llama.cpp b10001 fixes segfault, expands hardware support across platforms

ggml-org has rolled out llama.cpp b10001, a maintenance release that addresses a critical segfault in the test-export-graph-ops utility when invoked without arguments. This bug fix enhances stability for developers and users running local large language model inference, particularly those relying on automated testing or scripting. The release maintains llama.cpp's reputation as a robust, cross-platform library for running models like Llama, Mistral, and others on consumer hardware.

The standout feature of b10001 is its expanded platform support. Builds are now available for macOS (Apple Silicon with and without KleidiAI, Intel x64, iOS XCFramework), Linux (Ubuntu x64/arm64/s390x with CPU, Vulkan, ROCm 7.2, OpenVINO, SYCL FP32/FP16), Windows (x64/arm64 CPU, CUDA 12/13, Vulkan, OpenCL Adreno, OpenVINO, SYCL, HIP), and Android (arm64 CPU). This broad coverage enables enthusiasts and enterprises to deploy AI locally using their preferred GPU or CPU architecture. The release also includes new UI assets, signaling ongoing investment in the project's front-end. While b10001 is a minor version bump, it reinforces llama.cpp's position as the go-to tool for lightweight, high-performance LLM inference.