AnswerDotAI/gpu.cpp

Health v2 · maintenance only

Adoption not part of the score

3984 stars · 199 forks observed · 2026-08-28

What it is AI-extracted, prompt v1, taxonomy v1, 2026-08-29, confidence not recorded

A lightweight, header-only C++20 library for portable low-level GPU compute built on Dawn's WebGPU implementation. It provides a minimal API for creating tensors, kernels, and dispatching compute shaders across Metal, Vulkan, and browsers.

Use cases

• run gpu compute kernels in c++
• portable webgpu compute across metal and vulkan
• write wgsl compute shaders with a thin c++ wrapper
• run gpu workloads in the browser via webassembly
• learn low-level gpu programming without heavy frameworks

When to choose

• you want a small, close-to-WebGPU C++ compute library with minimal abstraction
• you need the same GPU compute code to run natively and in browsers
• you prefer explicit WGSL shaders and control over resources

When to avoid

• you need high-level tensor operations or automatic differentiation like PyTorch or JAX
• you want graphics rendering rather than compute
• you need CUDA-specific features or NVIDIA-only optimizations

Facets

library · maturity active

gpu-computing developer-tools gpu-computing cross-platform developer-tools cpp browser wasm webgpu wgsl header-only dawn compute-shaders gpgpu macos linux

Member repositories

For agents

markdown · JSON · MCP: product_card(name="AnswerDotAI/gpu.cpp")