Path: g2news1.google.com!news2.google.com!news.glorb.com!feeder.erje.net!eweka.nl!lightspeed.eweka.nl!zen.net.uk!hamilton.zen.co.uk!feed4.jnfs.ja.net!jnfs.ja.net!gemini.csx.cam.ac.uk!news.cam.ac.uk!not-for-mail
From: n...@cam.ac.uk
Newsgroups: sci.math,comp.lang.c,comp.lang.fortran
Subject: Re: 64-bit KISS RNGs
Date: Sun, 1 Mar 2009 09:45:43 +0000 (GMT)
Organization: University of Cambridge
Lines: 26
Message-ID: <godlg7$a4j$1@soup.linux.pwf.cam.ac.uk>
References: <d0d9069e-cfff-4520-a0fe-96715b25852d@j8g2000yql.googlegroups.com> <603fda23-428d-42f9-b4ad-869a30e8adae@40g2000prx.googlegroups.com> <godkjb$kn6$1@online.de>
NNTP-Posting-Host: soup.linux.pwf.cam.ac.uk

In article <godkjb$kn...@online.de>,
Phillip Helbig---remove CLOTHES to reply <hel...@astro.multiCLOTHESvax.de> wrote:
>In article
><603fda23-428d-42f9-b4ad-869a30e8a...@40g2000prx.googlegroups.com>,
>galathaea <galath...@gmail.com> writes: 
>
>RANLUX is slow, but at the highest "luxury level" all 24 bits of the 
>mantissa are chaotic.  So, one could just stick these together to create 
>numbers containing more bits.

That wasn't the issue she (I assume) was addressing - it was one that
I did.  Yes, that technique works, for both RANLUX and 32-bit KISS.
I use my own double-precision generator, of course, which has some
theoretical advantages over both and is marginally simpler than (and
similar to) KISS.

Galathaea's concern was about the period, and she is very right to
be so concerned.  While a long period does not guarantee pseudo-
randomness, it is a prerequisite for it - in particular, the pseudo-
random properties in N dimensions are often limited by the Nth root
of the period.  And, despite common belief, that is NOT solely true
for multiplicative congruential generators.


Regards,
Nick Maclaren.