SSup: Problem size in Swarm

[Kerry Hanson]

I have a quick question for anyone who may have a feel for the practical
limits of the simulation problem size in Swarm.  I've been chatting with the
people over at Sandia about their Aspen economic simulation model.  They run
an agent based model of the national economy written in C on their Paragon
machine (a teraflops parallel computer).  Most of us, unfortunately, don't
have access to a Paragon down the hall.

What is the practical limit to the size of the simulation problem in Swarm
now, say, in number of agents or some other measure?  I realize this is a
crummy angels-on-the-head-of-a-pin kind of measure but I'm interested in
just order of magnitude. 

In addition to the current capability I understand that the development team
plans to have a parallel version of Swarm using the MPI protocol available
sometime before September.  Has there been any thought on how the practical
problem size would scale on networked workstations?


Kerry


                  ==================================
   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.
                  ==================================