Forgot your password?
typodupeerror
Supercomputing PlayStation (Games)

Eight PS3 'Supercomputer' Ponders Gravity Waves 293

Posted by CmdrTaco
from the because-you-can dept.
Jamie found a story about a inexpensive supercomputer being used by an astrophysicist to research gravity waves. The interesting bit is that the system is built using 8 PS3s. Since nobody is actually playing games on the system, it makes sense to use them for research projects like this, but I really wonder now what is defining 'Supercomputer'... I mean, a hundred PS3s sure, but 8? I think we are de-valuing the meaning of the word 'super' :)
This discussion has been archived. No new comments can be posted.

Eight PS3 'Supercomputer' Ponders Gravity Waves

Comments Filter:
  • by EveryNickIsTaken (1054794) on Wednesday October 17, 2007 @10:51AM (#21010471)

    I mean, a hundred PS3s sure, but 8? I think we are de-valuing the meaning of the word 'super' :)
    I think we are de-valuing the meaning of the word 'inexpensive' :)
    • by The13thSin (1092867) on Wednesday October 17, 2007 @11:03AM (#21010671)
      For a supercomputer that's pretty cheap. Also I find the statement in the summary that there are no games to be played on the PS3 a bit childish. The PS3 has not been out for a year yet and there are multiple great games to get for it right now and even more coming very soon. I expected more from the Taco.
      • Re: (Score:3, Informative)

        by Bandman (86149)
        It's not just in the summary; read the article, it dishes out the abuse
      • by NeilTheStupidHead (963719) on Wednesday October 17, 2007 @11:18AM (#21010913) Journal
        I think that the article imples that PS3s aren't flying off the shelves as fast as Sony might like and thus are sitting in a warehouse somewhere, otherwise going unused. Even the article claims that this was done mostly because of the open platform presented by Sony and the fact that this researcher was able to get the consoles free from Sony. This is great for Sony because a sold console is money in Sony's pocket regardless of who buys it and what they do with it. If they can convince researchers to buy PS3s then it's probably a better deal than selling them to gamers. Few gamers would buy the equivalent of 7 PS3s (about $2800) worth of games and accessories. Some will, but most won't; even when they do, it's spread over the life of the console. A researcher goes and gets 8 consoles, cash up front and there's $3200 for Sony (less taxes, mfg. costs, etc).

        Maybe it's jsut me, but that sounds like a pretty good deal from Sony
        • Re:Inexpensive, eh? (Score:5, Interesting)

          by smussman (1160103) on Wednesday October 17, 2007 @11:22AM (#21010987)

          I think that the article imples that PS3s aren't flying off the shelves as fast as Sony might like and thus are sitting in a warehouse somewhere, otherwise going unused. Even the article claims that this was done mostly because of the open platform presented by Sony and the fact that this researcher was able to get the consoles free from Sony. This is great for Sony because a sold console is money in Sony's pocket regardless of who buys it and what they do with it. If they can convince researchers to buy PS3s then it's probably a better deal than selling them to gamers. Few gamers would buy the equivalent of 7 PS3s (about $2800) worth of games and accessories. Some will, but most won't; even when they do, it's spread over the life of the console. A researcher goes and gets 8 consoles, cash up front and there's $3200 for Sony (less taxes, mfg. costs, etc).

          Maybe it's jsut me, but that sounds like a pretty good deal from Sony
          IIRC, Sony sells consoles at a loss, and then gets that money back over the life of the console with license fees from games. So selling 8 consoles which will not generate money from game license fees, but still having to take the loss is not a good deal for Sony.
          • Re: (Score:3, Insightful)

            Giving away 8 consoles that will not generate money from game license fees and getting an article in Wired that's linked to by Slashdot is a good deal for Sony marketing.
      • by Rogerborg (306625) on Wednesday October 17, 2007 @11:55AM (#21011561) Homepage
        If you thought that was childish, then this should drive you into a perfect fit of apoplexy [penny-arcade.com].
  • It would've been cheaper to just buy a Cray.

