Re: GNOME needs VA's help?

[Peter Petrakis <peterpan wn net>]

<snip, Summary Miguel wants an Alpha :)>

Yeah, but it's still provides access to 256 individual Alphas that
anyone could use to develop their stuff. To give you an idea as to speed
of this machine I'll compare it to mine which is faster. When I play
with E cvs if i compile starting with the latest glib all the way to E
and eterm it doesn't take me more than 15-20 minutes depending on whats
running. My system is a 533LX 2 meg cache with 128 megs and UW scsi. You
pick up something llike mine with a Permedia 2 8 meg for around $2,500.
The SX's they're using are slightly slower with "only" 1 meg of cache
but they have the MVI instruction set which is like MMX
but more versitile and well better :). Similarly configured an SX would
cost $1,900-$2,000.
Check out www.dcginc.com , www.microway.com , www.harddata.com , and
www.aspsys.com for Alpha systems. If you have any questions about Alphas
I'd be happy to answer them.

Alpha Powered,
Peter Petrakis 

-- 
STCC DEC Alpha Resource Page
http://math.stcc.mass.edu/petrakis
PeterPan@wn.net  AlphaGuy =)
"BLAM! Who the hell are you!? Name's Ash <click clock> Housewares..."

Erik Andersen wrote:
> You realize of course that most apps won't run any faster at all
> on a cluster, since they will only run on one box.  You need to write
> custom code using MPICH (or some other message passing interface lib),
> and make sure you work on inherently parallilizable stuff.  Unfortunatly,
> clusters simply don't act at all like an SMP box (threads don't migrate,
> memory isn't shared, etc).
> 
> That said, I'd like to burn some cycles on that cluster as well... ;-)
> 
>  -Erik
> 
> --
> Erik B. Andersen   Web:    http://www.inconnect.com/~andersen/
>                    email:  andersee@debian.org
> --This message was written using 73% post-consumer electrons--
> 
> --
>         FAQ: Frequently-Asked Questions at http://www.gnome.org/gnomefaq
>          To unsubscribe: mail gnome-list-request@gnome.org with
>                        "unsubscribe" as the Subject.