Dynamic schedules and execution speed

[Roger M. Burkhart]

Rob Kewley writes:

> My model uses a dynamic schedule, and I iterate that model many times.
> My current symptom is that the 1st iteration of the model takes 3
> seconds and the 30th iteration takes 20 seconds.  It just gets slower
> from then on out.  I create no new entities for later iterations.  I
> simply reset all objects to their initial state and execute again.  I
> don't believe this is caused by a memory leak because I created my own
> concurrent group type as discussed in the "Bug in AutoDrop" postings to
> this list arount 25 June.  This slow speed may be caused by the fact
> that I terminate my dynamic schedule, remove all old action from it, and
> reactivate it again in the model swarm for the next iteration.

This sounds like horribly slow speed and I know of nothing in the schedule
machinery that could produce such slowdown (even if Autodrop weren't
working), but that's not to say there's not something somewhere.  Please
send me your template code example so I can look at that example for a
start.  Unfortunately, I'm currently traveling, so I can't say how fast
a response I can give myself, but I'll at least try to look at the code.
Dynamic schedules are at the heart of what Swarm is about and while there
can be some penalty relative to a static one, it should only be a relatively
minor percentage.  Can you give a little idea how many actions of what kind
you are scheduling, and removing, from your dynamic schedule on each
iteration?

--Roger


                  ==================================
   Swarm-Support is for discussion of the technical details of the day
   to day usage of Swarm.  For list administration needs (esp.
   [un]subscribing), please send a message to <address@hidden>
   with "help" in the body of the message.
                  ==================================