A family of enhanced Lehmer random number generators, with hyperplane suppression, and direct support for certain physical applications
Over two hundred congruential pseudorandom number generators, each with a different multiplier, are built into a single assembler routine that returns 32-bit integer and floating-point values. This gives a Monte Carlo user the opportunity of selecting a combination of sequences, so as to provide a g...
Gespeichert in:
Veröffentlicht in: | Computer physics communications 1997-12, Vol.107 (1), p.258-280 |
---|---|
Hauptverfasser: | , |
Format: | Artikel |
Sprache: | eng |
Schlagworte: | |
Online-Zugang: | Volltext |
Tags: |
Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
|
Zusammenfassung: | Over two hundred congruential pseudorandom number generators, each with a different multiplier, are built into a single assembler routine that returns 32-bit integer and floating-point values. This gives a Monte Carlo user the opportunity of selecting a combination of sequences, so as to provide a greater appearance of chaos. The software makes use of extended 64-bit arithmetic on Intel 386/387 (or higher) chips, thus attaining a period of 2
62 for each of the individual generators. The routine also features entry points that more directly support certain applications, such as well logging in nuclear geophysics. In addition to the customary uniform (0,1) “white noise” generator, the package provides values distributed according to the exponential and Gaussian distributions, random unit vectors in two and three dimensions, as well as Klein-Nishina and neutron scattering distributions. |
---|---|
ISSN: | 0010-4655 1879-2944 |
DOI: | 10.1016/S0010-4655(97)00101-X |