Forgot your password?
typodupeerror
Apple Businesses

More on Virginia Tech G5 Cluster: 17.6 Tflops 390

Posted by CmdrTaco
from the thats-a-lotta-flops dept.
daveschroeder writes "BBC World's Click Online has a video report (with text transcript) on Virginia Tech's new 1100-node dual 2.0 GHz G5 Terascale Cluster. The report quotes the performance as 17.6 Tflops. As a point of reference, the cluster would be number 2 on the most recent June Top 500 list, behind only Japan's Earth Simulator, and considerably more than doubling the performance of the current number 3 1152-node dual 2.4 GHz Xeon MCR Linux cluster. Assuming the performance figure accurately reflects the LINPACK score (which it should; since the deadline for submissions for the upcoming list of Oct 1 has already passed, one would imagine VT would quote that figure), and depending on new entries for November's upcoming list, the cluster should almost certainly rank in the top 5 - all for only US$5.2 million. The video report is available in Windows Media 9 and Real formats; the relevant portion starts at 13:00."
This discussion has been archived. No new comments can be posted.

More on Virginia Tech G5 Cluster: 17.6 Tflops

Comments Filter:
  • by Mwongozi (176765) <(gro.revolgdivad) (ta) (eerhthsals)> on Sunday October 12, 2003 @11:07AM (#7194244) Homepage
    You can watch just the report itself, no skipping required, by following the links on this page:

    http://www.bbcworld.com/content/template_clickonli ne.asp?pageid=666&co_pageid=3 [bbcworld.com]

    • Windows Media and Real Player about a G5 cluster? Don't you mean a nice Quicktime movie? (yes I have both players on Mac, sans spyware and other undesirable "system integration". But Apple will be sad)
  • Heist (Score:2, Funny)

    by DarkHazard (713597)
    Surley they only need 1099 G5s.
  • Yes, but, (Score:2, Funny)

    by Anonymous Coward
    can it defeat an iMac in Apple's Photoshop benchmarks ;-)?
  • Twice as fast...? (Score:3, Interesting)

    by suwain_2 (260792) on Sunday October 12, 2003 @11:14AM (#7194278) Journal
    ...considerably more than doubling the performance of the current number 3 1152-node dual 2.4 GHz Xeon MCR Linux cluster.

    If I understand this correctly, it's saying that a G5 is more than twice as fast as a dual 2.4 GHz Xeon? (1152 dual 2.4 GHz Xeons vs 1100 dual 2.0 GHz G5s -- there are fewer G5s and they run at a slower clock speed.)

    This is a pretty staggering statistic. I hadn't really believed the hype about how fast the new G5s were.
    • Re:Twice as fast...? (Score:2, Informative)

      by adam872 (652411)
      The two clusters are different enough that making accurate comparisons is difficult. The new G5's have a more recent PCI architecture, they use Infiniband as the interconnect and it's possible that they made use of the AltiVec (though I hear that this may not be the case because of 32 bit limitations). I believe none of these apply to the Xeon's. In high speed computing, the interconnect is vital, so that alone may push this cluster ahead for the time being. I don't doubt that the individual G5 processor ar
    • As a lot of others have pointed out, there is a lot more to "speed" that the frequency of the clockchip. I use a lot of IBM "P-series" (what used to be called RS/6000) machines at work. The clock speeds are generally low (sub 1Ghz) but in certain situtations these machines absolutely smoke. In other situtations, such are running Java, they drag ass. A lot has to do with how well the application using the processor takes advantage of the low-level capabilites.

      Think of computers like cars. The Honda S20
    • Re:Twice as fast...? (Score:3, Informative)

      by mangu (126918)
      17.6 Tflops in 2200 processors results in 8 Gflops/processor. I don't know about the Xeon, but I have benchmarked my own 2.4 GHz Pentium4 at 6 Gflops, multiplying two 1000x1000 random matrices using Lapack. So, yes, 8 Gflops at 2.0 GHz is faster than 6 Gflops at 2.4 GHz, but only slightly. Also, there is the overhead in the matrix multiplication. The peak performance in the 2.4 GHz P4 would be 9.6 Gflops, so one can say there's no magic other than Apple marketing in the G5. The diference in performance betw
  • by slyfox (100931) on Sunday October 12, 2003 @11:46AM (#7194375)
    The new "top 500" list will be announced right before SC2003 [sc-conference.org] and discussed in detail at a session of SC2003 on November 18 [sc-conference.org].

    Look for another (less speculative) story on Slashdot around then.
  • What's been left behind is wether or not these systems are using Panther as the OS. It would seem that with this kind of performance, an Apple supplied OS -- as opposed to Yellow Dog would -- only be capable of performing well on the G5 since Panther has processor optimizations for the G5.

    If the original XServers were too costly and low performance (since they came with a G4) wouldn't a G5 server (since the performance is apparently much better) be a great option for small/medium size businesses for a web/
    • -- as opposed to Yellow Dog

      This bothers me as much as people synonymizing Red Hat with linux for x86, and Internet Explorer with the internet; and yes, you sound just as dumb when you jump into a discussion on PowerPC based supercomputers running linux and equate the entire PowerPC world with Yellow Dog as somebody who jumps into a conversation about the internet and equates it with AOL version 7.0.

      Just FYI, Red Hat is the only major linux vendor out there that doesn't support PowerPC. Very few people u
      • In a recent interview (in German) [berlinonline.de] (actually in a future one, since it's dated 13th of October 2003 ;-), Suse Boss Richard Seibt says "In the future, Computers from Apple will run Linux. I can envision that in 10 years their comfortable user interface will be standard on all Linux systems."