    *ducks*
    • Re:Obligatory (Score:5, Interesting)

      by threaded (89367) on Wednesday October 17, 2007 @11:09AM (#21010767) Homepage
      Not true. I know of at least one place where they have a Cray and can't afford the electric bill to switch it on. They cost a fortune just to sit and look pretty too: it's taking up room on campus that could be used for other things.
      • Re: (Score:3, Insightful)

        by Maller (21311)
        Please enlighten me. Who is stupid enough to by a million+ dollar computer without factoring in facility costs?
    • Re:Obligatory (Score:5, Informative)

      by adisakp (705706) on Wednesday October 17, 2007 @02:33PM (#21014097) Journal
      It would've been cheaper to just buy a Cray.

      If you read the article and followed the link to his PS3 Gravity Grid [umassd.edu] site, you'd know a couple things about the cost (FREE) for this computational power:

      #1) The total cost of purchasing an entire "PS3 Gravity Grid" supercomputer for yourself is less than the cost of a single simulation run on a BlueGene. In other words, you can buy the cow, the pasture, and a barn for the price of a gallon of milk.

      #2) Sony *DONATED* his 8-node cluster (albeit with 20GB PS3's which they were closing out at the time) so he actually got a "supercomputer" for nearly free.

      #3) The power of the 8-node PS3 cluster is roughly the same as a 200 node partition on a BlueGene SuperComputer (1 PS3 = 25 Blue Gene nodes). With 8 Cell CPUs, he has 56 SPU's running at ~3GHz to crank through his computations. This would mean a single CELL SPU is roughly 4X more powerful than a single BlueGene node which isn't unreasonable considering that it runs at a higher clockspeed (the supercomputer has to worry more about heat dissipation with hundreds or thousands of cores).

      #4) I believe that by the US Gov't's somewhat outdated standards, a PS2 qualifies as a supercomputer. The FPU power in a PS3 is on ther order of 200 times higher than that of the PS2 for single precision and considerably more for double precision (which is emulated in software on the PS2).
  • by zifferent (656342) on Wednesday October 17, 2007 @10:53AM (#21010501)
    64 cpu's. That seems supercomputerish enough for me.
    • by andphi (899406) <<moc.liamg> <ta> <maspillihp>> on Wednesday October 17, 2007 @10:56AM (#21010555) Journal
      Should be enough for anyone
    • Re: (Score:3, Funny)

      by purpledinoz (573045)
      The PS3 is the most powerful blue-ray player in the world. 8 PS3's makes a blue-ray player that has the power of a supercomputer.
    • by Waffle Iron (339739) on Wednesday October 17, 2007 @11:07AM (#21010737)

      64 cpu's. That seems supercomputerish enough for me.

      It depends. For those problems that fit within the PS3's cramped memory, this is a supercomputer. For those problems that don't, this is a set of 8 matching doorstops.

    • by Kupek (75469) on Wednesday October 17, 2007 @11:08AM (#21010765)
      It's only six cores available to Linux per Cell processor on a PS3. One is reserved for the Game OS, and one is disabled to achieve a higher yield on fabrication. (The Game OS is always running, since Linux actually runs on top of a hypervisor.)
      • 9 cores? (Score:5, Funny)

        by oblivion95 (803698) on Wednesday October 17, 2007 @12:27PM (#21012097)
        They should have used 9 cores on the Cell. Then, they'd be using Seven of Nine.

        It's only six cores available to Linux per Cell processor on a PS3. One is reserved for the Game OS, and one is disabled to achieve a higher yield on fabrication.
        • Re: (Score:3, Informative)

          by adisakp (705706)
          If he's using the PPU core (hyperthreaded-not multicore) and 6 SPU cores, he is using 7 of 9 (1 disabled and 1 reserved of the 8 SPU's + 1 PPU) :-)
      • Re: (Score:3, Informative)

        I code on a PS3 running Linux. There are 7 cores available to Linux. 1 PPU + 6 SPU cores.
        Ok the PPU is not as powerful as an SPU, it's a basic in-order dual-threaded PowerPC core with the AltiVec instruction set, but you shouldn't ignore it.
    • by flaming-opus (8186) on Wednesday October 17, 2007 @11:23AM (#21011007)
      64 cores does not make a supercomputer. There are database servers with more cores than this, and have been for years. Technical computer, sure. Maybe even high performance computer. Definately NOT supercomputer. 8 systems, that's what? 4GB of RAM? There are laptops that can hold that much memory.

