Forgot your password?
typodupeerror

World's First Polymorphic Computer 113

Posted by ScuttleMonkey
from the good-kind-of-unstable dept.
tdelama writes to mention Raytheon Company has developed the first polymorphic computer named the Morphable Networked Micro-Architecture (MONARCH) for the US Department of Defense. "'Typically, a chip is optimally designed either for front-end signal processing or back-end control and data processing,' explained Nick Uros, vice president for the Advanced Concepts and Technology group of Raytheon Space and Airborne Systems. 'The MONARCH micro-architecture is unique in its ability to reconfigure itself to optimize processing on the fly. MONARCH provides exceptional compute capacity and highly flexible data bandwidth capability with beyond state-of-the-art power efficiency, and it's fully programmable.'"
This discussion has been archived. No new comments can be posted.

World's First Polymorphic Computer

Comments Filter:
  • Information free (Score:3, Informative)

    by SnowZero (92219) on Wednesday March 21, 2007 @05:45PM (#18435219)
    Thanks for the information free summary...
    • by rrohbeck (944847)
      You mean the FREE information summary?

      Well they can't write something like "We built yet another piece of programmable hardware" can they?

      • Re:Information free (Score:5, Interesting)

        by dch24 (904899) on Wednesday March 21, 2007 @06:42PM (#18436019) Journal

        Well they can't write something like "We built yet another piece of programmable hardware" can they?

        No, but they should. Not that I dislike Raytheon inherently, but they are certainly spinning this press release pretty hard. It's just programmable hardware. It's an attempt to catch the attention of the government because there are two Military-Industrial coalitions bidding right now for the military's next generation satellite system (which will be a contract worth tens of billions of dollars for about the next decade).

        Since the press release is so light on detail, obviously the actual hardware isn't that impressive. Note things like these quotes:

        In laboratory testing MONARCH outperformed the Intel quad-core Xeon chip by a factor of 10

        Oh, really? And how many libraries of congress per fortnight is that?

        for such purposes as global positioning systems, airborne and space radar and video processing systems

        Target audience, right there.

        64 gigaflops (floating point operations per second) with more than 60 gigabytes per second of memory bandwidth and more than 43 gigabytes per second of off-chip data bandwidth.

        This is at least a little bit of information. However, those numbers are similar to current generation CPUs. I think the PS3 Cell can outperform this chip, so unless we have some power numbers it's unimpressive.

        It's not a big surprise. It's just a press release and a slashvertisement.
        • Re: (Score:2, Insightful)

          by Anonymous Coward
          It is important to understand that current generation CPU's, GPU's....etc, cannot be flown in space. Energetic particles wreak havoc on such small feature sets. SEU's and latch up are serious issues.
        • Re: (Score:2, Insightful)

          by mkramer (25004)
          The problem with existing high-throughput processors are, as you pointed out, power consumption, plus achievable throughput, I/O throughput, and space readiness.

          For most front-end type signal processing, the MONARCH design approaches that of an FPGA in terms of utilization efficiency. When it comes to the next-gen sensors for DoD applications, the Cell doesn't have near the I/O capability required (or, more correctly, the I/O options don't match the processing resource requirements, so you lose efficiency)
        • by mrmeval (662166)

          This presentation has more information. It was developed *with* DARPA and Raytheon sponsored work
          http://www.it.lth.se/courses/DARK/novelarchitectur es/monarch-overview.pdf [it.lth.se]
          "Leverage DARPA-sponsored DIVA Project results, Raytheon IRAD-sponsored HPPS and Mercury Stream Co-procesing Engine"

          Some explainations of the concepts
          http://www.mil-embedded.com/columns/industry_analy sis/2005/10/Dingee/ [mil-embedded.com]

          Just the core design has functions a stock CPU would lack. Also not that before it's even built they have to design in ce
    • by Durinia (72612) on Wednesday March 21, 2007 @05:58PM (#18435405)
      He didn't have much to work with - the press release (err..."article") was information free, too!
    • that free info was wrong.

      The first Polymorphic computer [rwebs.net] was introduced in 1976.
    • The text starts out making it sound like it's got a cluster of FPGAs, reconfigurable on the fly, for the bulk of its structure (or at least the data path).

