Newsgroups: comp.parallel
From: aluru@top.cis.syr.edu (Srinivas Aluru)
Subject: Re: parallel pseudo random number generator
Organization: Syracuse University CIS Dept.
Date: 10 Apr 1995 22:17:58 GMT
Message-ID: <3mje0n$srg@usenet.srv.cis.pitt.edu>

In article <3m95kj$kln@usenet.srv.cis.pitt.edu>, koot@sara.nl (J.J.A. Koot) writes:
|> May this is a FAQ, may be other news groups are better suited for this
|> question. If you think so, please mail me personally (koot@sara.nl).
|> I am looking for a pseudo random number generator for use in a
|> parallel program using message passing. Efficiency requires that each
|> task has its own generator and that this generator can produce most of
|> the random numbers without communication with other tasks. The
|> sequences produced in different tasks must be statistically
|> independent from each other.  I have asked this question earlier, but
|> received no answer. 
[rest of previous article deleted]

I think I have what you are looking for:

1) Generator with period 2^607
2) Guarantee that the sequences will never overlap
3) No communication at all required
4) Perfectly scalable to any number of processors (restriction: must be 
   powers of 2)
5) A start up time that actually decreases when you increase the number of 
   processors
6) Each processor generating numbers at the same speed as the corresponding 
   sequential generator
7) Architecture independent code available (easy since no communication)

If you are interested, please read the article:

S. Aluru et. al, A Random Number Generator for Parallel Computers, Parallel 
Computing (Journal), Vol. 18, (1992) pp 839-847.

If you like the generator after reading the article, you can get the code from 
me by email.

-Srinivas