      If you went to a technical conference like, for example, Supercomputing '07, you would get laughed off the floor calling that a supercomputer. Supercomputer is a changing definition, but I don't think I'd call anything a supercomputer that didn't have at least 1TF of peak double-precission performance, and at least 200GB of RAM.
  • Not surprising... (Score:3, Insightful)

    by grocer (718489) on Wednesday October 17, 2007 @10:55AM (#21010521)
    I believe that people were clustering PS2 for research shortly after the release of the linux kit...cheap processing power is cheap processing power.
    • Re: (Score:2, Insightful)

      by JamesRose (1062530)
      I could be wrong, but isn't it actually quite expensive, because within those 8 PS3s, you're buying 8 very high end graphics (possibly integrated, but still there), which surely would bump up the price by quite a large amount, would it not have been easier to buy components?
  • Strange... (Score:4, Funny)

    by Anonymous Coward on Wednesday October 17, 2007 @10:55AM (#21010529)
    Noone has wished for a Beowulf cluster of these yet. Has the world order changed and no notification issued?
    • by trongey (21550)
      I think the real /. has been turned to stone by Natalie Portman's hot grits. This is just an imitation from Soviet Russia.
  • by szyzyg (7313) on Wednesday October 17, 2007 @10:56AM (#21010537)
    Just over a decade ago the fastest computers in the world were barely breaking the teraflop mark, today in theory the XB0x and Ps3 with their multitude of cpu cores and finely tuned graphics cards can top that. So 8 Ps3's - if you believe sony's hype could clock in a >10teraflops if the hardware was well utilised.

    I had a freiend who wrote a book 'Nemesis' which was a spy thriller involving a killer asteroid - it was published in the UK 1998, and back then he was talking about 'the teraflop box' as being the fastest computer in the world, unfortunaly it took 8 years to get the book released in the US and by that time a lot of the computer jargon had dated significantly, and you could get a teraflop box in the form of a turbocharged graphics card or cutting edge games console.
    • by edmudama (155475)
      Please mod the parent up.

      The PS3 is a pretty advanced processing platform, much like today's top-end video cards, especially when working with large sets of data doing floating point math. I'm not surprised at all that it can match the performance of 200 or so pentium-grade cores. (After all, joe blow researcher doesn't get time on one of those top-5 boxes when he signs his check for $5k.... he gets yesterday's tech)
    • The teraflop won't be usable for general computing (not even close) but if you add up all the little floating point units on the chip you'll get a teraflop.

    • by Stefanwulf (1032430) on Wednesday October 17, 2007 @12:55PM (#21012511)
      Sony's estimate of 1 teraflop per unit is more than a little bit optimistic, and also counts all the operations performed by the graphics system, which the average researcher won't have access to. For a slightly more realistic estimate of what 8 ps3's running linux can do, I'd point you towards http://www.terrasoftsolutions.com/store/index.php?submit=software&submitimg%5Bhardware%5D%5Bsolutions%5D=1, [terrasoftsolutions.com] who sell turnkey ps3 clusters. They are claiming that 8 units together break the theoretical teraflop mark, which seems more realistic to me. While that isn't among the fastest computers in the world by any stretch, it's still solidly in the realm of what I'd call a supercomputer.

      The cell is a fantastic piece of equipment - Dr. Dobb's has what I think is an excellent analysis of the kinds of performance benefits that it offers at http://www.ddj.com/hpc-high-performance-computing/197801624 [ddj.com]. I'm currently running one at home in a PS3 (for neural networks that drive an AIBO - I love Sony's tendency to dump hugely expensive hardware at mass-market prices), and I have every intention of picking up more used ones over the coming months to cluster together as the networks continue to grow. Even all by it's lonesome with code that's far from optimized, the one I have is running about 10 times faster than my main desktop for roughly equivalent computations.

      (Note that your mileage may vary - I just happen to like playing with systems that parallelize really well)
  • Devalued super (Score:4, Insightful)

    by Teese (89081) <beezel@@@gmail...com> on Wednesday October 17, 2007 @10:56AM (#21010545)

    I think we are de-valuing the meaning of the word 'super'
    I'm pretty sure we devalued super when the PowerMac G4 [youtube.com] was claimed as a supercomputer all by its lonesome.

    Super is a relative term, what was a super computer is now a computer that I hand-me-downed to my mom so she could check her email and browse the web.

