Please create an account to participate in the Slashdot moderation system

 


Forgot your password?
Close
typodupeerror
×
Hardware

New Optical Chip Claims 8 Trillion Operations/sec. 315

Posted by timothy from the each-century-is-a-matter-of-weeks dept.
Richard Finney writes "Lenslet is announcing the 'World's First Commercial Optical Processor.'. Reuters has the story here. The Inquirer has a cool graphic here on it. The processor is specified to run at a speed of 8 Tera (8,000 Giga) operations per second, one thousand times faster than any known DSP. When Lenslet releases its Enlight processor in a matter of weeks, a unit using the technology will be 1.7 centimetres high and measure 15 by 15 centimetres."
This discussion has been archived. No new comments can be posted.

New Optical Chip Claims 8 Trillion Operations/sec. More

New Optical Chip Claims 8 Trillion Operations/sec.

Comments Filter:

  • what a ripoff! (Score:5, Funny)

    by proj_2501 ( 78149 ) <mkb@ele.uri.edu> on Wednesday October 29, 2003 @10:15AM (#7337724) Journal
    It can't handle 8192 Giga Operations per second?

    • Re:what a ripoff! (Score:5, Insightful)

      by Directrix1 ( 157787 ) on Wednesday October 29, 2003 @10:47AM (#7338090)
      I was almost impressed by this, until I read up on the technology on their website. It will have a pretty limited use as it only has 8-bit precision vector/matrix MAC which is where the 8 teraflops come from. This will be fine and all for just video but it isn't much of a quantum leap for anything else (besides having an optical core). I mean it has power, but there are other chips out there that do more with greater precision numbers.
    • Talk about ripoff -- try downloading the white paper on .... 1.5 kbps.... wish they had their fast processor servin' up the data and not some non-optical 386.

  • Picture (Score:5, Informative)

    by r_glen ( 679664 ) on Wednesday October 29, 2003 @10:16AM (#7337745)
    There's a nice picture of the processor here [lenslet.com]

    • Re:Picture (Score:5, Informative)

      by r_glen ( 679664 ) on Wednesday October 29, 2003 @10:19AM (#7337793)
      Also, a demo video here [lenslet.com]
    • These things are NOT NEW. in 1985 I was a Jet Propulsion laboratory. A caltech professor there was using a light modulator to perform convolution matrix a operations to decode synthetic aperature radar data. THe design is identical.

      and of course this at instance was not likely to have been the first since this was something that was textbook knowledge at that time--fourier processing of signals could be done optically. his was just a particularly advance version, doing more advanced matrix multiples

      • I think the 'new' thing about this is the 8 terraops/sec part
      • YOU were a JET PROPULSION LABORATORY????!??? amazing....didn't know that was possible.
      • These things are NOT NEW. in 1985 I was a Jet Propulsion laboratory. A caltech professor there was using a light modulator to perform convolution matrix a operations to decode synthetic aperature radar data. THe design is identical.

        I was a tech in Emmet Leith's "Radar and Optics" lab at the UofMich and one of the first things I did was run an optical processor using essentially this hack - again to process synthetic aperture radar data. This was in 1967.

        Multi-megapixel 2-D FFT plus some geometry correct
  • would you be able to link this in a Beowulf-type manner?
    • would you be able to link this in a Beowulf-type manner?

      Yes. You can apply the black box theory to optical vs electronic processors. The internals are different, using different logic to form the gates. But the function and operation would be no different.

      The only exception would be how you communicate with the processor. If you interact with it optically, you have to alter the rest of the computer. Or you have to have an optical to electronic converter- which will cost you a little speed, but ma
    • Not only could you make a Beowulf out of it, but you could spread the nodes around the world using the ~80% of fiber that is currently dark.

      If you thought Broadband was fast, wait until you see the applications this device enables.
  • Sorry, COP just isn't cutting it for me. We need a much cooler acronym for this thing...
  • It seems this processor can be useful in lots of applications if for no other reason because of its speed. Why does half the announcement on Lenslet have to talk about how this will revolutionize the defense industry and homeland security?
    -N
    • Ummm because the company is in Israel, a country that has to deal with terrorist type attacks on a daily basis? I thought the same question till I saw

      "...said Major-General (Ret.) Isaac Ben- Israel, former head of the R&D Directorate of the Israeli Ministry of Defense."

      What else is the former head of the R&D Directorate of the Israeli Ministry of Defense going to say about a new chip like this one?
    • Why does half the announcement on Lenslet have to talk about how this will revolutionize the defense industry and homeland security?

