Re: [Swarm-Modelling] SWARM on Clusters

[Sunwoo Park]

Dear Marcus,
I really appreciate for your comments.

As my understanding, you claimed that message passing is not good for SWARM
mainly because communication overhead between agents might be nontrivial.
Right ? 
But, for example, 
(1) if models perform considerable amount of computation in a single
simulation phase (or cycle) with minimum communication between agents
(2) if the size and/or scalability of simulation models is much important
than performance (or execution time)
do you still think multi-threading approach is better than message passing
because of the nature of SWARM even in above cases ?

Also, could you give me some information (or link) on SWARM models/projects
that are already done or planned for 'large-scale' simulation  ?
For example, is there any SWARM project that uses more than 256 (or 512)
nodes in MPP machines ?

Sincerely,
Sunwoo

> From: "Marcus G. Daniels" <address@hidden>
> Reply-To: address@hidden
> Date: Thu, 29 Jan 2004 10:44:46 -0700
> To: address@hidden
> Subject: Re: [Swarm-Modelling] SWARM on Clusters
> 
> Sunwoo Park wrote:
> 
>> I just joined in this mailing list.
>> I have a simple question regarding SWARM software.
>> Is there any SWARM implementation that runs on cluster machines (or MPP
>> machines) based on Message Passing Paradigm (e.g., MPI) ?
>>  
>> 
> Swarm has a fine-grained knowledge of concurrency during a simulation.
> When multiple agents do something at the same timestep, Swarm knows
> this.   But that's just a little atom of the whole simulation execution
> sequence.  What this means is that in order for Swarm to exploit this
> knowledge on a parallel computer, it is necessary to be able to
> efficiently get that atom of computation to a physical processor.  A
> cluster, like a Beowulf arrangement of PCs, can't do this because the
> communication expense of getting the atom to the processor not amortized
> by the computation done.   A SMP or NUMA system can do this because the
> communication/overhead expense of getting the computation to the
> processor is small.  So if you have a two or four or eight way Opteron
> or Sun system or a big NUMA system like a SGI Altix, the interconnect
> between processors could reasonably slurp up these atoms and there would
> be a scalability win.
> 
> I think it would be hard to make a message passing system scale very
> well based on an architecture like Swarm.   You'd need low-latency
> interconnect, maybe Myrinet.
> 
> In any case, Swarm doesn't implement either.  A multithreaded Swarm
> would be feasible, but would assume a shared memory system like I mentioned.
> _______________________________________________
> Modelling mailing list
> address@hidden
> http://www.swarm.org/mailman/listinfo/modelling
>