    • it's not a super computer but the UI looks ok.
    • by oneiros27 (46144) on Wednesday October 17, 2007 @11:23AM (#21011005) Homepage
      It met the government's definition of super computer at the time. (1.5Gflops ... well, technically, 1500MTOPS).

      The designation is part of the "Dual-Use" restrictions on exports (basically, things which could be used for both military and non-military applications).

      The 1Gflop threshold was set as the necessary processing power to calculate balistic trajectories for missile systems.

      I can't find the documentation, but my understanding is that the current threshold is 190Gflop (since Jan 2002).
    • I think some people in the press referred to the Intel 286 processor as a supercomputer on the desktop, but I don't know if such terminology was ever used for marketing purposes.
  • by eaglesnax (238705) on Wednesday October 17, 2007 @10:58AM (#21010581)
    Now we know who bought all the PS3s!
  • devaluing super (Score:4, Insightful)

    by mihalis (28146) on Wednesday October 17, 2007 @11:00AM (#21010607) Homepage

    Well the guy used to use a 200-node parallel supercomputer, but now he prefers to use 8 PS3s. That to me proves that 8 PS3s is like a supercomputer TO HIM.

    I'm sure there are faster setups available if had the money, but 100% of 8 PS3s indefinitely is preferable, from what he says, to the costly little slices of "real" supercomputers he tried to rent before.

    I wonder if Sony could offer a "HPC PSP3" which provided a stripped down processor board without the shiny case, graphics memory etc. It would be interesting if the Cell processor could get better economies of scale.

    • Re:devaluing super (Score:4, Insightful)

      by TargetBoy (322020) on Wednesday October 17, 2007 @11:03AM (#21010665)
      Wouldn't it rather be IBM that might offer this, since they actually make the cell?
      • Re:devaluing super (Score:4, Informative)

        by mihalis (28146) on Wednesday October 17, 2007 @11:48AM (#21011425) Homepage

        Wouldn't it rather be IBM that might offer this, since they actually make the cell?

        Yes, actually I think you are correct. If I recall correctly it's Sony, IBM and Toshiba in the cell consortium, and the most ovious vendor of a "compute-node Cell module" would indeed be IBM, not Sony, good point.

        By the way, I had a typo, it would not be an "HPC PSP3" of course, the Cell is way too hot and power hungry! Although ... of course with sufficient shrinks and price reductions the current Cell might well one day be in a portable game console. Then we could have another round of speculation on personal clusters. I love the "wheel of reincarnation" in digital technology!

      • Re: (Score:3, Informative)

        by wilsonjd (597750)
        They already do, but it is a bit more expensive than 8 PS3s: http://www-03.ibm.com/systems/bladecenter/cell-based.html [ibm.com]
    • by akb (39826)
      Even more impressive is the monetary savings he achieved. The 8 PS3's cost less than even one of the 200 nodes he was using. That's a 99.5 percent reduction in cost without even considering power, cooling and networking.
    • Re: (Score:2, Informative)

      by Ant P. (974313)
      You mean like this? [ibm.com]
      • Re: (Score:3, Insightful)

        by mihalis (28146)

        You mean like this?

        Thought I replied to this, but can't see it.

        Anyway, yes, that's jus the ticket, except it's $19k!!

        All of a sudden racking up actual Sony PS3s with their curved shiny cases, graphics chip etc seems eminently sensible

  • You haven't read the Sony press releases about how powerful the PS3 cell processor is.
  • Defining "super" (Score:3, Interesting)

    by NeilTheStupidHead (963719) on Wednesday October 17, 2007 @11:04AM (#21010689) Journal
    How do we define a 'super' computer?
    Is it simply FlOps? Then at some point, every computer will be a super computer unless you scale the amount of operations with the speed of computers
    Is it the 'classical' image of a huge room of boxes chugging away? Then as individual computers get faster and smaller, these rooms will be filled with more computing power as time goes on.
    What about parallel processors? The PS3 has some form of parallel processing capability as I understand, so linking eight together isn't just 8 parallel processes it's 8*(parallel processes in one PS3)

    Since some 'super' computers of ages past have less power than some modern desktops, I think that the first is more likely if you scale the threshold of a 'super' computer, e.g. the fastest 1-2% of computers out there. More generally, I think that most people conveice of a super computer being any computer system that can perform tasks that would take an unreasonable amount of time on a single, off-the-shelf machine.

