randn benchmarks

[David Bateman]

The fact is that R has already implemented ziggurat in its SuppDists package,

SuppDists: Supplementary distributions
    Ten distributions supplementing those built into R. Inverse Gauss,
    Kruskal-Wallis, Kendall's Tau, Friedman's chi squared, Spearman's
    rho, maximum F ratio, the Pearson product moment correlation
    coefficiant, Johnson distributions, normal scores and generalized
    hypergeometric distributions. In addition two random number
    generators of George Marsaglia are included.

    Version:    1.0-9
    Date:       2004/1/7
    Author:     Bob Wheeler <address@hidden>
    Maintainer: Bob Wheeler <address@hidden>
    License:    GPL vesion 2 or newer
    URL:        http://www.bobwheeler.com/stat

Though a better URL seems to be

http://www.stat.ucl.ac.be/ISdidactique/Rhelp/library/SuppDists/html/ziggurat.html

Note its under a GPL license, and it is based on the article you quote. So I
really don't see that there is a problem with us using this code. I attach
an example implementation, which is faster than matlabs randn :-)

Cheers
David

-- 
David Bateman                                address@hidden
Motorola CRM                                 +33 1 69 35 48 04 (Ph) 
Parc Les Algorithmes, Commune de St Aubin    +33 1 69 35 77 01 (Fax) 
91193 Gif-Sur-Yvette FRANCE

The information contained in this communication has been classified as: 

[x] General Business Information 
[ ] Motorola Internal Use Only 
[ ] Motorola Confidential Proprietary

zigurat.cc
Description: Text document