Slashdot Mirror
Highest Resolution Wall Around
akhaksho writes "NCSA (the National Center for Supercomputing Applications) is in the
process of building the highest resolution display wall in academia. This is similar to the previous story about the wall at Sandia, but the intention of this wall is to get very high resolutions at a reasonable cost using off the shelf technology (for the most part). All of the code to run it and plans for the physical infrastructure will be available as part of the Display Wall in a Box effort. I'm
one of the guys that built this sucker (and have the scars to prove
it!) "
and the image on the wall would come from ... where exactly? Those 20 graphic cards are rendering the image fed to the LCD projectors that, uhm, project them into the wall.
It's not a screen, it's 20 projected images on the wall. Just remove 10 and you have "half" a display.
Having the ability to zoom in and out on a high resolution monitor isn't enough. Neither is having a huge picture. They want both of these attributes in the same device. It makes it easier for a group to crowd around or for someone to do presentations with. They keep the detail and have it the size of a wall.
Also (and I'm serious here) having big expensive toys that people look at and think "WOW THAT'S SO COOL" draws attention, which draws clients, which generates funds. Hey, they got posted on Slashdot, and now a million geeks know about their work. If even a fraction of them put any interest into the company, it will fund the "Highest Resolution Wall" project and pick up a few more participants.
Just my thoughts,
~LoudMusic
No sig for you. YOU GET NO SIG!
Yeah, but chances are, you've not going to be sitting 12 inches away from this thing. You're gonna be sitting a good ways back.
Heck, I'd pay $50 for an hour of quake on this. They could rent it out on nights and weekends and it would pay for itself in a month or two. Well, maybe a year or two. But it would still be a good investment...
Neh
... and there is no doubt, that one day he will be
where the eye of his telescope has already been
Considering that they're projection, and you could put the projector into something that contained the radiation. I don't think it would be much of a problem.. if at all.
Free Mac Mini
The 9-foot-high, 12-foot-wide screen can project images more than 20 times better than the typical computer monitor. The display surface--a screen divided into 20 sections--can display images of 4,096 x 3,840 pixels either as one large, high-resolution image or as several side-by-side images and information nodes.
OK, let's do some simple math. Let's say I run 1024x768, so this wall is going to be 4x as wide and 4x as tall, pixel-wise, correct? Now, if the screen is 9 feet tall and 12 feet wide, we get a diagonal of 15 feet (thank you, Pythagoras). The resolution is up about 4x, but it's spread over an area that is maybe 12x as large? (I'm using a simple 15" screen to keep the math simple.) We're talking about 1/3 reduction in dot pitch, aren't we? What's going on here?
I guess the application is for use a video wall to be seen from far away. Contrast this with that other enormous high-res display from that other article. It's clear that this wall will be more for public viewing, and not real "work," like medical imaging or whatever.
Nonetheless, though, I think it'd be cool to watch my DVD's on....
Where the wind blows, the tumbleweed goes.
dude and dude above you: I totally missed that part of the article! I thought it was like the previous slashdot article had been. Just a really massive (but conventional) display. Sorry.
Presumably each Linux PC is responsible for generating a subrectangle of the display. Yes, that would make it one of those "wussy grids of normal displays", but since they are using LCD projectors to throw up the image, they can calibrate them so as to make the tiled display fairly seamless.
I don't care if it's 90,000 hectares. That lake was not my doing.
We're using WireGL, which is a distributed OpenGL implementation out of Stanford. It has some code in it to synchronize the displays. Since we're using Myrinet (gigabit network with low latency), the pipes sync up pretty well. As for the stereo, we don't plan to do either. There are no projectors of this cost and form factor that are capable of active stereo and the complexity of mounting the projectors for passive stereo are too horrible to contemplate! :) We've got the other 20 projectors and plan to build another identical block next to the one we've already got.
and the story about VisuaLABS groutfree wall display technology is related?
You mean a Circuit City TV-in-a-Box? Or a Circuit City Opened-TV-without-a-Box, but covered under the same warranty?
--
"Outlook not so good." That magic 8-ball knows everything! I'll ask about Exchange Server next.
In Issac Asimov's third novel in the Foundation series, Second Foundation, mathematicians used a video wall projector to display their equations. As described in the book, the Prime Radiant did not cast a shadow, yet the walls were covered with equations. The coolest thing was, you just thought about a part of the equation and those lines marched down the wall to eye level. So combine this video wall with the mind-activated cursor talked about in this issue of Wired and we're almost there.
For stereo, that is...
I will be making a few assumptions (my bad), so bear with me.
Imagine that you have all 40 of the projectors set up, but instead of 4 x 5, it is now 8 x 5, like so:
RLRLRLRL
RLRLRLRL
RLRLRLRL
RLRLRLRL
Now, imagine that the R's are one set of projectors, aligned to project onto the screen as the system currently is set up (4 x 5), and the L's are set up the same (so that an adjacent R and L project onto the same area, overlapping perfectly). Throw a set of polarized filters in front of each (or, for that funky 70's effect, red/blue filters - or just tweak the colors), then wear the proper glasses.
The drivers (and the cluster) would have to be set up to throw the proper image to the proper sections - I don't know if you would have to divide the cluster in half, or what, to do this (maybe even need driver mods - ouch)...