      But by the time they're done they could even be talking about the Siemens/Infineon TriCore (an embedded processor core with an feature-rich instruction set suitable for process control, serious crunch, or DSP).
    • Thanks for the information free summary...
      Here's some information about polymorphs for you [imdb.com].

      You've been warned.

    • by rjpotts (152026)
      No they didn't! The Polymorphic Personal Computer (1978) [computermuseum.li]
  • But... (Score:4, Funny)

    by rrohbeck (944847) on Wednesday March 21, 2007 @05:45PM (#18435223)
    does it support multiple inheritance?
  • I'm convinced it's a trend these days to build names around clever sounding acronyms and not the reverse. It seems like 'cheating' to me.
    • Re: (Score:3, Insightful)

      by richdun (672214)
      Eh, better than redundant acronyms. That's just lazy.

      It also seems to be a trend to using clever sounding words without actually imparting any useful information. It's like everyone thinks everyone else is either too dumb to understand the complexity or subject matter of what we're doing, or too smart to fall for whatever we're trying to pass off as "new" or "innovative."
    • Re: (Score:2, Funny)

      by Yoooder (1038520)
      Hmm... I need to think of a name that will neatly fit into Antidisestablishmentarianism
    • by x2A (858210)
      "It seems like 'cheating' to me"

      What, Creating Handy, Efficient Acronyms To Inform of New Gear? That's exactly what it is!

    • I think it's called a bacronym. As in backwards acronym.
    • by spun (1352)
      It seems like 'cheating' to me.

      Creatively Hacking Extended Acronyms To Integrate Names Good?
  • Its made for the Department of Defense, it probably costs $5,000 a chip.
    • by Merc248 (1026032)

      I don't care about that DoD! What about... Day of Defeat ?!? Does it run it better than a fast PC with NOS logos on the side?

      On another note, I also want to know if "its ability to reconfigure itself to optimize processing on the fly" means "I can overclock this bitch on air cooling"

      • by x2A (858210)
        "On another note, I also want to know if "its ability to reconfigure itself to optimize processing on the fly" means "I can overclock this bitch on air cooling"

        No, it basically has a lump of human brain tissue from a "willing volunteer" slapped on the side of it. It can reconfigure itself on the fly, but it has to want to.

    • by headkase (533448)
      Buddy, a screwdriver costs $500 when your dealing with the Government. DoD unit prices for the chips are probably in the high tens-of-thousands per one chip. Interestingly, this chip would be a "Universal Turing Machine" for chip architectures (while today the vast majority of microprocessors are von Neumann types).
      • by c6gunner (950153)
        Common misconception based on the fact that government acquisitions, and especially military acquisitions, usually involve an extended support and warranty contract. People repeat this nonsense all the time, without having even the slightest understanding of how government purchasing works. Also, the crazy Jewish guy in "Independence Day" didn't help much with his "You don't actually think they spend $20,000.00 on a hammer, $30,000.00 on a toilet seat do you?" quote.
      • Me thinks you misunderstand UTM [wikipedia.org], Von Neumann architecture is already an impementation of UTM.
    • More like $50,000 per chip today. But in years to come the chip will be available for mass market and it will be only $50.
  • by ivan256 (17499) on Wednesday March 21, 2007 @05:48PM (#18435275)
    Does that mean it's vaporware? If it exists, how can it be beyond state of the art?
  • by PetoskeyGuy (648788) on Wednesday March 21, 2007 @05:49PM (#18435293)
    Low Power, Able to adapt / optimize itself as needed. Sounds like the old Transmeta designs. It would compile and execute code in the processor to emulate x86 commands as needed instead of hard wiring them.

    Of course it also sounds like terminator chip but I think that was from another company and should have already happened by now. ;)
    • by EmbeddedJanitor (597831) on Wednesday March 21, 2007 @05:57PM (#18435401)
      Reconfigurable computing using a bunch of FPGAs.

      All FPGA vendors now offer CPU cores (or you can get others from opencores.org). These cores can do a slew of different functions from DSP to straight CPU functions... and yes they do run Linux!

      For example, Xilinx FPGAs can be reconfigured to run at least 5 different CPU cores, including Java processors etc in single or multi-core arrangements. They can also be reconfigured to do hardware DSP (eg. GPS receivers, sonar processing...). They can implement any peripheral function you care to think of. This makes them pretty versatile for military applications: instead of having to carry a whole raft of different hardware, you can carry one set of boards which can be reconfigured as required.

