On 8/26/07, James G. Sack (jim) <[EMAIL PROTECTED]> wrote:
> Bob La Quey wrote:
> > On 8/23/07, James G. Sack (jim) <[EMAIL PROTECTED]> wrote:
> >> I wonder when these displays will be available at frys?
> >>   http://ucsdnews.ucsd.edu/newsrel/science/08-07HIPerSpaceDR-.asp
> >>
> >> Maybe UCSD has open houses? (.. housen? hice?)
> >>
> >> Regards,
> >> ..jim
> >
> > Is there something that I really do not get here?
> >
> > I can understand that "clusters of low cost displays"
> > may well be a path toward ultra high resolution displays,
> >
> > But ...
> >
> > Other than spending a ton of money what is worthy of
> > research here? Are there really any architectural or
> > serious theoretical issues about how to do this? I
> > suppose this will impress the Bigfoots and thus is
> > important for that reason.
> >
> > On the surface this just looks to me like "Gee whiz"
> > science of very little real importance. Are they not
> > simply building a "white elephant" that will be completely
> > obsolete in a few years?
>
> You may, of course, be right...
>  ..but..
> I am willing to believe that there are [some] justifications.

OK.

> Isn't visualization one of the big problems with large-scale modeling
> problems (like weather).

Visualization is an important problem. With Frys selling terabyte
disks for $500, and the web continuing to grow like crazy there
obviously is going to be an opportunity to have more and more
stuff to see.

Plenty of big science can use advanced visualization as well.

> And, I bet the devil-in-the-details problems add some long-term
> engineering benefits.

Well that is what I would like to know. On the software side I would
think a _lot_ of those details are best handled by serious architecture
studies and simulation, and not by actually building a large array of
displays.


>From their website
http://vis.ucsd.edu/mediawiki/index.php/Research_Projects:_HIPerWall
<quote>
While the hardware infrastructure of HIPerWall is challenging and
matches the state of the art, the data handling and distributed
visualization capabilities needed to support HIPerWall's capacity are
well beyond current practice.
</quote>

I find myself in more and less agreement with this
statement. Lees with "the hardware infrastructure of
HIPerWall is challenging" and more with the comments on
data handling etc.

Maybe we should estimate the display costs. I find
Apple 30" Cinema Displays going in quantity one for
about $1,800. They need 55 for the actual display.
So say they buy 60 x$1500 = $90,000.

Not much. So maybe it is worth building. Apple may well
have given them a deal. Add maybe another $100,000 for
the servers behind it.

Add  a few grad students, a post doc,  and a prof to
make another $200,000.

Say $400,000 ~ $500,000 ... It might be worth it.
Especially if they are Open Sourceing the software.

> Your question surely seems appropriate, and while I hope that that
> question was asked in the NSF funding process, I wouldn't be surprised
> if there was some razzle-dazzle(!) involved in selling the idea.

I do think that is the _main_ thing. Other than a number
of serious software research aspects there really is no need
for the hardware. But given that the hardware is a small part
of the cost, why not?

> Then too, maybe Nvidia and/or Apple underwrote some costs?

Yeh, I would hope so. But I do hope that does not mean that
the software openness is compromised. It would be interesting
to ask them for a snapshot of the software :)

BobLQ


-- 
[email protected]
http://www.kernel-panic.org/cgi-bin/mailman/listinfo/kplug-list

Reply via email to