, arm64, s390x), Vulkan, ROCm 7.2, OpenVINO, and SYCL (FP32/FP16) variants for Ubuntu.
  • Windows binaries support CPU (x64/arm64), OpenCL Adreno, CUDA 12.4/13.3, Vulkan, OpenVINO, SYCL, and HIP.
  • Android arm64 (CPU) and openEuler x86/aarch64 (310p/910b ACL Graph) builds are also available.

This release expands hardware compatibility for llama.cpp users, particularly with the new SYCL operator support and updated CUDA versions.