Advice for 2nd cluster installation

[Robert G. Brown]

On Fri, 10 Jan 2003, Tim Wait wrote:

> > If your task mix is embarrassingly
> > parallel (as it sounds like it might be given that you use a batch
> > queue) 
> 
> Um, what? Sounds like you might be a taxi driver given
> that you own a car.
> 
> Tim
> 

I'd have said more like you might be a bus driver because you own a bus
-- not a sure thing, but not crazy.  Or maybe it is crazy, I dunno.
What do you expect from a bozo?  Anyway, here's my logic:

A lot of folks who do moderate to fine grained computations own their
own dedicated, specifically engineered systems and don't necessarily
need a batch queue.  Batch queues seem (to me, anyway) to make the most
sense for people who share resources.  Resources that tend to be shared
are often "generic" compute resources not really suitable for fine
grained/synchronous computation.  I tend to think of batch queues as
"mostly" being used by folks for whom resource allocation on a
one-at-a-time-basis is fine, not "N out of M at a time, starting all N
at this specific time and holding them until that". Maybe this isn't
really the case anymore.

It's probably just a hangover from the old days, when a batch queue
meant reading your cards into this machine and waiting until a computer
got around to running them, then retrieving output from this lettered
box an hour later.

   rgb

Robert G. Brown	                       http://www.phy.duke.edu/~rgb/
Duke University Dept. of Physics, Box 90305
Durham, N.C. 27708-0305
Phone: 1-919-660-2567  Fax: 919-660-2525     email:rgb at phy.duke.edu