      • Re: (Score:3, Informative)

        by DragonWriter (970822)
        Yeah, that's what I was thinking. In fact, I was thinking that, brand names aside, the marketing-speak in the summary sounded exactly like the marketing-speak for the Starbridge systems FPGA-based "hypercomputers" when they were announced in the mid-late 1990s.

        • I first thought of SRC Computers [srccomp.com], the late Seymour Cray's last company, which makes dynamically reconfigurable computers (using x86 microprocessors). But they're more into reconfiguring the connections, not the processors themselves...

          Really, this doesn't sound any different than loadable microcode, which nobody's cared much about for the last 20 years. Processing power is pretty cheap these days, so there's no reason to make a processor "retargetable", it just increases the cost and reduces the perform

          • Even in consumer goods space you often need some gates for a particular logic/dsp function + you need a micro for supervisory/user interaction type activity. You could use an FPGA + a normal micro or you could just run the micro in the FPGA. These days the gates required to run a low-end micro in an FPGA only cost 50c or so, about the same cost as a very low end 8-bit micro. A $3 FPGA will run an 8-bit CPU core + do a whole raft of other functions. The FPGA-based solution has some interesting benefits: (1)
          • by julesh (229690)
            But they're more into reconfiguring the connections, not the processors themselves...

            Really, this doesn't sound any different than loadable microcode, which nobody's cared much about for the last 20 years.


            That's not what it sounds like at all. The article described the devices as an containing 6 processors and a reconfigurable "computational array". Sounds to me like a hybrid approach where standard processors are connected to each other through a device somewhat like an FPGA that can be configured to pe
      • by MoxFulder (159829)
        Yeah, this reminded me of FPGAs too... a bit less configurable though. Of course, you could accomplish that with FPGAs as well, by having a system that combined an ASIC processor and support chips with an FPGA or two.

        FPGAs are very exciting, versatile, and fun, so I don't want to knock them.

        This "innovation" doesn't seem to be anything all that new, other than the efficiency and performance numbers, which I find hard to believe.
      • Re: (Score:3, Interesting)

        by julesh (229690)
        Reconfigurable computing using a bunch of FPGAs.

        All FPGA vendors now offer CPU cores (or you can get others from opencores.org). These cores can do a slew of different functions from DSP to straight CPU functions... and yes they do run Linux!


        '"In laboratory testing MONARCH outperformed the Intel quad-core Xeon chip by a factor of 10," said Michael Vahey, the principal investigator for the company's MONARCH technology.'

        I don't think you can achieve that with current generation FPGAs. At least not for the "b
    • Who says it hasn't already happened? They were just better at keeping it hush-hush this time after seeing what happened in the move (a la Spaceballs style) ;)
    • by geekoid (135745)
      How do you know it didn't and that Sarah Conner was successfull?

      Ha!

  • by hey! (33014) on Wednesday March 21, 2007 @05:53PM (#18435331) Homepage Journal
    There is very, very little new under the sun.

    Back in the early 1970s there was a mini computer called the "Meta 4" whose microprogramming could be changed on the fly. The purpose was to let you run software written for other vendors' instruction sets.

    While the chip being discussed may do other spiffy stuff to optimize its performance in different roles, you really can't call it the first "polymorphic" computer.

    • Indeed.

      I remember working on (an emulator of) a "computer with a dynamic architecture"... way back when.
    • by take5 (561870)
      The first polymorphic computer was built much earlier than
      that by RW (later TRW). I do not know when, but UCSB had one
      in 1974, when I was a student there.

      Quote from:

      http://catless.ncl.ac.uk/Risks/9.80.html#subj8.1 [ncl.ac.uk]

      "The conceptual design of this Polymorphic Computer, as they called it, was
      attributed to Sy Ramo, who had earlier helped lead Hughes Aircraft and
      Ramo-Wooldridge (later called TRW) to fame and fortune. The architecture of
      this new machine was an interesting bad idea. The basic idea was to u
  • Vs. FPGA? (Score:4, Interesting)

    by yeggman (599487) on Wednesday March 21, 2007 @05:54PM (#18435347)
    How does this compare to a Field Programmable Gate Array?

