Vulkan
A cross-platform, low-level GPU graphics and compute API developed by the Khronos Group. It provides direct control over GPU hardware with low CPU overhead, high parallelism, and cross-platform portability, and is widely used in gaming, graphics rendering, and GPU computing.
Core Facts
GPUI uses Metal (macOS) and Vulkan (cross-platform) as rendering backends.
llama.cpp supports GPU acceleration backends including Metal, CUDA, Vulkan, and SYCL.
Proton is a compatibility tool deeply customized by Valve based on the Wine project, used to run Windows games on Linux.
Timeline (last 90 days)
Vulkan is a cross-vendor graphics and compute API; through its compute extensions, inference engines can run on AMD integrated GPUs, offering broader compatibility but slightly lower peak performance than ROCm.
This patch only targets the Metal backend and does not constitute a general solution for similar library compilation errors on other platforms (such as the CUDA and Vulkan backends).
Video2X's AI engine runs on NCNN and Vulkan, making hardware requirements relatively friendly—even a ten-year-old graphics card can run it.
When mapping GameCube's GX graphics instructions to modern graphics APIs (Vulkan, Metal), AI can assist in generating shader conversion code and state management logic, but developers still ultimately need to verify the correctness of rendering results.
After the update, computers without an NVIDIA GPU or ROCm (such as Snapdragon laptops and Linux ARM boards) can run GGUF models on the GPU via Vulkan.
Pipit claims acceleration via local Vulkan GPU/NPU delegation, with search latency kept under 50 milliseconds.
MoltenVK is an open-source compatibility layer that translates and maps the Vulkan API onto Apple's Metal, part of the Vulkan portability implementations in the Khronos ecosystem; it does not translate OpenGL directly into Metal, but rather lets applications and engines originally targeting Vulkan run on macOS and iOS.
For legacy projects still using OpenGL, a common approach is to first convert OpenGL calls to Vulkan via solutions such as Zink and then land on Metal through MoltenVK, forming a multi-stage translation chain; this chain introduces extra CPU overhead and potential compatibility deviations, and debugging difficulty rises markedly, so it is better suited as a transitional measure than a long-term architecture.
The current version of Ollama is v0.40.3, and the official images support NVIDIA CUDA, AMD ROCm and the built-in Vulkan runtime.
BLAS is a standard interface for a set of basic linear algebra routines, and matrix multiplication is the computational core of LLM inference; llama.cpp can be compiled and linked against different acceleration implementations: NVIDIA GPUs use CUDA, while AMD, Intel, or machines without a discrete GPU use Vulkan or CPU-side OpenBLAS, etc.
40 more timeline events
All Facts (20)
GPUI uses Metal (macOS) and Vulkan (cross-platform) as rendering backends.
80%Verifiedllama.cpp supports GPU acceleration backends including Metal, CUDA, Vulkan, and SYCL.
80%VerifiedProton is a compatibility tool deeply customized by Valve based on the Wine project, used to run Windows games on Linux.
70%Verifiedllama.cpp supports GPU acceleration via CUDA for NVIDIA GPUs, ROCm for AMD GPUs, and Metal for Apple Silicon, and also supports pure CPU inference.
70%VerifiedThe graphics APIs used by the PS5 are named GNM (low-level) and GNMX (high-level wrapper), which are Sony proprietary interfaces with no publicly available specification documents.
65%Verifiedllama.cpp has evolved into a general-purpose local inference framework covering multiple backends including CPU, CUDA, Metal, and Vulkan.
65%VerifiedVulkan is a cross-platform graphics and compute API used by projects such as llama.cpp as a general-purpose GPU inference backend, with the advantages of broad compatibility, simple deployment, and almost no dependency on vendor-specific driver stacks.
65%UnverifiedGodot引擎本身使用C++编写，核心代码约150万行，支持Vulkan、OpenGL ES 3.0和兼容性渲染器三种渲染后端
80%UnverifiedWhen the ROCm version cannot run, the system falls back to a Vulkan sd.cpp build.
70%UnverifiedSwiftShader is a high-performance CPU renderer developed by Google, supporting Vulkan and WebGPU.
70%UnverifiedVulkan is a cross-platform graphics and compute API maintained by the Khronos Group, supporting systems including Windows, Linux, and Android.
70%UnverifiedWebGPU introduces compute shader capabilities comparable to Vulkan/Metal/D3D12, enabling general-purpose GPU computing (GPGPU).
60%Unverifiedllama.cpp was introduced in early 2023 to run LLaMA models using only a CPU, and subsequently added support for hardware acceleration backends such as Metal, CUDA, and ROCm.
60%UnverifiedVulkan is a cross-vendor graphics and compute API; through its compute extensions, inference engines can run on AMD integrated GPUs, offering broader compatibility but slightly lower peak performance than ROCm.
50%UnverifiedThis patch only targets the Metal backend and does not constitute a general solution for similar library compilation errors on other platforms (such as the CUDA and Vulkan backends).
50%UnverifiedVideo2X's AI engine runs on NCNN and Vulkan, making hardware requirements relatively friendly—even a ten-year-old graphics card can run it.
50%UnverifiedWhen mapping GameCube's GX graphics instructions to modern graphics APIs (Vulkan, Metal), AI can assist in generating shader conversion code and state management logic, but developers still ultimately need to verify the correctness of rendering results.
50%UnverifiedAfter the update, computers without an NVIDIA GPU or ROCm (such as Snapdragon laptops and Linux ARM boards) can run GGUF models on the GPU via Vulkan.
50%UnverifiedPipit claims acceleration via local Vulkan GPU/NPU delegation, with search latency kept under 50 milliseconds.
50%UnverifiedMoltenVK is an open-source compatibility layer that translates and maps the Vulkan API onto Apple's Metal, part of the Vulkan portability implementations in the Khronos ecosystem; it does not translate OpenGL directly into Metal, but rather lets applications and engines originally targeting Vulkan run on macOS and iOS.
50%