An implementation of Takuji Nishimura's and Makoto Matsumoto's Mersenne Twister pseudo random number generator in Go.

Copyright (C) 2013 Jochen Voss

This program is free software: you can redistribute it and/or modify it under the terms of the GNU General Public License as published by the Free Software Foundation, either version 3 of the License, or (at your option) any later version.

Please send any comments or bug reports to the program's author, Jochen Voss <[email protected]>.

The Mersenne Twister is a pseudo random number generator (PRNG), developed by Takuji Nishimura and Makoto Matsumoto. The Mersenne Twister is, for example, commonly used in Monte Carlo simulations and is the default random number generator for many programming languages, e.g. Python and R. This package implements the 64bit version of the algorithm.

This package can be installed using the go get command:

go get github.com/seehuhn/mt19937

Detailed usage instructions are available via the package's online help, either on pkg.go.dev or on the command line:

go doc github.com/seehuhn/mt19937

The class MT19937 represents instances of the Mersenne Twister. New instances can be allocated using the mt19937.New() function. A seed can be set using the .Seed() or .SeedFromSlice() methods. MT19937 implements the rand.Source interface from the math/rand package. Typically the PRNG is wrapped in a rand.Rand object as in the following example:

rng := rand.New(mt19937.New())
rng.Seed(time.Now().UnixNano())

Note that MT19937 is not safe for concurrent accesss by different goroutines. If more than one goroutine accesses the PRNG, the callers must synchronise access using sync.Mutex or similar.

Go has a built-in PRNG provided by the math/rand package. I did not find any information about this built-in PRNG except for a comment in the source code which says "algorithm by DP Mitchell and JA Reeds". In contrast, the MT19737 generator provided in this package is a well-understood random number generator. Relevant references include [Ni2000] and [MatNi1998].

The unit tests for the mt19937 package verify that the output of the Go implementation coincides with the output of the reference implementation.

The mt19937 generator is slightly slower than the Go default PRNG. A speed comparison can be performed using the following command:

go test -bench=. github.com/seehuhn/mt19937

On my laptop, using go version 1.15.5, I get the following results:

method	time per call	thoughput
MT19937.Uint64	5.63 ns/op	1422.14 MB/s
MT19937.Int63	5.69 ns/op	1405.04 MB/s
builtin Uint64	4.05 ns/op	1973.47 MB/s
builtin Int63	4.15 ns/op	1929.62 MB/s

This shows that, on my system, a call to the Int63() method of the built-in PRNG takes about 73% of the time that MT19937.Int63() takes.