> The club has been looking for an excuse to build a cluster and has been
> trying to find some hardware. I thought this Seti project would be a
> good place to start. I don't know anything about Seti so I thought I'd
> offer the opportunity here. If Seti is the limitation, then I'll let
> him know his original idea ( several Windoze boxes ) is the best one.
>
I concur, it seems the SETI project is already doing the redistribution
cycles, so why mess with a good thing(?) Fun, maybe, to see how fast a
beowulf could tackle those packets... For the purpose of understanding how
these systems work, and even to test their stability, it sometimes is
desirable to generate otherwise-ridiculous jobs: I was speaking with a
friend recently about the feasibility of generating an (albeit slow)
ad-hoc beowulf simulation by bringing nodes up inside a VMware session
or the like... any suggestions?
I have also heard about some showing-off by running emulators within
emulators within emulators... sounds fun if it could be responsive.
Anywhoo, it seems that you could find a better test for a cluster than
SETI, or not do a cluster : ) good luck.
ben
