[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
RE: [Swarm-Support] memory problems
From: |
Steve Railsback |
Subject: |
RE: [Swarm-Support] memory problems |
Date: |
Wed, 4 Jun 2003 09:49:34 +0100 |
> -----Original Message-----
> From: address@hidden
> [mailto:address@hidden On Behalf Of Paul E Johnson
> Sent: Tuesday, June 03, 2003 7:36 PM
> To: address@hidden
> Subject: Re: [Swarm-Support] memory problems
>
>
> I agree that the drop on the EZDistribution object should
> give all the
> memory back, but I don't think most people have tried to use
> that class
> in this way and so, if there is a memory leak, we have not
> been alerted
> to it yet. I found a memory leak, which I still have not
> solved, in a
> method that used probes to retrieve double-valued variables,
> and if you
> are doing that, I expect you could see the same issue.
Yes, we have also found that instances of Swarm classes do not always
give everything back when you drop them- at one time Marcus provided an
explanation. Another person to explain it and the work-around we use is
Steve Jackson, but he's on vacation. I vaguely recall it might have
something to do with the zone you put an object in still retains some
information after you drop the object.
I agree with Paul that the obvious solution is to not drop the
EZDistribution- just repeat the [a update] statement each time step.
If you do not unnecessarily create and drop too much stuff, this memory
issue should never be a problem. (But always be careful you clean up
sufficiently when re-using an object.) However, if you have a big model
that runs for a long time, or you use an experiment manager to run the
model many times at once, you can have problems. But they can be solved
with a little more sophisticated memory management (using a class
"ZoneAllocMapper" which is on our software page at
http://math.humboldt.edu/EcoModeling - I can't explain it but Steve J.
can...so if anyone is interested, post another question next week when
he's back.
Steve Railsback