  • by Intellectual Elitist (706889) on Wednesday October 17, 2007 @11:13AM (#21010829)
    > Since nobody is actually playing games on the system

    You can always count on Slashdot for a fair and balanced presentation of information.

    Funny that I've bought 4 disc-based games and at least one downloadable game since the beginning of July, and have been using my PS3 almost exclusively for gaming since then. I'll be buying at least 4 more games before the end of the year, too.

    Sorry to burst your bubble, but the PS3 game drought has been over for a while now...
    • Re: (Score:3, Insightful)

      by Jarjarthejedi (996957)
      Yeah, but in this case it's not /. doing the PS3 bashing, it's the actual article from Wired instead, just look at the first sentence

      "Suffering from its exorbitant price point and a dearth of titles, Sony's PlayStation 3 isn't exactly the most popular gaming platform on the block."

      Looks like /. isn't the only PS3 hating news source out there eh :P
      • You guys are saying two different things. That the PS3 isn't the most popular gaming platform is demonstratebly true. This is not at all the same as saying that the people who have them aren't using them. I know I'm certainly using mine all the time.
    • by suv4x4 (956391)
      Question: why do you read Slashdot.
      • > Question: why do you read Slashdot.

        For the linked articles and the comments from the few people who actually have some insight into the issues at hand.
    • Re: (Score:3, Insightful)

      by Fross (83754)
      8 games? Yikes, either you play *everything* or you've got some real crud in there. Care to elaborate? (I hope one of them isn't Lair ;) )

      To give an idea, the top 8 games on PS3 get metacritic scores of 85 or more ( http://www.metacritic.com/games/ps3/scores/ [metacritic.com] ). Only one of those is over 90.

      To compare, the 360 has *27* games at 85 or more ( http://www.metacritic.com/games/xbox360/scores/ [metacritic.com] ) 9 of which rate 90 or more.

      For me, of those 8 games I'd be interested in 4, 2 of which are also available on PC.

      I'
      • by Intellectual Elitist (706889) on Wednesday October 17, 2007 @12:43PM (#21012315)
        > 8 games? Yikes, either you play *everything* or you've got some real crud in there. Care to elaborate?

        The four games I was referring to were Ninja Gaiden Sigma (88 [metacritic.com]), Skate (85 [metacritic.com]), Stuntman: Ignition (75 [metacritic.com]), and Warhawk (84 [metacritic.com]). The downloadable game was Super Stardust HD (84 [metacritic.com]). None of those games are even remotely close to "crud".

        The four games I referred to having an interest in purchasing before the end of the year are Ratchet & Clank Future: Tools Of Destruction, Uncharted: Drake's Fortune, Rock Band, and the collector's edition of Stranglehold. I might also consider Army Of Two, Assassin's Creed, and Call Of Duty 4, depending on the reviews.

        > To give an idea, the top 8 games on PS3 get metacritic scores of 85 or more [...] Only one of those is over 90. To compare, the 360 has *27* games at 85 or more [...] 9 of which rate 90 or more.

        The original post had nothing to do with the 360 -- it was about the insinuation that no one uses the PS3 for gaming, which is ridiculous.

        You're also making an apples to oranges comparison, because the 360 has been out longer and has a much larger base of titles. But if you want to compare, as of October 13th Metacritic's aggregated ratings for the 360 [metacritic.com], PS3 [metacritic.com], and Wii [metacritic.com] show that the 360 has 264 rated games, the PS3 has 82, and the Wii has 87. Since the PS3 and Wii came out later than the 360 and around the same time as each other, this makes sense.

        If you look at the percentage of each console's library that has a metascore of 75 (out of 100) or higher, the PS3 leads with 54%, followed by the 360 at 44%, then the Wii with only 16%. If you go with a metascore of 80+, the PS3 has 34%, the 360 has 27%, and the Wii has only 8% above that level. At 90+ the Wii has 3%, the 360 has 3%, and the PS3 trails with only 1% of its library at that level.

        Going by percentages, the PS3 and 360 libraries are of roughly equivalent quality, while the Wii's lags far behind.

        > the general sentiment is the PS3 needs a killer app (like a halo, gears of war, or some other really good exclusive title) to make it worth getting.

