FastNoise

FastNoise is an open source noise generation library with a large collection of different noise algorithms. This library has been designed for realtime usage from the ground up, so has been optimised for speed without sacrificing noise quality.

This project started when my search to find a good noise library for procedural terrain generation concluded without an obvious choice. I enjoyed the options and customisation of Accidental Noise Library and the speed of LibNoise, so many of the techniques from these libraries and the knowledge I gained from reading through their source has gone into creating FastNoise.

I have now also created FastNoise SIMD, which utilises SIMD CPU instructions to gain huge performance boosts. It is slightly less flexible and cannot be converted to other languages, but if you can I would highly suggest using this for heavy noise generation loads.

Features

  • Value Noise 2D, 3D
  • Perlin Noise 2D, 3D
  • Simplex Noise 2D, 3D, 4D
  • Cubic Noise 2D, 3D
  • Gradient Perturb 2D, 3D
  • Multiple fractal options for all of the above
  • Cellular (Voronoi) Noise 2D, 3D
  • White Noise 2D, 3D, 4D
  • Supports floats or doubles

FastNoise Preview

Performance Comparisons

Using default noise settings on FastNoise and matching those settings across the other libraries where possible.

Timings below are x1000 ns to generate 32x32x32 points of noise on a single thread.

  • CPU: Intel Xeon Skylake @ 2.0Ghz
  • Compiler: Intel 17.0 x64
Noise TypeFastNoiseFastNoiseSIMD AVX2LibNoiseFastNoise 2D
White Noise14113111
Value635160364
Perlin9643421409476
Simplex1189340875
Cellular29331472569601074
Cubic29331393872

Comparision of fractal performance here.

Examples

Cellular Noise

Cellular Noise

Cellular Noise

Cellular Noise

Fractal Noise

Fractal Noise

Value Noise

Value Noise

White Noise

White Noise

Gradient Perturb

Gradient Perturb

Gradient Perturb

Gradient Perturb

Git Repository

More info available on the GitHub readme
Code released under MIT license

https://github.com/Auburns/FastNoise

You may also like...

Leave a Reply