Actually, it shouldn't be too hard to set up - other than requiring double the horizontal space (plus there might be distortion issues as well, due to space between the projectors, I would imagine). Besides, you already said:
We've got the other 20 projectors and plan to build another identical block next to the one we've already got.
Could this work?
Reason is the Path to God - Anon
Oops.. my mistake.. that was a ceiling, not a wall.
Ratguy
The other thing that gets me is the use of the term resolution. In raw terms, this display actually has very poor resolution: about 28 dots per inch. If you stand way back from it, it might have a high number of dots per degree of arc of vision. But then, how bright is it from 100 feet away?
Yeah, that moderator crack hits ya pretty hard, doesn't it?
"Technically, a cat locked in a box may be alive or dead." -Kurt Cobain
The US Air Force has been working on the Interactive Datawall (http://www.rl.af.mil/tech/programs/ADII/adii_dw.h tml), which includes laser pointer tracking and voice recognition among other things.
Eventually they hope to have a portable version so various mil units can just cart them around to whatever theater they're needed in.
We've come a long way from LeMay's old "Big Board"...
:wq
A resolution of 28 DPI sounds low until you figure that a typical display is only in the 72 to 100 dpi range, and is viewed from only about two feet away. A screen 18 feet wide by nine feet tall is unlikely to be looked at that closely -- It's more likeley to be viewed from about 10 feet or more away - any closer would effectively prevent someone from seeing the whole picture. And at that distance, that's the same apparent size per pixel, or better, than even high end monitors and displays.
As for brightness - as you get further away from a uniformly emitting planar surface, the brightness per unit angular area remais constant - merely the apparent total area decreases. If it's bright up close, it's bright enough at any distance at which it is a significant part of your field of view.
And with a fairly high powered cluster to generate the graphics, this can probably render animations of various problems (e.g. turbulent airflow over a surface) in real time on a 1:1 scale - not to mention 3d walkthroughs of complex structures and simulations of advanced weaponry.
Quake, anyone?
Liquor
Sanity is a highly overrated commodity.
finally, someone has made a beowulf cluster of monitors.
-- free as in swatantryam - not soujanyam.
Hi, in case you guys want to spend some time answering questions here, how are you syncing the graphics cards across the wall? SGI claims you need an Infine Reality to achieve decent syncronization, which otherwise is very noticeable. Personally I've seen one CAVE type of installation driven by PCs syncronized over a RS-323 port and I can't honestly say I noticed the projectors being out of sync.
Which takes me to the other question: are you using or do you plan to implement active stereo projection or do you want to install another 20 projectors and use a passive system?
I have been trying to think of what types of data display you could do on this type of screen
;]
some kick-ass fractals,
"My mother never saw the irony in calling me a son-of-a-bitch." - Jack Nicholson
I don't know... huge video screens, like VR, kind of freaks me out. Too much of an opportunity to forego reality.
Me too! I've played Quake2 in the Cave, and it's really intense. I saw the Quake3 graphics demo on the wall, and I've never seen a sharper picture in my life. As for those wondering about the cost justification.... It's research! They need very little of that. www.lan-fest.com
That what happens when you get computer guys trying to cut, drill, and tap aluminum extrusion!
Very large dataset display, a couple of things come to mind:
Pictures of the universe, it is hard to get things visible, and yet to scale from other things.
Very large graphs, the graph of the distribution of hosts across the internet.
Modeling of large molecules.
If you get scars from building it, I don't wanna build one!
Free Mac Mini
This is an honest question, not a troll. I have been trying to think of what types of data display you could do on this type of screen that you couldn't do with more conventional technology. If it is sheer size you want projection technology can do that, and if you want fine grained imaging you can use high resolution computer monitors and zoom in and out. How does this really large, finely detailed display benefit the research? Again this is an honest question. The cost of this thing cannot be trivial. What do you see here that you can't see anywhere else?
:-) So, anybody have any ideas?
Don't get me wrong, I'm all for technology for it's own sweet sake. But I'm curious how this is cost justified. Those reasons/rationalizations could prove very useful. (I doubt anyone I am likely to work for would consider Quake sufficient motivation, now that the golden days of the dot coms are over. Maybe these guys are just the coolest folks on the planet
"I object to doing things that computers can do." -- Olin Shivers, lispers.org
Does this by any chance use VisuaLABS GroutFree(tm) technology?
Wow!
"... 8,096 pixels across and 3,840 pixels high on an 18-foot-wide screen ... "
Porn wouldn't be any fun...
"Good Lord, I think I can see her kidneys!"
comment 61
Sorry.
I guess i'd have to be more familiar with the WireGL and how it renders frames, but as long as most of the crunching is done on the video cards, I would think that you could do this with many fewer systems. Especially since you are only running at sub-1024x768 resolution per screen. How about 4 dual athlon PC's with 1 AGP and 4 PCI dual-head cards each? That would still give you the same number of pixels, at a significant savings in hardware cost.
Although you probably couldn't play quake on it, could you do most other things?
Neh
... and there is no doubt, that one day he will be
where the eye of his telescope has already been
It is transparent to OpenGL apps. We're using WireGL out of Stanford which is a distributed OpenGL implementation.