        The general sentiment is also that Iraq was involved in 9/11 and that Britney Spears's personal life is somehow newsworthy. I'll think for myself, thanks.

        That said, every console gets a "killer app" eventually. I'm sure the inevitable God Of War III will fill that void if nothing else does beforehand.
    • by aztektum (170569)
      If you want Fair and Balanced, you shouldn't be reading /.

      I hear Fox News is the place for that.

      Only one of those statements is intended as sarcasm, but I'll let the reader decide which :)
  • Memory limitations (Score:4, Interesting)

    by Kupek (75469) on Wednesday October 17, 2007 @11:13AM (#21010845)
    I'd like to know more details about his code, because a PS3 only has 256MB of RAM. That's a serious performance obstacle, since most high performance applications that do anything interesting need much more than that. I know it's a problem our group has had, and we've heard the same from others.
    • by Joce640k (829181)
      That's main RAM. The individual cell processors only have 256kb each (IIRC) for holding their data.

      • by Kupek (75469)
        No, each SPE has 256kb of local storage. In general, a Cell processor has 8 of them, but in the PS3 only 6 are usable. But that has nothing to do with my main point; the 256k of local storage for each SPE is a problem, but you can code around it. (It's not trivial, but it can be done.) You can't code around having a small amount of RAM and still maintain high performance.
        • by chrysrobyn (106763) on Wednesday October 17, 2007 @11:39AM (#21011267)
          256k of local storage for each SPE is a problem, but you can code around it. (It's not trivial, but it can be done.) You can't code around having a small amount of RAM and still maintain high performance.

          Let me see if I get this straight, you can imagine a piece of code that doesn't mind churning on itself within 256KB, but you can't imagine having to keep 256MB of main memory fed from a network or disk? In my experience, any piece of code that can both benefit from extreme parallelism and fit both the code and enough data to be worth working on within 256KB can handle a few reads from a disk or the network once in a while. If it can't, then 256KB of memory isn't enough to keep the (sub)processor fed, and you need a machine with more on-die memory (many of which can be found).

          Cell is very good at integers and single precision floats for workloads that are parallelizable and fit within 256KB. If you stray from any of that, there are plenty of interesting competitors.

          • by Kupek (75469)
            The transfer time between the SPEs and main memory is many orders of magnitude smaller than between main memory and disk. Further, communication between SPEs and main memory can be overlapped with computation, completely covering the latency.

            Keep in mind that the SPE's local storage is basically a software managed cache. So your argument of "churning on itself within 256kb" would also apply to an L1 or L2 cache.
        • You can if you need to do to a lot of calculations involving a small amount of data. (So that it all fits in RAM, or at least you can batch large portions of it into RAM at a time for extended periods.)
        • Re: (Score:3, Insightful)

          by lena_10326 (1100441)

          You can't code around having a small amount of RAM and still maintain high performance.

          I wouldn't agree with that. That's only true if the algorithm relies on access to the entire data set because it requires random access or multiple table scans. Lots of algorithms can operate on small independant chunks or can be rewritten to use sequential data access, which is chunk friendly. I think it's apparent his algorithm works on small chunks due to the relatively small amount of RAM, unless his entire data set

  • I can't think of anything clever to say other than the subject, though...
  • My old xbox is now a media centre, so why not use a set of consoles for data crunching. It's all just math and these things are optimized for it.
  • by LoudMusic (199347)

    I think we are de-valuing the meaning of the word 'super' :)
    I believe it was originally Apple who brought on the devaluing of the word "super" with their "The PowerMac G4 is more powerful than a Super Computer" campaign. Sure it is - your 1999 desktop computer is more powerful than ... a super computer from 1983. Congratulations! You're only 16 years late.
  • Only 256 Megs of RAM (Score:2, Informative)

    by hweimer (709734)
    The CPU power of the PS3 is indeed very impressive, however, for most real-world supercomputing tasks the 256 MB RAM per node are way too low. One Gig per core should be the minimum, meaning you would have to increase the amount of RAM in the PS3 by a factor of 24.
    • It is a shame IBM and/or Sony haven't released a box that is essentially a PS3 without the pretty case, blue-ray, the graphics chip, PS2 emulation chip, game DRM, the game OS and with lots of RAM.
  • When do i get the cell on a standard ATX form factor with an open BIOS for personla Linux home use?

