gpujs/gpu.js

Health v2 · maintenance only

Adoption not part of the score

15370 stars · 661 forks observed · 2026-08-28

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

GPU.js is a JavaScript acceleration library for GPGPU that automatically transpiles JavaScript functions into shader language (GLSL) and compiles them to run on the GPU, with automatic CPU fallback when no GPU is available. It works in both the browser (via WebGL) and Node.js, often delivering 1-15x speedups for parallelizable computations like matrix math.

Use cases

• accelerate matrix multiplication in javascript
• run parallel computations on the gpu from the browser
• speed up heavy math loops in node.js
• transpile javascript functions to webgl shaders
• add gpu acceleration with cpu fallback
• compute large 2d arrays faster in javascript

When to choose

• you need GPU acceleration for parallelizable numeric JavaScript code without writing shaders
• you want a single codebase that runs in browser and Node.js with automatic CPU fallback
• your workload involves matrix math, image processing, or large array computations

When to avoid

• your code is inherently sequential or branch-heavy, where GPU gains are minimal
• you need WebGPU or CUDA-specific features rather than WebGL-based kernels
• you need a full machine learning framework with autodiff and model training

Facets

library · maturity active

gpu-computing math compiler machine-learning gpu-computing web-development developer-tools performance browser cross-platform gpgpu webgl glsl shader-transpilation cpu-fallback parallel-computing algorithms nodejs javascript web

Member repositories

For agents

markdown · JSON · MCP: product_card(name="gpujs/gpu.js")