Re: [Swarm-Modelling] SWARM on Clusters

[Andy Cleary]

At 02:53 PM 1/29/2004 -0800, you wrote:
> Andy Cleary writes:
>  > You just can't do any reasonable level of parallelism without taking a 
> more
>  > explicit "partitioned/domain decomposed" approach, and for that a new
>  > generation of ABM infrastructure is needed...
>
> It's true that you have to [de]compose things at the domain level.
> But, I don't think it's true that you need a whole new infrastructure.
> It's more likely that all we need is a new piece of infrastructure.

:-) There's some subtlety between "a whole new infrastructure" and what I 
said. I am not necessarily saying that all existing infrastructure must be 
thrown away; a "new generation of ABM infrastructure" could very well be an 
evolution of existing infrastructure, and in fact more or less, that's what 
I envision. Many of the existing pieces could stay as they are and serve as 
the "sequential pieces", and yes some pieces need to be added; but I also 
suspect that some existing things need to be changed to "bridge" the old 
and the new.

I realize I'm speaking in broad swaths; I have more detailed notions in 
mind, but the project that was paying me to do this design work ceased 
doing so several weeks ago, and so I am forced to put it aside in an 
incomplete state until that day that a more "enlightened" funding agency 
resurrects that work. ;-) [Actually, the discussions over the last day or 
two about getting this kind of project funded come a day late and a dollar 
short for my use, though with my 20/20 hindsight, I'm in thorough 
agreement.][And the comment about national labs being a place where 
interesting but not scientifically meaningful projects can hideout was a 
LOL hoot!]


> Perhaps as we develop ABM patterns, some attempts could be made to
> describe this new piece so that we can begin building it.

This is the way forward IMHO. My design work was based on this, although 
apropos the national lab comment, since we don't actually have any ABM work 
at this lab I pretty much had to infer what the common ABM patterns are 
from very indirect sources and our own prototype infectious disease model...

Andy