    Any plans for that?
    • Re: (Score:3, Informative)

      by for_usenet (550217)
      Check out Mercury [mc.com] systems. These aren't exactly for home use, but they are shipping Cell + Linux computer systems. However, given the economics, it might just be better to go with a PS3, if you can live with the memory limitations. There may be other companies, but checking at TerraSoft [terrasoftsolutions.com] (one vendor for PowerPC Linux software and hardware) takes you right back to IBM and Mercury hardware.
  • Disclaimer: I hate the PS3 (though I love the cell, but not for gaming, because that's too complicated for most game programmers to handle). I love my XBox 360 and Wii (as long as they both continue to function and don't break).

    Since nobody is actually playing games on the system, it makes sense to use them for research projects like this

    Yes, because ~4 million people count as "nobody". But seriously, am I the only one that's tired of troll article summaries around here? It's either a flippant comm
    • by trongey (21550)

      ...Slashdot is where i go for excellent commentary ...

      Wow. I can't imagine the level of frustration you must feel.
  • by Midnight Thunder (17205) on Wednesday October 17, 2007 @11:33AM (#21011165) Homepage Journal
    Looking at this page:

        http://www.answers.com/topic/supercomputer?cat=biz-fin [answers.com]

    they define a 'supercomputer' as being "A mainframe computer that is among the largest, fastest, or most powerful of those available at a given time". This is suitably vague, since the point of reference changes all the time. On the other hand there is no point of reference in the definition. For example, does it have to be in the top 100 or 100x more powerful than the current top of the line PC? Without a suitable reference point anyone could call their cluster amongst, the "largest, fastest or most powerful".
    • by jvkjvk (102057)
      Well, that definition isn't going to work very well unless we get to redefine mainframe computer , is it? Clusters are definitely out then, so what does that leave use as far as 'supercomputers'?

      Other than that, the defn' seems pretty good. I don't get this need to quantify so precisely. "Among" seems a perfectly good term.

      For example, let's say you have a bona fide super computer, - one of the top 100. A guy down the street has #193 but it doesn't make much sense to say that his isn't a supercomputer u
  • That is 50 to 500 teraflops in 2007. Everything else is a "last generation" supercomputer and marketing noise. My cell phone is as fast and has as much memory as a 1970s Cray supercomputer (60 MFlops).
  • 1.2 TFlops (Score:5, Informative)

    by rockmuelle (575982) on Wednesday October 17, 2007 @11:36AM (#21011201)
    8 PS3s gives you 1.2 teraflops of single-precision performance or a similar number if you stick to integer operations (6 SPUs/PS3 gives ~150 GFlops). 1.2 teraflops is a supercomputer in my book.

    Using Jack Dongerra's single-precision algorithms that do half the work in single and the other half in double precision, you can maintain a high level of performance and precision. And, the unique architecture of the Cell opens up some interesting algorithmic research issues, allowing scientists to publish twice for the same work: once for the science results, once for the computer science results. :)

    On the flip side, the Gigabit ethernet on the PS3s isn't really 1GB - the PPU can barely keep up. So, extra care must be taken around communication points. And, a similar Intel/AMD-based rack would run about $20k and is much easier to develop for, so if your labor is expensive (i.e., you're not in academia), PS3 clusters may not make much sense.

    -Chris
    • The simple solution is to use the "real" [ibm.com] thing from IBM. All the power and performance of the 2 Cell processors instead of Sony's 1 on the PS3 and 2GB of RAM (1GB per core) in a high density blade format. Subsequently you are not wasting a core on the PS3's OS running underneath and you get a real NIC.
  • but I really wonder now what is defining 'Supercomputer'... I mean, a hundred PS3s sure, but 8? I think we are de-valuing the meaning of the word 'super' :)

    Well, I don't know what qualifies as a supercomputer nowadays ...

    But, to some of us, any computer made in the last decade at one point would have qualified as a supercomputer. I seem to recall any machine which had > 1GHz of CPU speed used to be classified as munitions grade equipment and illegal for export. Something to do with being able to design

  • I wonder if the US governments' restrictions on exporting super computers covers game consoles as well. If it does the Xbox 360 is most certainly restricted.