      Because that's the two applications most likely be willing to quickly pay a pretty penny for this development - that might matter to an investor.

      Those two applications can put to immediate use very-high-speed logic. Real-time and ver-high-speed filtering of radar (and other) data allows discerning objects that are much smaller, or much farther away, or are much closer to t

  • FYI (Score:5, Interesting)

    by Doesn't_Comment_Code ( 692510 ) on Wednesday October 29, 2003 @10:19AM (#7337783)
    Interstingly, optical processors aren't faster because light is faster than electricity. They are faster because they have much faster rise and fall times between digital on and digital off.

    • Re:FYI (Score:5, Funny)

      by Detritus ( 11846 ) on Wednesday October 29, 2003 @10:30AM (#7337908) Homepage
      Really, I thought it was the cross-polarized emission of tachyons in a Potrezebie field.
    • I believe it also has something to do with the heat associated with electronic processors which should not exist at all on an optical.

    • Re:FYI (Score:5, Interesting)

      by QuantumFTL ( 197300 ) * on Wednesday October 29, 2003 @10:54AM (#7338161)
      Interstingly, optical processors aren't faster because light is faster than electricity. They are faster because they have much faster rise and fall times between digital on and digital off.

      While your statement may indeed be correct, that is not why this chip is faster. The reason is that they are doing analog signal processing using the physics of various optical elements to perform computationally intensive mathematics.

      Think of it this way: We can use large, expensive mathematical operations to simulate optical components, which means we can also do the reverse - using optical components to perform the expensive mathematical operations.

      I read about this about 2 years ago, and it was really quite fascinating to me. It turns out that with a simple lense, you can compute a fourier transform just by focussing the light (it doesn't focus down to an infinitesimally small point).

      I managed to find an article about this, hopefully it should be apparent why this chip doesn't run quake:
      Check it out here. [acesinternational.org]

      They are certainly not the only people doing this. I've seen plenty of references of this being used in missile guidance systems (turns out a simple fourier transform trick can be used to track objects in a camera). Someone I met while working at the Jet Propulsion Lab was working on this Optical Signal Processors. They prove to be very big in the next 10 years.

      Cheers,
      Justin
      • They used to use regular electronic circuits to solve differential equations and similar problems too. They didn't get an exact solution, but they got a usable value. I think that's what you're talking about here.

        • Re:FYI (Score:5, Informative)

          by QuantumFTL ( 197300 ) * on Wednesday October 29, 2003 @12:37PM (#7339109)
          They used to use regular electronic circuits to solve differential equations and similar problems too. They didn't get an exact solution, but they got a usable value. I think that's what you're talking about here.

          You're talking about old analog electronic computers... yeah those weren't very precise (one of the reasons they are no longer used).

          What I'm talking about is a little different. Those electronic ciruits would solve differential equations in the time domain (requiring a bit of time to compute) whereas these optical processors process information in the frequency domain (almost instantly, the bottlneck is as you say how fast they can moduate the light from an electronic signal).

          Frequency domain computing is fundamentally different from the time domain computing in that in time domain analog computers, tiny errors accumulate very rapidly. For instance, an operational amplifier that is used to perform an integration will have a small bias current which will slowly charge the integrating capactor(s), requiring the integration to be rezeroed every so often (at least every few seconds, if not many times a second). In frequency domain computation, the error is not accumulative like that. There is error, and it does add up, but its pretty much orthogonal (the error is spread throughout the frequency space, rather than adding up towards the end of the time space in a time domain computer).

          A really great article I found (this is the one I originally read back in 2001) is here [eetimes.com]. Anyone interested in the more technical side of the processor should read it. It explains why the processing is so fast (because it's essentially parallel rather than serial, along with being based on photons rather than electrons).

          That's where I got most of my information from, along with my optics and mathematical physics classes :)

          Cheers,
          Justin
          Disclaimer: I'm still a semester away from my BS in physics

  • Article (Score:2, Informative)

    by Anonymous Coward
    Israeli Processor Computes at Speed of Light
    Wed October 29, 2003 05:03 AM ET

    By Tova Cohen

    HERZLIYA, Israel (Reuters) - An Israeli start-up has developed a processor that uses optics instead of silicon, enabling it to compute at the speed of light, the company said.

    Lenslet said its processor will enable new capabilities in homeland security and military, multimedia and communications applications.

    "Optical processing is a strategic competitive advantage for nations and companies," said Avner Halperin, vice

  • User availability... (Score:3, Insightful)

    by Beatbyte ( 163694 ) on Wednesday October 29, 2003 @10:19AM (#7337792) Homepage
    This innovative new product will enable revolutionary, new applications in the fields of defense, homeland security, multimedia and communications. The exhibition being held at The World Trade Center, continues until October 15th, 2003.

