randommw v1.0

randommw is a small, monolithic, 64-bit C library providing the basic functionality for random number generation in scientific computing applications, in particular the generation of sequences of random floating point numbers with uniform or Gaussian distributions. It is an integrated and curated collection of tried & tested code described in the literature. The library provides a choice of five modern and well-tested random number generators (RNGs): MWC8222, Lehmer64, PCG64DXSM (as used in Numpy), Xoshiro256+, and MELG19937 (a variant of the Mersenne Twister). There is a ziggurat algorithm, ZIGNOR, by J. A. Doornik (2005), for obtaining random floating-point numbers with a Gaussian distribution using these RNGs. The generators are amazingly fast, enabling, for instance, simulation of large numbers of Brownian particles with long trajectories.