    http://en.wikipedia.org/wiki/FPGA [wikipedia.org]

    Is this a bunch of those plus some BIOS like program to optimize it?
    • Re:Vs. FPGA? (Score:5, Informative)

      by Jake73 (306340) on Wednesday March 21, 2007 @06:25PM (#18435791) Homepage
      An FPGA could provide the same functionality, but at a tremendous loss of efficiency. FPGAs have a very high overhead to support reconfiguration. But the reconfiguration is considerably more than this processor (likely) offers. (I couldn't read the article - dead link)

      But with a claim of incredible power efficiency, it's decidedly not an FPGA. I imagine they borrow some of the concepts, but not entirely.

      As a hybrid, FPOA (field-programmable object arrays) provide small programmable "objects" which are less granular than typical FPGA offerings. In the right application, an FPOA can achieve higher speeds and better power efficiency. In the "wrong" application, they're horrible.

      It seems that this device would switch between the high computational efficiency of DSPs and things like graphics processors and the better branching / decision-making performance of general-purpose CPUs.
      • by imsabbel (611519)
        Take a truckload of salt.
        They may claim power efficiency, but im sure they mean compared to those 10 quad-xeons they use as benchmark (i.e. 5kW).
        Lets say you give those FPGAs a power envelope of 250W. And glue your benchmarked algorithms optimized into them.
        I think its doable (the Grape people did some variations of MDGrape on FPGA, and still had near gordon-bell price Gflop/$ ratios for the specific workloads...)
      • by cotijas (1078725)

        I am not sure what exactly is going on with MONARCH, but the idea of polymorphism is not necessarily restricted to reconfigurable computing (like on FPGAs). I think that TRIPS (UT Austin) was the first proposed polymorphic architecture. I am not sure if they have actually built a prototype. The idea is that you build a chip with a bunch (100s or 1000s) of small cores/tiles. Then the tiles can be grouped into larger "virtual cores" depending on the type of parallelism in your workload. A good description is

        • by dave1g (680091)
          They do have prototype on real silicon now. They got them a couple months ago. ( I know a guy working with the project ) Im not sure if they have worked all the bugs out or not yet.
    • Re: (Score:1, Informative)

      by ubergeek65536 (862868)
      Other companies have already been working on using standard AMD servers for reconfigurable computing. Using a CPU socket is a step forward but this has been available for years using FPGAs on PCI cards.

      http://www.xtremedatainc.com/xd1000_brief.html [xtremedatainc.com]
      http://www.drccomputer.com/index.html [drccomputer.com]
  • by aardvarkjoe (156801) on Wednesday March 21, 2007 @06:07PM (#18435567)
    "The MONARCH zaps itself with a wand of polymorph. The arch-lich hits! Oh no, it's using the touch of death! You die..."
  • I, for one, welcome our MONARCH overlords.

    Wait...
    • Re: (Score:2, Funny)

      by Taleron (875810)

      Brock: Don't you have anything else to do but harp on Dr. Venture? Why haven't you tried the World Domination thing, you afraid of the big leagues?

      The Monarch: Please. How stupid do I look to you? World Domination. I'll leave that to the religious nuts or the Republicans, thank you.

    • by rrohbeck (944847)
      >I, for one, welcome our MONARCH overlords.
      I hate them [wikipedia.org]. Make a damn mess on the windscreen in summer.
  • not new (Score:3, Insightful)

    by Anonymous Coward on Wednesday March 21, 2007 @06:14PM (#18435657)
    doesn't sound like a massively new idea at all...

    Cypress semi currently making a MPU that has digital and analog blocks and can reconfigure itself on the fly, its call PSoC.

    example: Coke uses it in their new vending machines, the chip is configured as a mpu during the day and runs the interface, at night it reconfigures itself into a modem to upload data to coke.

    all these people have done is take 6 FPU cores and slapped them on top of a FPGA (or similar programmable logic bank)....good idea? yes. revolutionary? no...

    and its not a computer, its a high speed DSP chip "In laboratory testing MONARCH outperformed the Intel quad-core Xeon chip by a factor of 10," wow, so you built a chip designed for a specific purpose and compared it to a general CPU, good job. You can build an algorithm into a $15 FPGA and have it out-perform a quad core xeon....so?

