Newsgroups: comp.theory,comp.ai.genetic,comp.ai
Path: cantaloupe.srv.cs.cmu.edu!das-news2.harvard.edu!news2.near.net!news.mathworks.com!hookup!news.moneng.mei.com!uwm.edu!math.ohio-state.edu!jussieu.fr!univ-lyon1.fr!swidir.switch.ch!news.unige.ch!usenet
From: sylvere@divsun.unige.ch (Silvere Martin-Michiellot)
Subject: Re: wanted: FAST binary random number gene
Message-ID: <1995Mar16.135500.22980@news.unige.ch>
Sender: usenet@news.unige.ch
Reply-To: sylvere@divsun.unige.ch
Organization: University of Geneva, Switzerland
References: <3k6tp7$t13@vixen.cso.uiuc.edu>
Date: Thu, 16 Mar 1995 13:55:00 GMT
Lines: 34
Xref: glinda.oz.cs.cmu.edu comp.theory:11980 comp.ai.genetic:5280 comp.ai:28181

In article t13@vixen.cso.uiuc.edu, dputnam@prairienet.org (Dan Putnam) writes:
>Byung-Ro Moon (moon@maui.cs.ucla.edu) wrote:
>: I'm looking for a very FAST and reasonable binary random number generator.
>: As I need fairly many calls for this, random()%2 (or random() mod 2)
>: takes too much time as the Unix routine random() or rand() are not so cheap. 
>
>: Any help will be appreciated.
>
>: -- BR 
>
>Random number generation algorithms tend to be multi-bit creatures.
>
>How about saving a call to random() in a variable and accessing
>the bits sequentially?  You should probably check to see if the
>random number generator is really giving you independently distributed
>bits.

Just an idea...
On thing that can be done is to read sequentially the bits (or even bytes)
of the memory of your computer (the RAM, boys).
Of course, you have to choose addresses of programs, not data, and, then,
there is a great probability that your average random will be .5 (or 127.5 for
random bytes).

Call me I you think about something faster :-)


-----------------

"Is anyone alive down there ?"

Silvere MARTIN-MICHIELLOT


