Computer Simulation Question
czarmr
Does anybody know where I can get Arena 3.5 Simulation questions answered?
For example:
ERROR:
Entity: 3933
A runtime error was detected at time 607.612 at the following block:
* 4054 6357$ EXIT:SpecialHandlingConv:NEXT(2106$);
No cells are allocated to this entity.
0183:CNVOWN
Thank you
-Bodnar,B.L.
In article <8bjjto$e8d$1...@leopard.it.wsu.edu>,
If your maintenance contract is up-to-date, try Systems Modeling Corp -- the
manufacturer of ARENA. I use ARENA 3.5 (actually, I program the simulations
in the command language -- SIMAN V) for packet switching modeling. I ran
across a weird problem like this in ARENA 1.2 about two years ago -- after
several weeks of troubleshooting, another engineer and I figured out the
problem was in the memory manager. An upgrade took care of these problems.
The upgrade to ARENA 3.5 is ARENA 4.0. I dislike it since it locks onto some
disk serial number instead of the dongle (consider what would happen if you
(1) have an urgent modeling project and (2) have your computer crash... how
many days will take you to restart the work???).
If you're running this under Windows 98, be sure that your operating
environment is "clean." I've had interesting problems (e.g., probability
calculations exceeding unity) if W98 had too many programs operating
concurrently.
Finally, you can do what I do: activate tracing in your .exp file, dump the
trace file, and look through it. If it's a programming problem, you'll
eventually track it down.
Drop me a line if you think I can be of further assistance.
Regards,
Bohdan Bodnar
bbo...@lucent.com
ne...@unitz.on.ca
Try their website at www.sm.com
Neil
Sent via Deja.com http://www.deja.com/
Before you buy.
From s...@eecs.wsu.edu Mon Mar 27 08:55:40 2000 -0700
Status: R
X-Status:
X-Keywords:
Received: from moderated.news.pipex.net (moderated.news.pipex.net [158.43.192.79]) by thalia.eecs.wsu.edu with SMTP (8.9.3/)
id IAA27350 for <s...@eecs.wsu.edu>; Mon, 27 Mar 2000 08:55:39 -0800 (PST)
Received: (qmail 15793 invoked by alias); 27 Mar 2000 16:55:37 -0000
Received: (qmail 15789 invoked from network); 27 Mar 2000 16:55:36 -0000
Received: from mail.deja.com (205.238.143.34)
by moderated.news.pipex.net with SMTP; 27 Mar 2000 16:55:36 -0000
Received: from nnrp1.deja.com (nnrp1.deja.com [10.2.2.16])
by mail.deja.com (8.8.7/8.8.5) with ESMTP id KAA31501
for <comp-si...@moderators.isc.org>; Mon, 27 Mar 2000 10:55:29 -0600
Received: (news@localhost) by nnrp1.deja.com (8.7.6/8.6.12) id KAA11622; Mon, 27 Mar 2000 10:55:17 -0600
To: comp-si...@moderators.isc.org
Path: not-for-mail
From: ne...@unitz.on.ca
Newsgroups: comp.simulation
Subject: Re: Comparison
Date: Mon, 27 Mar 2000 16:55:27 GMT
Organization: Deja.com - Before you buy.
Lines: 15
Message-ID: <8bo3pl$bb5$1...@nnrp1.deja.com>
References: <8aman9$er7$1...@leopard.it.wsu.edu> <8bjjs8$pki$1...@leopard.it.wsu.edu>
NNTP-Posting-Host: 207.236.97.125
X-Article-Creation-Date: Mon Mar 27 16:55:27 2000 GMT
X-Http-User-Agent: Mozilla/4.0 (compatible; MSIE 4.01; Windows NT)
X-Http-Proxy: 1.0 INCOPROXYSUDS1, 1.0 x42.deja.com:80 (Squid/1.1.22) for client 207.236.97.125
X-MyDeja-Info: XMYDJUIDrunci
You will find an annual simulation software review in the Industrial
Engineering Solutions publication. I think that it is usually published
in May/June. This review has very good comparison information about
most discrete-event simulation software.
Also many of the simulation software vendore have demo versions
available on CD for you to have a look at. You just have to ask them.
Hope this is of some help.
Neil
Sent via Deja.com http://www.deja.com/
Before you buy.
Jim Henriksen
>
> If you're running this under Windows 98, be sure that your operating
> environment is "clean." I've had interesting problems (e.g., probability
> calculations exceeding unity) if W98 had too many programs operating
> concurrently.
>
I am incredulous. The chance of other programs affecting probability
calculations within your simulation application is (or ought to be)
miniscule. Did you reach this conclusion on your own, or was this something
the vendor told you? In etither case, what evidence lead to the conclusion.
This sounds like another computing myth to me. The problem is almost
certainly in your program. Referencing unitialized variables is one way to
experience unintentionally random behavior. Might this have been the case?
Jim Henriksen
-Bodnar,B.L.
How about "personal experience" or "school of hard knocks"??? Rebooting the OS
eliminated the problem. The simulation also ran flawlessly under W95.
Incidently, the failure was on the third (independent) iteration of a
simulation which ran, otherwise, flawlessly. Multiple simulations were
running concurrently in multiple DOS windows with a bunch of other stuff
running in the background.
I've also had simulations fail on W95-based machines before memory-leak fixes
were put into the kernel. No, the problem is certainly *NOT* in my coding.
Try running W98, opening and shutting down programs, and then inform me just
how stable the OS is...
Dr. Bohdan Bodnar
Lucent Technologies, Inc.
Jim Henriksen
> How about "personal experience" or "school of hard knocks"??? Rebooting
the OS
> eliminated the problem. The simulation also ran flawlessly under W95.
> Incidently, the failure was on the third (independent) iteration of a
> simulation which ran, otherwise, flawlessly. Multiple simulations were
> running concurrently in multiple DOS windows with a bunch of other stuff
> running in the background.
>
> I've also had simulations fail on W95-based machines before memory-leak
fixes
> were put into the kernel. No, the problem is certainly *NOT* in my
coding.
>
> Try running W98, opening and shutting down programs, and then inform me
just
> how stable the OS is...
>
> Dr. Bohdan Bodnar
> Lucent Technologies, Inc.
>
>
Dear Dr. Bodnar:
These kinds of errors can be expremely difficult, and perhaps even
impossible to track down. However, what you've said can be paraphrased as
"I can't find the error, so it must be a Windows problem." I'm no great fan
of Microsoft's software, believe me, but in this case, I feel the
probability of user error is greater than that of operating system error.
DOS boxes under Windows run in independent virtual machines. Simulation
runs tend to place great demands for CPU time, and perhaps memory, but in
most other ways place relatively modest demands on an operating system. The
chances of multiple independent DOS tasks of this type interfering with
another is low.
You still haven't answered my earlier question as to how you reached your
conclusion. Did you talk to the vendor of the simulation software you're
using? What sorts of memory allocation strategies does the software use?
Is dynamically acquired memory zeroed before use? Is a "grab and never give
back" approach used for requests for dynamic memory?
My bet still is that you have uninitialized data values (probably pointers)
that are causing this problem. When you run multiple simulations at the
same time, the probability of taking a hit due to use of a bad value
increases.
Regards,
Jim Henriksen