    (not that it is really a super computer, but if you have ever had to deal with said restrictions you know that 10 year old desktops are considered supercomputers by the US gov)

  • by gamer4Life (803857) on Wednesday October 17, 2007 @11:58AM (#21011591)

    Since nobody is actually playing games on the system..


    I do work web site administration for a non-profit organization and it's amazing how much we'll bend backwards to accommodate the views of our sponsors. If a sponsor gives us money, we'll be sure to remove a reference to another organization, just to appease them.

    Since Microsoft buys lots of ad space across many Internet sites, including this one, it's no surprise that many of these sites will put an anti-Sony spin on their "news".

    These sites will call the 40GB PS3 "gimped", while calling the 360 Arcade "a deal", as well as other hypocritical bs.

    When your income depends on advertising money, you'll do whatever it takes to appease your sponsors.
  • by Shadow Wrought (586631) * <shadow...wrought@@@gmail...com> on Wednesday October 17, 2007 @12:05PM (#21011729) Homepage Journal
    Is it faster than a speeding bullet?

    Nope. It is far too unaerodynamic to reach such speeds without a prohibitive amount of initial energy. Certainly not unassisted.

    Is it more powerful than a locomotive?

    While it concumes about the same amount of raw fuel, it produces far too little in the way of mechanical enregy to pull even a single model RR caboose. Amtrak found this out to their chagrin.

    Is it able to leap tall buildings with a single bound?

    While it does acheive a much heralded TeraFLOP, it turns out that that word does not actually mean "hitting the Earth" as a casual guess at its derivation might assume. So, in a nutshell, no jumping, buildings or otherwise, without significant assistance.

    Finally, does it fight for Truth, Justice, and the American Way?

    The ultimate in guileless parroting, it will simply display whatever it is told, and will never consider the veracity of the content before micrying it. Justice is a bit trickier in that there is little about Justice that is agreed upon. Once GTA IV comes out however, there will no longer be any support for the notion of it supporting even justice with a little "j." As for the American Way... Well it does favor style over substance with an arrogant belief taht it will be Bought because it is Made. Which is about as close to the American Way these days as anything else. call it one out of three.

    So, no, I would have to say that it would not qualify, in any quantity, as a "Super-Computer."

  • If they just built a box with say, four NV G80 based computational engines (128 stream processors per card with thousands of threads per procesor) and used CUDA, even that would have been more impressive.
  • Wild guess here but it probably really is computationally faster than the what was the fastest computer in the world of say 20 years ago.
    Rather than a sense of the word 'supercomputer' being devalued, maybe its definition just needs to keep up with the times.
  • Good call (Score:3, Insightful)

    by styryx (952942) on Wednesday October 17, 2007 @12:25PM (#21012073)
    I had a look at using multiple PS3s for simulations a while ago. Purely based on the ass-rocking-ness of the CELL chip.

    There are servers that use the CELL chip, from IBM, see the Blade server. [wikipedia.org] But the Blade server is quite a bit expensive; that is 8 PS3's at the UK price was cheaper the last time I looked. On top of all that is the 'pooling' that the CELL chip does, while this won't be that good for simulation (with current, popular implementations, e.g. MPI2), it will be awesome for games: succinctly, any process that requires extra 'power' can request another node from the 'pool' and release it back when it is under less strain. The transport latency (often the biggest latency in Parallel, even with fibre optic switches, unless its a purely Monte Carlo sim...) is much reduced by having all processors on a single die. The architecture is a mix with vector based operations as well.

    Prima facie it would be perfect to use multiple PS3s. After speaking to some HPC chaps, at Edinburgh Uni,they informed me that the memory on the PS3's is pretty low (512MB split between video and the conventional) which can be a pain if you want to perform REALLY big simulations (which, when scaling is accounted for, is pretty much the point of using supercomputers... not _necessarily_ speed, lets not make this the point of debate, it is simulation dependent.). I will also add that the memory, though small, is bloody fast. If you can code to keep bloat completely removed, you won't need many BG processes; and split memory requirements between each of the PS3s then it is a really, really nice system. Takes a bit of effort and a learning curve, but there are many resources online, native Linux support is an Uber Bonus for Sony (though I am considering NOT buying a PS3, or many, due to their Media departments behavior!).
  • It also devalues the meaning of "inexpensive" if you ask me...

What this country needs is a good five dollar plasma weapon.

Working...