    -xian
  • by Hubec (28321) on Wednesday March 21, 2007 @06:15PM (#18435683)
    Get a 100k of these running in parallel, give em a self organizing and threading algorithm and run for cover. On the plus side Schwarzenegger's armageddon would be much more interesting than Gore's alternative.
    • I think the one solves the other quite nicely. Saltwater will absolutely destroy all those fancy exposed-parts humanoid robots. And the beauty of it is that the homicidal killbots will take care of our gorilla problem before that.
    • by ShakaUVM (157947)
      If the chips are made by Intel though, the world ending in heat death would be quite likely.
  • Does it run linux?
  • by n6kuy (172098)
    They've invented Reconfigurable Computing [wikipedia.org]!

    What a novel idea....
  • Linux and Monarch sounds like a perfect MS replacement.
     
    • by donglekey (124433)
      Linux and Monarch sounds like a perfect MS replacement.

      This quote sounds like a cliche template replacement.
  • by nomadic (141991) *
    Now we can build units with heavy artillery!

    None of you got that, did you...
  • A few years ago I started writing a novel. I never finished it, although...

    People there were using very small (smaller than a coin) thought-controlled self-modifying computers connected in a peer-to-peer wireless network to communicate with each other and the environment. You think "open the door" and the door opens. You think "it's too bright out there" and your sunglasses dim. You think to that guy: "hey, dude!" and he can "hear" you. You can talk to each other not by opening your mouth, but your minds, t
  • a beowulf cluster of these?
  • As Norman O. Brown may have put it...."Is it polymorphically perverse?"
  • I agree with many comments, this seams so not new... ...on first read it reminded me of "Hypercomputers" by Starbridge systems which [at the time c.2000] had a desktop based on FPGAs boards that matched Cray performance.... and could withstand a Magnum shot through the system (excluding PSU) and still run with minor perf hit.

    At the time you could get a demo video of this 'experiment' in action, however I've drawn short on evidence now we're in 2007.... my only trace via Google is this ol' /. article by Lord
    • by x2A (858210)
      "One of they're earlier claims was it was so dynamically reconfigurable you should shoot it with a .358 magnum and it'd still function"

      Could? :-p

      I can just imaging part of the installation guide that recommends the unloading of a firearm into the computer *lol*

  • do we say 'check mate' ?
  • The Burroughs could reconfigure itself also. But that was the 70's and this new machine is... new so it must be the first!

    http://www.answers.com/topic/burroughs-b1700 [answers.com]
  • by bconway (63464) on Wednesday March 21, 2007 @07:47PM (#18436727) Homepage
    Don't forget, it's from the company that brought us the Patriot Missile (TM), so you know it's gonna be good.
  • Buzzwords aside, a chip that could, rather than require a hypervisor to translate machine code, change to execute the code for the VM whose task is being executed that quantum, would bring a large speed increase.

    For example, one can have a RS/6000 partition running AIX, a simulated ARM processor running a version of Windows Mobile, Solaris on SPARC, and Windows Server 2003 on x86, and when the task switcher changes to the next VM, the chip can natively execute that platform's instructions. No JIT caching n
  • According to Dr. Daystrom, who developed this technology, it is capable of producing a computer far more powerful than the older, duotronic systems which we're accustomed to. Others have suggested, however, that the new technology may yield program execution that is less deterministic, perhaps even less reliable than the conventional systems.
  • Does it run Linux?

    I for one welcome our new microelectronic polymorphic overlords

  • It's a breakrhrough; but we've seen hundreds posted; we're still in silicon, not-light, no amoebas on our chips...has anyone seen a post on Slashdot of something anyone's actually taken to market? (It reminds me of the reports of flying cars in the early 60's...)
  • Since when did slashdot become a corporate mouthpiece?
    Could this article have added anything to a serious discussion of technology?

    Maybe somebody can tell me. Until then I'm a bit disappointed. This really doesn't deserve to be on the front page. --Peter
  • Raytheon Company has developed the first polymorphic computer named the Morphable Networked Micro-Architecture (MONARCH)
    So, it will be able to change shape and connect with the great link?
  • The MONARCH chip is designed to fit into the new DR GIRLFRIEND deep and raspy socket.

Top Ten Things Overheard At The ANSI C Draft Committee Meetings: (9) Dammit, little-endian systems *are* more consistent!

Working...