[Pipet Devel] computing at the Field Museum

J. Steinbachs stein at fmppr.fmnh.org
Wed Jun 16 14:58:13 EDT 1999


"J.W. Bizzaro" wrote:
> 
> 
> Hi Jennifer.  I'm a bit confused to see "GCG," "administratively simple," and
> "user friendly" used in the same sentence :-)  Are we talking about the same
> GCG, the Wisconsin Package?  I think there is a Web interface, if that's what
> you're referring to.

I'm embarrassed to say that I have been too lazy to really learn HTML -
my husband sorta taught me a few things, but why not let him do my HTML
stuff for me? :) (his stuff is at http://alloy.net)  So, the GCG SeqWeb
would be administratively simple in comparison to the alternative: my
own home-built web interface to software of our choice (a la
http://bioweb.pasteur.fr/intro-uk.html#log).  We don't have a yearly
operating budget for the beast (doh!), hence my reluctance to completely
take the GCG plunge.  And then the users likely wouldn't make use of all
the nifty tools - just sequence analysis and phylogenetic analysis...
 
> > The other system will likely be some sort of Linux Beowulf cluster
> > without GCG, and for the time being, without a customized friendly user
> > interface - unless of course Loci comes into fruition soon :)
> 
> Wow.  Is the cluster going to be set up in-house or come pre-built by a Linux
> system builder?  As you may have read, Loci can gains computational "power" by
> distributing analyses to more powerful systems.  It would be nice to see it
> running on a Beowulf cluster.

We are currently contemplating two systems, both somewhere in-house:
1. A Sun E3500 or a comparable Alpha-based system, with ETA being the
end of the summer - designed around GCG, for now.
2. Tentatively, a Microway Alpha-based 16-node Linux Beowulf cluster.

Our serious constraint is space - or lack thereof, otherwise I'd just
update UPenn's Phylofarm or NIH's LOBOS system.  . o O (who gets money
for an expensive system without thinking about where they're gonna put
it or how they're gonna cool it?!?!)  We had scoped out a small room,
currently being used as a computer junkyard - but I'm getting the
impression from the building engineers that a/c in that space will be
very challenging given the fact that there is little to no direct
ventilation currently.  Oh, and then there is the problem with the teeny
tiny LEAK in the ceiling.  

> 
> > When I get one or both of these systems, I would be more than happy to
> > alpha and/or beta test (and I can think of a few users who might be good
> > guinea pigs for the user front-end).
> 
> Great.  Do you have a system running GNOME at the moment?  If so, it shouldn't
> be much to get the GUI running that I've been working on.

Right now, I don't have anything here (windowmaker on my linux box at
home).  They've got me in temporary office space (Nothing like arriving
and being told "oh... don't unpack... you'll be moving in a few
weeks...") and I don't even have my own computer - just this nasty NT
thing.  I'm waiting to hear back from Sun to see if they'll be giving me
an UltraSparc 10 through their academic grant program.

BTW, I'm on a gcg-managers mailing list (http://molbio.unmc.edu for
details) and I mentioned Loci on there :) so you might see a few extra
hits on the web page over the next couple of days.

cheers,
-j
--------------------------
Jennifer Steinbachs
Computational Biologist
Dept. of Botany
The Field Museum
Chicago, IL 60605-2496

office: 312-665-7810
fax: 312-665-7158
--------------------------




More information about the Pipet-Devel mailing list