    The fact that...
    1. its at the WTC
    2. they mention defense and homeland security
    3. its immensely powerful

    ...makes me question whether or not this is going to be available to end users.

    besides the lack of a huge marketing campaign.

    Anyone know anything different?
    • First off, this thing is going to cost a bazillion dollars, and will be a "commercial beta" product for years.. They'll sell them, but theres still a ton of work to improve the thing.

      So who would buy one? Someone with lots of cash - the DoD.. You wanna sell to the DoD, you have to show how it will fight terrorism.. It's just the way things are selling to the government.

      I work writing code for public safety systems - records and dispatching for cops. We did a search and replace on all our marketing lit

  • Gotta Love the spin (Score:5, Insightful)

    by nsingapu ( 658028 ) on Wednesday October 29, 2003 @10:20AM (#7337796) Homepage
    "Processing at the speed of light, you can have safer airports"
    Its really quite sick and disturbing that the aftermath of 9/11 has degraded to a marketing ploy.

    • "Processing at the speed of light, you can have safer airports"
      Its really quite sick and disturbing that the aftermath of 9/11 has degraded to a marketing ploy.

      Well, Lenslet is an Israeli start-up and from what I've heard over the past several years, Israel has had some very strict policies in regards to their airports due to terrorist acts/threats. I really don't think that this has to do with 9/11 explicitly, but the world-wide (or at least Israeli) terrorist threat situation in general.

      I think they're

    • "Processing at the speed of light, you can have safer airports"
      Its really quite sick and disturbing that the aftermath of 9/11 has degraded to a marketing ploy

      It's also disturbing that they (and you) don't seem to realise that the electrical impulses in silicon chips also move at the speed of light.

      Then again, most everything marketing types write makes me want to spew. I presume if they ever say anything that's not deceptive they get banned from the industry for life or something.

    • Re:Gotta Love the spin (Score:3, Informative)

      by Suidae ( 162977 )
      Whats really sad is that you automaticly assume that 'having safer airports' refers to terrorism, as if that is the only thing that makes airports dangerous.

  • dsp vs processor (Score:3, Insightful)

    by redtoade ( 51167 ) on Wednesday October 29, 2003 @10:22AM (#7337824) Homepage Journal
    "the world's first commercial optical digital signal processor"

    When I read the lead post, I thought it was an actual processor like on a PC motherboard... not a DSP. These aren't the same things are they? The possible applications listed on the press release seem to be entirely communications oriented. (ie. fiber optics)

    Now a NAND gate using only optics (not electro-optics) would be fantastic. Maybe using some sort of wave interference to generate the logic table... and as you know you can build all of the other logic gates from a NAND!

    • Re:dsp vs processor (Score:2, Informative)

      by Anonymous Coward
      Generally speaking a DSP is just like any other processor but optimized for certain types of calculations, like fourier transforms and matrix multiplications.

      This thing probably has some very specialized optical processing elements that can do thousands of "ops" in parallell if your code can utilize it fully.

      Remember it's Tera-operations per second, not tera-instructions.
    • and as you know you can build all of the other logic gates from a NAND!

      That's not entirely accurate... you also need a way to link the logic gates together. And a non-destructive way of sampling the signal. (You can only build every logic gate from an NAND if you can hook the NAND's inputs together to create a NOT, meaning one signal, two sinks)

      And if you have an OR gate, the process is MUCH easier...
  • The creation of light normally includes the production of heat, Plus every time a light particle hits something and isnt reflected heat is created..

    Our current use of fiber optics is pretty limited, mostly used for transist of data from point a to point b, in a optic circuit you are going to have billions of particles of light being created and absorbed in a little chip, When ever you change electric energy into light there is loss of energy through heat transfer.

    • .. and everytime an electron hits a defect in a material, a phonon is produced (well, not always, but often) and thus heat is generated. The point is that for any (sanely designed) optical system, the heat generation will be much lower than for a comparable electrical system.

      Heat generation from optical components will not be an issue for the foreseeable future.

  • Environment (Score:3, Interesting)

    by Whispers_in_the_dark ( 560817 ) * <{rich.harkins} {at} {gmail.com}> on Wednesday October 29, 2003 @10:24AM (#7337840)
    What sort of environment would this sort of device need in order to operate? Glancing at the picture I looks like the device internals need to be very precisely aligned to work. How does it react to vibration? Temperature? Phase of the moon (kidding)? Would a regular CEV style environment be sufficent or does it require uber-protection?

