Slashdot Mirror
100 Million Pixels of Virtual Reality
Roland Piquepaille writes "It's ironic that Iowa State University (ISU) announced a big upgrade of its C6 virtual reality (VR) room the same day as SGI filed for bankruptcy. Back in 2000, this 10x10x10 foot room was powered by SGI Onyx2 computers. The new version of this six-sided VR room will use 96 graphics processing units from Hewlett-Packard. And with its 24 Sony digital projectors, the researchers at ISU will immerse themselves into images of about 100 million pixels in the most realistic VR room in the world. Of course, this upgrade is not cheap. But with this $4 million addition, this new C6 should lead to new advances in urban planning, genetics, engineering or unmanned aerial vehicles."
No, this concerns real genetics - primarily agricultural typing and visualization. And, yes, I am here at ISU.
All the previous posts are lame as hell. I shouldn't add another one, but I have to point out the misuse of the word "ironic". Somebody seems to think that "ironic" means "sad coincidence". No, it means "incongruous circumstances". (There's actually several meanings of "ironic", but this is the one that comes closest to applying.) There's nothing incongruous about this. SGI went bankrupt because their specialized hardware got replaced by commodity hardware. The new VR room uses commodity hardware. No irony here, move along.
Does Hewlett-Packard actually make GPUs? I would think they would go with some off the shelf chips from Nvidia or ATI, surely those can push more pixels than anything else and they would have the advantage of a relatively standard API (opengl for example).
Is there some very specialised requirement I'm not seeing here?
Search first, ask questions later.
There actually are things you can do in the C6 that you cannot do with goggles. For one - and to name something that I know is implementable and implemented - you can track body posturing and position within the C6 to make the experience more engaging/real. Any pictures just do not do this justice; the "seams" shown in the picture are not nearly as obvious in the real thing. On that note, I will say that I've nearly walked into the wall before (on the old system), and missed walking into the screen by a matter of about 6 inches.
With respect to your other comment, the part about interoperability (The goggles would also have the benefit of being runnable on relatively standand class hardware), sometimes you want and need specialized solutions to do great things. Just because you or I cannot hope to afford such a system doesn't invalidate the system.
I just dont get why multi-million dollar visualisation equipement create better research. And I've work in a HPC research center where we have a very nice 3D screen powered by a massive SGI.. And never saw it used to any significant research, sure its a nice toy and its a nice way to blow research dollars. But what a waste. And anyways, most of the time, most researchers where doing their visuation in their offices with their PCs and nvidia/ati cards and their consumer grade crts.. And I'm sure they could see plenty.
#1 -
Yes, though head-tracking is typically only done for 1 user, there are ways to set it up with multiple head-tracks and render/shutter multiple times per application-frame. Further, the difference between tracked/non-tracked users is really only an issue for objects that are near the 'screen' or would be 'inside' the walls. Large-scale or large-distance viewing is not affected since the binocular disparity is so small.
#2 -
The floor of the old C6 could handle 7-8 people safely, which is about as many as you can pack into a 10'x room anyways.
At least while they're alive and/or comfortable.
Of course such a room is very interesting for information visualisation. I think the next best thing is to hardwire a computer to our brain so we don't need a room for so much resolution. And this also would benefit better use of augmented reality
I went to ISAR in 2000, in those days even SGI's weren't getting close to get all the computing force AR typically needs. I wonder how AR is now developing. AR is maybe more interesting for interaction designers to make virtual interfaces for objects from the real world. I have experience many AR applications on ISAR and it gave me a deep impression which VR never has given me. People who find AR interesting (the next really big thing) should follow this link
PS: if someone wants to prove me the VR experience of this thing I might say "hmm.. maybe that's an interesting offer"
I used to teach system admin and hardware repair courses for the Origin2000 and Onyx2 at SGI, and when the class was in Mountain View one module was to visit the "Reality Wall". That screen had only 24 Megapixels projected onto a 120 degree wrap around screen, but even at that the flight simulator was so realistic that students would fall out of their chairs when the plane took a curve.
Poor old SGI. They built amazingly excellent hardware, bleeding edge software, paid their workers well, treated employees like kings and customers like emporers, and donated heavily to the open source movement.
So, of course they went bankrupt.
Done in by the Microslop-ization of technology.
We who were once the high preists of the cult of technology, wizards of electronic wonder, have become the janitors of the Microsoft plumbing, fit only to plunge out the cr@p that clogs the email pipes.
By allowing slackers in our ranks to use shrink-wrap scumware to badly execute business functions cheaply, we have fallen from grace.
"Sic Semper Path of Least Resistance"
...toy, plain and simple.