Models & Labs
b9070 release adds Q4_0 MoE GEMM for Adreno
llama.cpp Releaseshigh confidence
Why it matters
- →Enhances performance for Adreno GPU users, crucial for mobile development.
- →Streamlines codebase with technical adjustments, improving efficiency.
- →Expands platform support, making the tool more accessible.
The latest b9070 release of llama.cpp brings Q4_0 MoE GEMM support for Adreno GPUs, enhancing performance for mobile developers using Qualcomm's graphics. This update, co-authored by Li He from Qualcomm, includes technical refinements like whitespace fixes and code cleanup. The release supports a wide range of platforms, including macOS, Linux, Windows, and Android, ensuring broad accessibility. This update focuses on optimizing existing capabilities rather than introducing new models.
Read originalMore from llama.cpp Releases
Open Sourcemodels
llama.cpp b9056 Release Expands Platform Support
The latest b9056 release of llama.cpp continues its trend of broadening platform compatibility, now including support for macOS Apple Silicon with KleidiAI enabled and a variety of Linux configurations such as Ubuntu with Vulkan and ROCm 7.2. This update also enhances Windows support with CUDA 12 and 13 DLLs, making it more versatile for developers working across different environments. While there are no groundbreaking new features, the release solidifies llama.cpp's position as a flexible inference runtime across diverse hardware setups. Developers can now leverage these updates to optimize performance on their specific systems, whether they're using Apple Silicon, AMD, or NVIDIA GPUs.
llama.cpp Releases
Open Sourcemodels
llama.cpp b9057 Release Expands Platform Support
The latest b9057 release of llama.cpp continues its trend of broadening platform compatibility, now optimizing for RISC-V CPUs with q1_0 dot support. This update enhances performance across a wide array of systems, including macOS, Linux, Windows, and Android, with specific builds for Apple Silicon, Vulkan, and CUDA environments. Notably, the inclusion of ROCm 7.2 for Ubuntu x64 and CUDA 13 for Windows x64 signifies a commitment to supporting diverse hardware configurations. While no new models are introduced, this release solidifies llama.cpp's position as a versatile inference runtime across multiple architectures.