    Just curious...
  • IIRC, it took AMD and Intel some five years to match the DSP speeds produced Texas Instruments in the late nineties (>1 GHz). If that analogy holds for anything, it would indicate we may have them on our desktops this decade!

  • Optical (Score:4, Insightful)

    by locarecords.com ( 601843 ) <david@lFREEBSDocarecords.com minus bsd> on Wednesday October 29, 2003 @10:28AM (#7337888) Homepage Journal
    No doubt that optical is fast but isn't the problem always going to be routing the light inside a processor (ie optical transisters) and the interface between the light and the electrical will always cause bottlenecks... I think a lot to solve before this becomes a workable technology...

    • The idea... ultimately... is to use optically based components for the whole computer. Even the peripherals would be connected to the computer via optical cabling (although you may need electrical wires running parallel to the cables for delivery of power).

      The issue of creation of an "optical transistor" is a solved problem, and has been for quite some time now. The real problem has been making them tiny enough to be able to do something that is actually useful in a manageably (and competitively) sized v

  • Ha! (Score:5, Funny)

    by ArmenTanzarian ( 210418 ) on Wednesday October 29, 2003 @10:29AM (#7337905) Homepage Journal
    I patented this idea already, give me money!
    My patent states:
    Using light to do stuff and/or calculate stuff.
    It's all there in black and white.
    • I patented this idea already, give me money!
      My patent states:
      ...
      It's all there in black and white.

       Sorry, they use colors in the light, and your patent only covers black and white.
      Cool use of blacklights, though

  • More Info (Score:3, Informative)

    by 4of12 ( 97621 ) on Wednesday October 29, 2003 @10:30AM (#7337911) Homepage Journal

    Exactly what operations were performed?

    The "vector matrix" multiply is attractive to a lot of people.

    But I doubt this includes fetching data, storing results in memory. And the operations might be more like one-bit XOR's than general Level 3 BLAS.

    Need more information...

    • Re:More Info (Score:3, Interesting)

      by Mozz Alimoz ( 245834 )
      I'll have a guess at how this works.

      The article says "The Ablaze(TM) is the Spatial Light Modulator (SLM) in the optical core of the EnLight256(TM)". Going by the graphic in the Inquirer [theinquirer.net] article, they shine a row of blinking lights through a LCD-like device [lenslet.com] (and some lenses and mirrors I assume) and collect the results in a column of light sensors on the other end.

      Each pattern of on/off elements on the LCD-like device gives them a different transformation running at however fast you could emit and sense t

  • Does anyone know how this tech works? Won't it be limited by the electronics it's attached to?

    Photonics [psc.edu] promises to give us an all photon path but I don't think anyone's close to making an entire processor with photonic crystals yet.

  • It scares me to think of the cataclysm that will occur when someone tries overclocking this thing too hard....

    In all honesty though, I can't imagine a reason anyone would try to boost this thing past it's current potential, considering the applications created thus far for ANYTHING computationally related aren't near this things capacity, I think it's safe to say this thing will hold a landmark speed for years to come without any thought of needing to "jack up the juice"...

    GAH! I CAN'T HELP IT!!!
    Jus

  • mirror of that cool pic is here (Score:2, Informative)

    by Anonymous Coward
    here is a mirror of that pic, cuz that site is already slashdotted by now.

    http://www.stuwo.net/temp/i_products_enlight.jpg

    http://www.stuwo.net/temp/i_products_enlight.jpg [stuwo.net]
  • Santa-beard [newgrounds.com] on that...

    My mind is alredy overflowing with potential applications. I want to smash one to see how it works. Actually I want to smash them all because I'm afraid of things that I don't understand. Although once I understand it that urge will go away. This is like, God's gift to DSP engineers. It's both exciting and frightening to contemplate it's potential applications.

    I guess when and where can I get one and how much will it cost?
  • Can it run Doom 3 at a good framerate?

    Seriously though, that sounds very cool and it is deffinitely the future of computing, but how much heat does it produce? And why do the articles talk about how "Bulky" it is. Sure it's big (although I've seen RAID cards that took up almost an entire full size case), but if there was a 7GHz athlon out that was the size of my head, I'd deffinitely buy it. I don't care that it takes up so much space.
  • The instruction set for the benchmark?

    00: [NOP]
    02: [NOP]
    04: [NOP]
    06: [NOP] ...
  • how does doom3 run on it?

