Bob Jenkins' small noncryptographic PRNG (pseudorandom number generator) ported to JavaScript.

Why not Math.random?

Sometimes you want repeatable results. With this lib, you can save the seed you used, and later initialize the PRNG with the same seed and get the same sequence of pseudorandom numbers.

Also, this PRNG beats the one built into V8 on ent's chi-square test, indicating that its output is, in a sense, "more random". (The results on ent's other tests are excellent for both.)

Is it fast?

Both BurtlePRNG and the native PRNG are incredibly fast and unlikely to be a bottleneck in your application.

You can use test.js, which prints 10MB of random bytes to the standard output using Node, to run your own benchmarks. On Node 0.10.28, there is no significant performance difference between generating 32-bit unsigned integers using test.js versus equivalent code using Math.random:

((1<<31) * 2 * Math.random()) >>> 0

Usage

If you're using Node.js, Browserify or Webpack, use the usual npm install burtleprng and require:

var BurtlePRNG = require('burtleprng');

Otherwise, simply include BurtlePRNG.js in a script tag. It will export a single class, BurtlePRNG.

API

new BurtlePRNG(seed)

Creates and returns a new PRNG instance. The seed argument is mandatory and should be an unsigned 32-bit integer. Reasonable ways to come up with a seed include:

Date.now() & 0x7fffffff

// or

((1<<31) * 2 * Math.random()) >>> 0

prng.next()

Returns a pseudorandom, unsigned 32-bit integer and advances the PRNG state.

prng.float()

Returns a pseudorandom floating-point number on the interval [0, 1), similar to Math.random(), and advances the PRNG state. Equivalent to prng.next() / 4294967296.0.

License

Bob Jenkins released his original C implementation of this PRNG as public domain. Similarly, burtleprng is released under the Creative Commons CC0 Public Domain Dedication. See COPYING.txt for details.