Re: [Swarmfest2006] Keynote speakers -> use of swarm funds

[Rick Riolo]

here are a couple of comments about marcu's suggestion:

we (cscs) recently tried to become an google soc of mentor organization,
but they said they had as many orgs as they could handle this year.
but maybe in future years.

or there is an item in there about how a student can try
to name their own mentor organization, because none of the ones
that exist are quite right. (its in the faq somewhere).
so we could try that.

as it turns out, an ex-student here at UM (Ed Baskerville)
would like to do something with Swarm or something related
to ABM tools.  (He has contributed to the swarm lists in the
past---he is a big mac os x user.)
he is going to try that student-requests-specific-new-mentor
approach to do some project htis summer, probably making
a kind of drone-like sim runner that uses modern stuff like
grid computing to distribute jobs.

i don't know anything about the approaches marcus mentions below.
building a new simulator sounds like more than a one summer project, tho.
so it might take coming up with a plan that breaks
the task up into summer-doable chunks.

- r

Rick Riolo                           address@hidden
Center for the Study of Complex Systems (CSCS)
4477 Randall Lab
University of Michigan         Ann Arbor MI 48109-1120
Phone: 734 763 3323                  Fax: 734 763 9267
http://cscs.umich.edu/~rlr

On Thu, 20 Apr 2006, Marcus G. Daniels wrote:

> Date: Thu, 20 Apr 2006 10:28:00 -0600
> From: Marcus G. Daniels <address@hidden>
> Reply-To: SwarmFest 2006 Organizers list <address@hidden>
> To: SwarmFest 2006 Organizers list <address@hidden>
> Subject: Re: [Swarmfest2006] Keynote speakers
>
> Jim Anderson wrote:
> > It does bring up
> > the issue of what directions SWARM should take in the next few years.
> >
> How about sponsoring a student for Google Summer of Code?
> http://code.google.com/soc
>
> Moore's Law is finally breaking down, and so computer systems are moving
> toward multi-core CPUs and multithreading.  The Swarm approach lends
> itself to this, and so we could oversee a project to adapt Swarm to
> multithreading or just build an all new simulator that was very
> lightweight and fast and really exploited all available concurrency in a
> model.
>
> If a new simulator were domain-specific, it could be easier to design
> and optimize and make useful.   Also if it was a relatively simple code
> base, it could be compiled for FPGAs using a compiler like Impuse C or
> Handel C.   Some slogan like Swarm Biosimulation on a chip!
>
>
> _______________________________________________
> Swarmfest2006 mailing list
> address@hidden
> http://www.swarm.org/mailman/listinfo/swarmfest2006
>