Re: [Swarm-Modelling] ABMs on Graphical Processor Units

[Russell Standish]

On Thu, Dec 27, 2007 at 08:01:44PM -0700, Marcus G. Daniels wrote:
> Russell Standish wrote:
> >I think the most useful outcome of this work would be a library of
> >data structures and algorithms with a C language interface (with a JNI
> >Java interface a relatively trivial proposition on top of that) that
> >provides useful building blocks for ABMs. 
> What blocks would be useful?
> 
> How about a library to manage agents and messages?   One might imagine 
> objects being handed out by the library and all messages would be sent 
> via remote procedure calls.  In this way, the logical model could be 
> mapped to physical resources much as a queuing system load balances 
> jobs.   In particular, messaging frequency could be used to migrate 
> agents to nearby memory.  Since all the objects would be tracked, there 
> would also be the possibility of using smart pointers / GC in the target 
> language(s).
> 
> Marcus

The blocks need to be implementable in SIMD hardware, which pretty
much rules out the sort of higher level functionality you're talking
about. The sorts of things I imagine might be useful are
implementations of various sorts of spaces of simple data types
(reals, ints etc), with different dimensionalities, neighbourhoods and
update rules. Other possibilities include artificial neural networks,
or fuzzy logic inference engines that could be used to implement agent
behaviour.

Cheers

-- 

----------------------------------------------------------------------------
A/Prof Russell Standish                  Phone 0425 253119 (mobile)
Mathematics                              
UNSW SYDNEY 2052                         address@hidden
Australia                                http://www.hpcoders.com.au
----------------------------------------------------------------------------