What is Digg?59 Comments
- richardleis, on 10/12/2007, -1/+16There is no hyperbole extreme enough to capture what this breakthrough means. Wow.
- tetsuwan, on 10/12/2007, -0/+11I'm at the Division for Solid State Physics in Lund. I work a lot with indium phosphide (which they used as an emitter for the laser), and our group frequently cooperates with Santa Barbara. I'd love to see the inside news of this. I work with indium phosphide quantum dots (and possibly wires), not lasing, but this is very interesting.
As for the hyperbole, it remains to be seen how this works out. - ChileanGoD, on 10/12/2007, -2/+13Does this mean we will have a laser fest inside our PC cases???.. Time to invent a USB fog producer :).
- inactive, on 10/12/2007, -0/+11It means that we can be re-programmed during LASIK! No amount of foil can shield us from this!
- Zippo, on 10/12/2007, -0/+7A cute thought, but these chips better be completely vacuum sealed, or else a single spec of dust could block or distort a beam of light could bring your computer crashing to a halt.
- rysolag, on 10/12/2007, -1/+8This sounds like an awesome break through. I think things might be increasing at a faster than exponential rate. UCSB - represent, represent!
- HeavyMetaler, on 10/12/2007, -0/+5I may be wrong but i think optical audio is LED light not laser light.
- EvilPenguin, on 10/12/2007, -0/+5If anyone is using one then they won't see your comment.
- doublebackslash, on 10/12/2007, -0/+5"Did anyone else get a Google ad for Green Laser pointers & Charles Chips? That's awesome!"
Strangely enough:No
https://addons.mozilla.org/firefox/10/
https://addons.mozilla.org/firefox/1136/
https://addons.mozilla.org/firefox/433/
http://www.google.com/tools/firefox/toolbar/index.html
And I can't imagine why. - Shorties, on 10/12/2007, -0/+4"The development is a result of research at Intel, the world’s largest chip maker, and the University of California, Santa Barbara. " Haha Hell yeah... UCSB Represent!
- sonofagunn, on 10/12/2007, -1/+5Yes, you can fit fiber optics inside small areas and still end up with a HUGE increase in bandwidth between components. They don't need to be connected in straight lines.
- greyfade, on 10/12/2007, -0/+4this is primarily useful not in replacing the traces that connect distant chips, but in replacing high-latency lines in large chip dies. for example, the Pentium 4 suffered from high latencies because of the scattered nature of the layout. replacing long-distance lines with optical connections would mitigate that problem.
i say mitigate, because this tech has a long way to go before it's actually useful in major designs.
also, i should mention that high-speed lasers for fiber optics has been around for decades. this is not news in the normal sense. - sonofagunn, on 10/12/2007, -0/+4@zediker - I don't think you'd be replacing all of those fine lines on your motherboard. The processor and RAM are still going to be purely electronic I think. Where optical chips/interconnects are going to help would be, for example, between your motherboard and your network card so your PC could read/write directly to a fiber optic network.
- marksy, on 10/12/2007, -0/+3i think this is different from optical discs. As the laser is directly on the chip and sending data from one chip to another.. Practicle use would be from RAM to CPU perhaps?
and i think optical audio is low wavelength red laser - like a pen laser. - fajita, on 10/12/2007, -0/+3it isn't an issue of them being able to using LED or laser light like a fiber connection.
The issue is that right now the laser or LED must be coded and decoded every time it passes through a system.
So it starts as an electrical signal, gets coded into light, transmitted (at the speed of light), then decoded back into an electrical signal.
From the sounds of it they have been able to produce the light encoding/decoding right onto a piece of silicon. This should speed up the process of coding and decoding. As previously it required additional circuitry and parts on a card. - Primedeath, on 10/12/2007, -4/+7Laser beams!
Pew pew! I hit you for 1d4+1 damage! You die! haha. - inactive, on 10/12/2007, -0/+3Just a quick calculation:
Speed of light (at best): 3E8 (m/s)
Distance traveled: d (m)
Time to travel distance, d: T = d / 3E8
Maximum frequency between components separated by d: 1/T = 3E8/d
So, let's say you are separated by 1 meter, your maximum frequency would be 300MHz. If you have two components operating on data w/ a clock frequency faster than that, your data won't transmit fast enough. Yeah, this helps stuff, but it's not going to make it so that you can have components spread way far apart (at least, not synchronously). - vdog, on 10/12/2007, -0/+3DVD players use laser light to read the disk, but it's converted to an eletrical signal at the head. After the information has been processed by conventional means, it's turned into light by an LED in the optical out port.
Optical buses are good, but we're a long way off optical processors and RAM. - aveyuen, on 10/12/2007, -0/+2@greyfade:
source? They most definitely have not been making Si laser diodes for decades. They've been making Si diodes, and in the last 10 years or so, Si LED's, but i do believe this is the first electrically-pumped Si-based package compatible with the CMOS process flow. There was a Raman laser made a little while back that was optically pumped but that isn't compatible with large-scale processing. I believe that paper was in Nature or Nature Materials; try looking it up if you're interested. - inactive, on 10/12/2007, -0/+2More than that, wire-paths are only part of the problem. Logic is where the real bottleneck is. Make a faster, colder, and lower-power transistor. That's where the money's at.
- inactive, on 10/12/2007, -0/+2I thought that lasers embedded in silicon were already being created...anyway, it's nice to see that this technology is coming up to speed. Light has many advantages over electricity when transferring large amounts of data, like having no crosstalk or AC coupling between adjacent electrical channels. It's also excellent for electrically isolating noisier components.
- Khamel83, on 10/12/2007, -0/+2i just told my co-workers about an article i read years ago that was about lasers replacing wires to speed up data transfer but they thought it i was crazy. now i can show 'em
- inactive, on 10/12/2007, -0/+2Did anyone else get a Google ad for Green Laser pointers & Charles Chips? That's awesome!
