llama.cpp 发布: b11126

来源摘要

vulkan: add IQ4_XS MMQ/MMV matmul kernels (#28415) * vulkan: optimize IQ4_XS matmul kernels Assisted-by: OpenAI Codex * vulkan: address IQ4_XS review nits - drop the dead LOAD_VEC_A != 8 branch in the IQ4_XS shmem load; iq4_xs is in lut_load_vec_a()'s "8" list, so that path is never generated - disable MMVQ for IQ4_XS on Intel (27.3% tg regression on A770) - remove a stray empty line in types.glsl Assisted-By: Claude Opus 5 **Website:** - **Attestations:** - **macOS/iOS:** - [macOS Apple Silicon (arm64)](https://github.com/ggml-org/llama.cpp/releases/download/b11126/llama-b11126-bin-macos-arm64.tar.gz) - macOS Apple Silicon (arm64, KleidiAI enabled) [DISABLED](https://github.com/ggml-org/llama.cpp/pull/23780) - [macOS Intel (x64)](https://github.com/ggml-org/llama.cpp/releases/download/b11126/llama-b11126-bin-macos-x64.tar.gz) - [iOS XCFramework](https://github.com/ggml-org/llama.cpp/releases/download/b11126/llama-b11126-xcframework.zip) **Linux:** - [Ubuntu x64 (CPU)](https://github.com/ggml-org/llama.cpp/releases/download/b11126/llama-b11126-bin-ubuntu-x64.tar.gz) - [Ubuntu arm64 (CPU)](https://github.com/ggml-org/llama.cpp/releases/download/b11126/llama-b11126-bin-ubuntu-arm64.tar.gz) - [Ubuntu s390x (CPU)](https://github.com/ggml-org/llama.cpp/releases/download/b11126/llama-b11126-bin-ubuntu-s390x.tar.gz) - [Ubuntu x64 (Vulkan)](https://github.com/ggml-org/llama.cpp/releases/download/b11126/llama-b11126-bin-ubuntu-vulkan-x64.tar.gz) - [Ubuntu arm64 (Vulkan)](https://github.com/ggml-org/llama.cpp/releases/download/b11126/llama-b11126-bin-ubuntu-vulkan-arm64.tar.gz) - [Ubuntu x64 (CUDA 12)](https://github.com/ggml-org/llama.cpp/releases/download/b11126/llama-b11126-bin-ubuntu-cuda-12.8-x64.tar.gz) - [CUDA 12.8 libraries](https://github.com/ggml-org/llama.cpp/releases/