  • Reading the fine print ... (Score:4, Insightful)

    by fygment ( 444210 ) on Wednesday October 29, 2003 @10:54AM (#7338159)
    ... at the Lenslet [lenslet.com] page, the unit actually has several components. The VMM (vector matrix multiplier) does 8000 MAC (matrix array calculations) but there is a VPU (vector processing unit) that comes in at 128 Giga-ops and which would be the bottleneck in the whole setup. No question this is a huge improvement BUT to put it in perspective, it is a DSP only, not a computer system (although some neural network weenies might see a way of turning this into something more than just a DSP). In any case, the bottlenecks will come from the equipment it has to operate with both onboard and off.

    Still, note that it's developed with Matlab [mathworks.com]. Now surely that is the Holy Grail of research, a bitchin' language with an awesome tailored processor. Imagine the logo Matlab [Lenslet Inside].

    • (although some neural network weenies might see a way of turning this into something more than just a DSP)

      I'll ignore the weenies bit. . .

      but yeah, that's actually the first thing I thought of when I saw it was basically only good for matrix operations. Granted, the precision is really low, but depending on how expensive this thing is it could make really fat neural networks a whole lot cheaper to run.
  • Okay, what the hell IS one of these things and how does it WORK?

    There is a diagram [lenslet.com] on lenslet.com that shows how their optical processor works. There are three parts, a row of lasers, a row of photodetectors, and a big grid of Multiple Quantum Well (MQW) Spatial Light Modulator. I assume this grid is where the matrix operations actually take place. I don't even care about the math, I never could understand it. But from a physical point of view, how is this thing constructed? What _IS_ it?

    I tried searching
  • Is this is a general purpose CPU with a full range of instructions, or just an analog light processor that can do some very specialised tasks very quickly?

  • Window of opportunity:The Open Source Optical RISC (Score:3, Interesting)

    by NZheretic ( 23872 ) on Wednesday October 29, 2003 @11:22AM (#7338368) Homepage Journal
    The introduction of this technology over the next five or so years offers a window of opportunity to collectivly develop Free/Liberty/Open Licensed optical RISC ( Reduced Instruction Set CPU) technology.

    The shift from electronic to optical results in a massive reduction in the time it take to change states, to the point where it possible to, once again, build a CPU from relatively widely spaced modular optical components. You can build a single optical CPU spread over a motherboard or even cabinet sized area and it will still be several magnitude times faster than the fastest silicon/electronic single chip CPU.

    No one but the biggest companies are going to have the capital nessary to collect and shrink the resulting designs down into single optical chip hardware and manufacture the result, with a further magnitude increase in performance. As with the existing CPU industry, it is likely that the market could maintain only a few such CPU companies. Opening up the design and development process, as with open source development, would result is a far more rapid pace of development. Relative obsolescence woul;d insure that there would plenty of opertunity for large profits for the large and small manufactures.

  • Never thought people would go this far just to play doomIII...
  • So an SLM could be just one of those TI-invented micromirror arrays right, like they use in projectors? Probably they already had to figure out how to calibrate those, to get smooth color blending for the projectors, so that for example if you set one mirror to the position 128 you get exactly half as much light as if you set it to 256, and half as much again at 64, and so on.

    Vector-matrix multiplication involves summing the products of the vector cells and the matrix column cells to get one output value

  • The power of the eye (Score:3, Funny)

    by Zhe Mappel ( 607548 ) on Wednesday October 29, 2003 @11:31AM (#7338458)
    What's really cool about this is how it harnesses the power of the eye.

    Optical, derived from the ancient Greek word optikos, literally means "the focal power to perceive fair Helena while she is sunbathing nude in yonder olive grove." That's eye power to you and me, dude.

    Some critics will say that a major drawback in these new systems is the need for a mechanized eyedropper next to the chip, keeping the core moist and supple at all times. You don't want this chip going red-eye on you during mission critical tasks.

    Still, modders are going to go wild. Within minutes, you can change the color of your CPU's iris using the very same dramatic contact lenses worn by today's biggest infomercial stars.

    Unfortunately, if you're into porn, excessive downloading can make your computer go blind. That's why I'll be recommending to my porn-intensive clients that they stay on Wintel systems.

  • The key to this architecture is that it's not Von-Neuman... You don't have billions of transistors waiting for their chance to interact with an instruction stream. The fact that it's an optical multiply-accumulator is just an distraction.

    Take a silicon die, build a 16x16 - 16bit (32 bit result) MAC on it, run it at 1 Ghz (all feasible with modern technology), you get 16x16*1Ghz - 256 Billion ops/sec. I'm guessing this could be done for less than a buck/chip in any kind of quantity. Stack up 256 of those..

Slashdot Top Deals

news: gotcha

Close