Virginia Tech’s ‘Big Mac’ Apple G5 cluster not on top 10 supercomputers list

“The Japan Marine Science and Technology Center’s Earth Simulator remains the world’s most powerful computer, according to the latest version of an industry ranking… One machine missing from the latest ranking was the cluster of 1,100 Apple Computer Inc. Power Mac G5 desktops built by Virginia Polytechnic Institute and State University. The machine achieved a maximum performance of 10.3T flops to rank third in the previous chart. This time the machine didn’t make the list because it was out of service for hardware upgrades, said the ranking authors,” Martyn Williams reports for IDG News Service.

The full Top 500 ranking can be found online at http://www.top500.org . It is compiled by Hans Meuer of the University of Mannheim, Erich Strohmaier and Horst Simon of NERSC/Lawrence Berkeley National Laboratory, and Jack Dongarra of the University of Tennessee.

Full article here.

26 Comments

  1. Yup that sucks. My dad works at Tech, so I’m not surprised to this – they’re upgrading to Xserves (if you didn’t already know).

    Shouldn’t be too much longer.

    Goooo Hokies!

  2. Did anyone really think Apple in the top 10 was for real. They patched together a system with super glue and duct tape.

    The “real IT world” doesn’t place Apple’s hardware in the realm of supercomputers. It is interesting that everyone assumes that the system is being rebuilt with new “Apple” hardware; their is no direct evidence from VA Tech. to support that claim.

  3. Sputnik-You’re pretty dense. Apple is powered by IBM’s processors. IBM has a system that is 4th in the world. How do you figure that Apple’s hardware doesn’t fit in the “real IT world”.

  4. Yes, a lot of people do think, and rightly so, that Apple in the top 10 is for real… because it is. It’s far from “patched together … with super glue and duct tape”; infiniband is expensive stuff.

    And as for the source to support the claim, read the top500.org news. The new release from Top500 states “VT is replacing hardware and the new hardware was not in place for the compiling of this TOP500 list.” And the Top500 wouldn’t just make up stuff for their news release; it’s obvious that VT told them that.

  5. Sputnik and your experince with the real “IT” world is…
    oh none I see. Considering that Reuters repoted that the Big Mac was being rebuilt with new hardware (a claim that is backed up by both Apple and VA Tech) I know I can safely say that the sack of excrement you pull your “facts” from is bottemless. Why dont you get a real life, and STFU.

  6. Just to inform those who can’t read… SPUTNIK!

    The VT supercluster was not in the running for this listing since it was offline for hardware upgrades. In other words it was in the process of being made MORE powerful and will be in the next lsiting.

    Putz.

  7. Virgina Tech’s is nowhere near the NEC 40 teraflop cluster or the other two fast Japanese clusters, but it was faster than the Dell cluster if memory serves me correctly. Even if/when IBM makes its 360 teraflop machine, this must be said of any computer: Purgamentum init, exit purgamentum [garbage in, garbage out]. Why we need a 360 teraflop garbage processor is beyond me. Consider the human brain….

  8. I’m not surprised at all by this. I read somewhere last week that they (whoever it is who determines the criteria for supercomputer ranking) were going to make changes to how list membership was determined. I suspected right from the start that this was aimed at removing the Mac G5 cluster from the list. I also suspect the prejudice against Apple runs much, much deeper than just a “Wintel in business makes the platform better” bias.

  9. Yeah, I doubt the conspiracy theory, Tom. The changes you’re talking about are not so much about this particular ranking — which involves a test of a specific ability of supercomputers to solve problems — but the addition of seven other tests that are designed to measure the performance of other attributes of a supercomputer. The effect will be that there will likely be 8 tests that will provide numbers that make the “fastest” designation a little less meaningful because not everybody is looking for the same thing in a supercomputer. Some problems require the transfer of massive amounts of data that is easy to process, some problems require the transfer of relatively small amounts of data that require intensive processing. A “slower” computer that is able to better handle huge amounts of data transfer will probably be less expensive to own and operate for the former type of problem than a “faster” computer that is less efficient in transferring data.

    The current test measurement, while no longer exclusive, will likely remain relevant.

  10. sputnik
    “It is interesting that everyone assumes that the system is being rebuilt with new “Apple” hardware; their is no direct evidence from VA Tech. to support that claim”.

    another dumbass comment from Sputnik (no surprise there)!

    From VT’s website
    http://www.tcf.vt.edu/

    I think it’s these words that give it away (bottom of page)

    “Current System Status: Offline for XServe G5 Upgrade”

  11. “… I suspected right from the start that this was aimed at removing the Mac G5 cluster from the list. …” – tom

    It hardly needs anything to remove the G5 cluster from the list. The cluster as we know it does not exist anymore. It was dismantled to be replaced with Xserve G5s. Since the thing does not exist, do you really expect that it ranks in the Top 500? Honestly.

    Relax, it should be back for the November list.

Reader Feedback

This site uses Akismet to reduce spam. Learn how your comment data is processed.