Slashdot Mirror

← Back to Stories (view on slashdot.org)

Largest Privately Owned Supercomputer

Posted by timothy on from the beats-the-basement-beowulf-rack dept.

GORMUR writes "IBM has launched its Watson Blue Gene system, the largest privately owned supercompuer seen by the press. The super computer is described reaching a whopping 91.29 teraflops. IBM has plans on giving Academic researchers access to some computing time. Some more info can be found the IBM site. All this makes you wonder what other supercomputers are out there, not known to the press, and if it's time to increase the size of your private key and strengthen your encryption."

17 of 292 comments (clear)

  1. NSA... by ThomasFlip · · Score: 5, Interesting

    I think it's safe to say that the NSA, with it's largest budget out of any intelligence agency in the U.S, has probably cracked the 100 TF mark ? It's a shame we will never no what kind of muscle they can flex.

    --
    If the dollar is an "I owe you nothing", then the Euro is a "Who owes you nothing." - Doug Casey
    1. Re:NSA... by slavemowgli · · Score: 3, Interesting

      More on this here; also see Simon Singh's The Code Book: The Science of Secrecy from Ancient Egypt to Quantum Cryptography, which, IIRC, has a section about this.

      --
      quidquid latine dictum sit altum videtur.
    2. Re:NSA... by Anonymous Coward · · Score: 1, Interesting

      I've heard 2 stories about the NSA.

      One was from a professor in College. He was a big person in the HPC (High Performance Computing) world. At one point he was on the board of some large public government super computers. He had talked about some 2000+ processor built buy Sun that went into the NSA facility. This was a few years ago, so who knows if it has changed or not.

      Second was maybe the NSA, but I know it was some high clearance government facility. When any of these highly classified places need tech support, it really isn't the easiest thing in the world to do. Any company that sells stuff to these organizations has to employee people with clearance high enough to enter the buildings. Plus, if you are going in to debug some hardware/software problem, you can't bring that many tools with you. You are not allowed to leave with any kind of electronic that you bring into their buildings. Most of these tend to be incinerated. Also, when you enter secure rooms, there are some big red lights above the doors that go off telling everyone else in the room that someone is entering the room that isn't allowed to see anything (ie: the tech support guys).

      They also won't let you work on the equipment directly. They bring it to you in a small room... really interesting stuff.

    3. Re:NSA... by Anonymous Coward · · Score: 2, Interesting

      I seriously doubt NSA can secretly keep a state-of-the-art processor fab running.

      Me too.

      It doesn't seem to be secret at all.

  2. SHA Collisions by Anonymous Coward · · Score: 2, Interesting

    People using 256-bit encryption algorithms should be safe for now, given the massive amount of computations needed for key exhaustion. However we should be working on implementing SHA-512 as soon as possible as it will soon become trivial to find collisions in SHA1

  3. What you can't buy with money by antispam_ben · · Score: 4, Interesting

    The NSA is the single biggest employer of mathematicians in the world, and it's probably safe to say that they are at least a couple of years ahead of the rest of the world as far as cryptography and cryptanalysis is concerned. ... but you can't buy advances in mathematics with money.

    Then what do they use to pay their mathematicians? Coffee?

    --
    Tag lost or not installed.
    1. Re:What you can't buy with money by Mao · · Score: 3, Interesting

      I do want to point out that if you are a mathematician with the ability to seriously advance mathematics, it would totally suck to not be able to publish any of your major results. If a high stature mathematician is willing to work for NSA and risk not being able to publish work which he/she has done in his capcity as a NSA researcher, he/she most likely is in it for more than just money.

      I do wonder, suppose some NSA guy proves the Riemann's hypothesis. What would they do? How far does patriotism go?

  4. Nice! by Duncan3 · · Score: 2, Interesting

    Almost 1/2 a Folding@home, I'm 1/2 impressed ;)

    Holy interconnect batman!

