From chemistry-request- at -ccl.net Thu May 29 17:29:55 2003
Received: from facepe.dqf.ufpe.br (facepe.dqf.ufpe.br [150.161.5.2])
	by server.ccl.net (8.12.8/8.12.8) with ESMTP id h4TLTsgC019003
	for <chemistry$at$ccl.net>; Thu, 29 May 2003 17:29:55 -0400
Received: from dqf.ufpe.br ([150.161.5.90])
	by facepe.dqf.ufpe.br (8.12.6/8.12.6) with ESMTP id h4TJ0wfD019275
	for <chemistry$at$ccl.net>; Thu, 29 May 2003 16:00:58 -0300 (BRT)
	(envelope-from sidney$at$dqf.ufpe.br)
Message-ID: <3ED67F2E.9090506$at$dqf.ufpe.br>
Date: Thu, 29 May 2003 18:44:14 -0300
From: Sidney Ramos de Santana <sidney$at$dqf.ufpe.br>
User-Agent: Mozilla/5.0 (X11; U; Linux i686; en-US; rv:1.0.1) Gecko/20020830
X-Accept-Language: en-us, en
MIME-Version: 1.0
To: CCL List <chemistry$at$ccl.net>
Subject: Random Number Generator Fail ...
Content-Type: text/plain; charset=us-ascii; format=flowed
Content-Transfer-Encoding: 7bit

Dear CCL Users,

I compiled the Shell-Dynamo 2.1 with G77(Gnu Compiler)
and with IFC(Intel Fortran Compiler).
I have been trying to perform a calculation of diffusion constants
of Ions in a oxifluoride glass.
In this program I am using Buckigham potential and the initial
coordinates of the ions in the simulation box was been filling
randomicly, because I choosed the option "fillbox random".
This initial coordinates in the program compiled with G77
are generating by "ranf" intrinsic function, and in the IFC are
generating by "random_number" intrinsic function using
"date_and_time" intrinsic function to generate a seed.
In the both cases the initial configuration generated was the worst
possible. This way, I will need a very, very big step numbers in
this MD simulation.
So, this question for you:
What is the best to generate randomicly an initial
coordinates  of the ions the simulation box:
an intrinsic function of Compiler to generate random
number (if yes, why?) or an algorithm(if yes, why) ?

Sidney Ramos
Graduate Program of Materials Science
DQF - CCEN - UFPE