That reminds me of the food and laser light show at Six Flags. :-) - zediker, on 10/12/2007, -2/+4optic fibers dont work at sharp angles, because it 'breaks' the angle of refraction.
Sure, you could make an optic computer, but you would have fiber everywhere, preventing cooling air flow. And I havent heard of an integrated optic fiber either. - sonofagunn, on 10/12/2007, -0/+2What really makes this awesome is that all the other components are in place to take advantage of this. The company mentioned in the article, Luxtera, has been mentioned on Slashdot a few times because they have working silicon-based optical interconnects. Soon, components will be communicating at the speed of light!
- cyssero, on 04/18/2009, -1/+3I just have to ask -- don't a lot of DVD players do this through Optical out? Either way, dugg for innovation.
- fantasticjon, on 10/12/2007, -0/+2Does this mean that a remote processor or component would be able to act like part of the computer with little or no latency. That would be awesome! Everytime you buy a new computer you could just hook it up to your network and your main computer gets faster. Or you could borrow unused cycles from the Internet and go really fast. This would really change everything if that is how it would work.
- cal01, on 10/12/2007, -0/+2At first, I thought this would be some sort of new breakthrough with silicon, since it is an indirect band gap semiconductor.
Then I read the article and it says that they just used an InP layer. BAH. I doubt this research is even that new or original. :( - inactive, on 10/12/2007, -0/+2I was just pointing out that there is still a (relatively) large delay between two points separated by a "large" distance even at the speed of light. Wire-path delays are usually measured in picoseconds in modern processors; this delay is much larger than that. It's better than copper, but it's not instantaneous as one tends to think when they see "speed of light."
- inactive, on 10/12/2007, -0/+2Encoding and Decoding? It's just a diode. That's how they work: light comes in, current goes out; or vice-versa, current comes in, light goes out.
- theron1n, on 10/12/2007, -0/+2ad? What's this 'ad' of which you speak?
- theron1n, on 10/12/2007, -0/+2zebov, don't confuse the photon's travel time with the maximum period. The correct period should be based on the hardware's ability to decode the photons back to bits since this is astronomically slower than the time spent by the photon in the optic cable.
- noodlez, on 10/12/2007, -0/+2yeah there is.
its kind of like the advent of the automobile.
people can travel between towns faster, once a company bothers to build an entire machine for this purpose.
as a side thought, it seems to me like there's a chance of a slow down in the encoding/decoding of the signal through the laser. - hurfydurfur, on 10/12/2007, -0/+2Only "omg laz0rz" go pew pew.
- bacon_skoda, on 10/12/2007, -0/+2I remember my professor talking about this 9 years ago in class. I thought he was out of his mind. Sounded like too much work for little gain if it was even possible.
- fantasticjon, on 10/12/2007, -1/+3@zediker
All these problems are good points, but they are minor hurdles and the payoff is worth the redesign. - josegutz, on 10/12/2007, -0/+2This means that when working with chips now you have to wear protective mirror shades... Ahhhh I get it! Like they do in that show CHiPs!
- justice7, on 10/12/2007, -0/+1there is a lot to be gained from this
heat is the biggest enemy of circuits, and photons would be much cooler temperature wise than electrons at the same data rate. (need a physics guy to confirm that)
i think its a great idea, another way of looking at circuits. - Raydr, on 10/12/2007, -2/+3You are aware that fibers are bendable, right?
- shredswithpiks, on 10/12/2007, -0/+1wouldn't this take all (or the vast majority) of heating issues associated with fast chips away?
- OnymousHero, on 10/12/2007, -1/+2Is there a firefox plugin to exclude comments from digg containing 'lasers and 'sharks' ?
- greyfade, on 10/12/2007, -0/+1i was mistaken. they're using indium phosphorus, not gallium arsenide.
but it's still not that new. - Mofo, on 10/12/2007, -0/+1Dugg for my alma mater.
- greyfade, on 10/12/2007, -0/+1wow, they figured out that the gallium arsenide layer on a silicon substrate makes light!
buried for being inaccurate and old.
(for those of you not aware, silicon-based laser diodes have been in production for decades. this is merely an update to the technology that simplifies interchip communication.) - cal01, on 10/12/2007, -0/+1@aveyun: it's uninteresting because they used a layer of InP instead of actually altering the Si structure itself. Yes, having a CMOS compatible flow is important, but the actual "breakthrough" level of this invention is very low. Adding additional layers is easy; all it requires is tweaking the process until it works. Color me surprised if they actually make a Si/Si-XX light emitter.
- nach0, on 10/12/2007, -1/+1“This is a field that has just begun exploding in the past 18 months,”
That's quite an unfortunate quote. - buffawhat, on 10/12/2007, -1/+1(to be spoken in an Igor drawl) Lasers... YES!! Lasers!
- RiemannLebesgue, on 10/12/2007, -3/+3This is pretty big news.
- JCGV, on 10/12/2007, -1/+1@zediker, instead of having x traces running to your GPU you'll have 2 optic fibers (one up one down). Ditto for your hard disk and your dvd burner. Instead of having a s-ata cable (or even worse a p-ata cable). An optic fiber is 0.4 mm in diameter including a protective shield. Let's say we put the 2 fibers in some extra protection on it so we're can handle them more or less normal, what would it be? something like a cable with a diameter of 2 mm (rough estimate).
- inactive, on 10/12/2007, -0/+0Imagine where we would be if the middle ages didn`t happen!
-
Show 51 - 59 of 59 discussions
Browsing Digg on your phone just got easier with our enhancements to the 
© Digg Inc. 2009
— Content created and posted by Digg users is