    --
    - Adam L. Beberg - The Cosm Project - http://www.mithral.com/
  5. DC? by ilyanep · · Score: 2, Interesting

    How come nobody counts Distributed Computing as Supercomputers? I'm sure many of the BOINC Projects (SETI@Home at berkeley, E@H at UWM, etc.) have close or even higher than that.

    --
    ~Ilyanep
    To get message, take amount of carrier pigeons at each stage mod 2. Then decode binary.
    1. Re:DC? by mvdw · · Score: 2, Interesting

      Perhaps a decent measure would be to average the processing out over a week or so; eg each seti@home unit is, say, 1e9 floating point operations, calculate how many units are processed in a week, divide by seconds in a week, there's your number. This method allows for the redundancy of the @home method, ie, each unit will be computed a number of times, if only the completed units are counted it gives a measure of true (sustained) performance.

  6. And what makes you think... by PaulBu · · Score: 2, Interesting

    ... that NSA would be interested in teraFLops? Last time I checked, their kind of processing required manipulating bits (in weird ways), not imprecise floating point numbers. Go to DOE to pay for FLOPS...

    Not that I'd know, but I can still guess... ;-)

    Paul B.

  7. Re:sure. by St.+Arbirix · · Score: 2, Interesting

    I really don't see the use of a supercomputer in the establishment of botnets. Since when is the creation of a zombie node dependent on functional security? Botnets are built on security vulnerabilities. The supercomputer attack would only be useful on a legitimately high strength encryption system.

    So, yes, no one interested in encryption would be involved in setting up a botnet.

    And when I say I make up my own encryption I'm not saying it's all that exciting or wonderful, it's just not what anyone would expect to be dealing with and so they wouldn't know how to attack it. For example, take a binary and replace every fourth byte with a random number and then dump the pulled bytes in reverse order at the end of the file. I just made it up, but I'd bet it'd confuse people.

    --
    Direct away from face when opening.
  8. "Strengthen(ing) your encryption" by Anonymous Coward · · Score: 1, Interesting
    All this makes you wonder what other supercomputers are out there, not known to the press, and if it's time to increase the size of your private key and strengthen your encryption.

    So instead of taking 3 billion years for all the known supercomputers to factor my 2048-bit RSA key, it will only take 2.5 billion years.

  9. Re:You mean GCHQ by finkployd · · Score: 2, Interesting

    Both did, and I doubt we will ever really know which did it first (we know when GCHQ did, not when the NSA did).

    Finkployd

  10. Wonderances by GoClick · · Score: 2, Interesting

    No but I bet it would only take it a day or two to compile it!

    Ever wonder how much processing power Google has between all of their systems and all of the Google tool bars running around?

    Has anyone ever wondered if MS or Yahoo has tried or is currently using their various browser bars to provide distributed computing?

    Has anyone ever wondered if they buy insurance on these things for stuff like faulty processor design? Like the Pentium bug? I mean how'd you like to build this thing and the find out all of the processors have a bug?

    Has anyone wondered if you have software on your machine that fouled a browser bar's data somehow if you're responsible?

  11. Re:Old Supercomputers by joib · · Score: 3, Interesting

    Well, the Cray T3E they used to have at the supercomputer center where I submit my stuff was dismantled and the pieces thrown into a big trash bin in the yard. *sniff*

    The life of a supercomputer is AFAIK really closer to 5 years than 10. It's not that they aren't impressive machines even 5 years old, it's just that they use _lots_ of power and floor space. Looking at how much computing per $ you can do, it's just cheaper to replace them with something new than to keep them running.

  12. Minor Mistake... by Bad+to+the+Ben · · Score: 2, Interesting

    Actually, the NSA didn't beat Diffie and Hellman to the punch, it was the Government Communications Headquarters (GCHQ) of the United Kingdom, in particular a man named James Ellis. It's mentioned in the "Science of Secrecy" book you linked to, page 166.

    I don't even think the NSA was around back in the 60's when this was going on.