        No idea why I brought that up now ;-)

    • No, the G5s shipped with 10.2.7, it's recompiled for the G5 to take advantage of the new architecture - should be reasonably optimized although that probably isn't a priority yet - and has some 64-bit support. Yellow Dog Linux is ready for the G5, but their support is only as good as gcc at this point.
      • No, it's not. Yellow Dog is /not/ shipping a G5 Linux, and as far as I know there's no distro out there that does. (The earlier report of Gentoo booting was a hoax.)
    • by daveschroeder (516195) on Sunday October 12, 2003 @12:20PM (#7194542)
      The project leader, Dr. Srinidhi Varadarajan, will be speaking at a session entitled Building Virginia Tech's G5 Supercluster [oreillynet.com] on Oct 28 at the upcoming O'Reilly Mac OS X conference [oreillynet.com].

      He'll probably reveal some of the technical details, such as the version of Mac OS X used, at that session.

      Also, according to a blog [oreillynet.com] at O'Reilly:

      Next year, all the little known details [about the cluster] will be revealed in a new book. By that time we'll know what the project means for supercomputing and for Apple.
    • If they shipped with 10.2.7, can they buy the Panther family pack? [apple.com] Or 220 family packs?
      • No, the family pack is only for home users; commercial and education users need to look elsewhere. Likely, Apple gave them a special contract(especially if they built a customized version of the OS) which cost a fair bit more.
        • which cost a fair bit more

          Or less...

          The power of this err... is nothing compared to the power of the force (marketing).

          Just think of all the free publicity they have got so far.
          • True, but still, the education/government prices are nowhere near the family pack prices. The family pack is $40/machine; no education contract I've seen gets below 2x that amount, although the contract does stipulate 3 years of free upgrades.
      • According to Apple, **all** G5 systems will get Panther free (well, $19.99 for a shipped CD) - they've been grandfathered into the program:

        When you fill out the form to check your G5 serial number, be sure to leave out any dashes.
    • They had a 1st of October deadline to meet for the cluster. If they had waited for Apple to bring out their G5 XServe, then they would have missed the deadline. The G5 towers were available and had been tested to a fuller extent, and they probably worked out to be cheaper. If you look at the photos, on the site, then you will notice that space was not an issue, so enclosure size was probably not important.

      I would not be surprised if the first generation of G5 XServes are 2U cases, so that there is enough s
  • MCR uses Quadrics - What does this cluster use?
    http://doc.quadrics.com/Quadrics/QuadricsHom e.nsf/ DisplayPages/7C18E51DBC215D3E80256D5900535959
  • Could someone who knows the going rate of these top 5 supercomputers please tell me how much less expensive 5.2 million is? I know that it sounds like a lot of money to me, but I'm guessing that it is orders of magnitude cheaper than the other top computers.
    • From this article: zdnet.co.uk [zdnet.co.uk]

      "Sun Microsystems, for instance, is designing a supercomputer under a $50m (31.6m) grant from the US Defense Advanced Research Projects Agency (Darpa) that will contain 100,000 processors, according to Jim Mitchell, who heads up Sun's labs. The whole thing could conceivably fit into a room."

      5.2 million is a lot less then 50 million.

  • I didn't see anywhere the source of these benchmarks - I believe what they have done is multiply the SPECfp_rate_base2000 for a dual processor 2GHz G5 (which according to Apple is 15.7) [apple.com] and multiply it by 1100.

    The thing is, that only comes to 17.27TFLOPS, and in addition it does assume that the original spec scores were accurate. [theregister.co.uk]

    Would anyone care to shed some light onto this?
    • Would anyone care to shed some light onto this?

      I can shed light to this extent: a linear scaling between processors and processing power is only realized in the most idealized of situations (those known as 'embarrasingly parallel'), where each job is small and completely independent of other jobs. The funny thing about embarrasingly parallel tasks is that they do not need a fancy parallel computer; they can just as easily be accomplished on N separate 486 machines, if N is sufficently large.

      The upshot

    • by daveschroeder (516195) on Sunday October 12, 2003 @12:40PM (#7194634)
      Easy: you yourself point out that 1100 * 15.7 = 17.27 ... not 17.6.

      Since the call for papers for the new Top 500 list was Oct 1, and the BBC show aired on Oct 9 with a companion BBC News story [bbc.co.uk] dated Oct 12, you'd hope that VT was simply regurgitating the figure that has already been sent to the Top 500 organization.

      And why are you trolling around with one of those super-old benchmarking stories? We've already established that every manufacturer does what they can to show their products in the best possible light. At least Apple documented their test [veritest.com] results [apple.com] and methods in full.

      So acually, your logic doesn't make any sense: you jump to the conclusion that it's not real results - even though real results already exist and have been submitted, and the entire story is pretty much about that process, making performance figures a critical piece to get accurate - and that they must have just multiplied some benchmark number by 1100. Then, even though the subject of your own post indicates your recognition that "it doesn't add up", you still apparently assume that the results are somehow doctored, this time for the worse, and you manage to weave in one of the stories that tries to make it look like Apple lied with its benchmarks - which it didn't - which is unrelated to the current issue! How does it "assume" the original scores were accurate?? YOU are assuming that they're just multiplying. You might have been onto something if the multiplication actually came out, but it doesn't, meaning that is NOT what they did.

      Bravo, +1 Troll.
  • http://www.bbcworld.com/content/template_clickonli ne.asp?pageid=666&co_pageid=20 [bbcworld.com]

    Skip ahead to 13:00

    Since the freakin' Windows Media files won't play on OS X.

    simon

  • Interesting math (Score:3, Interesting)

    by lexcyber (133454) on Sunday October 12, 2003 @01:11PM (#7194783) Homepage
    The VT cluster cost about $5.2 Million and get approx. 17TFlops - The NEC Earth Simulator gets 35TFlops and cost one billion dollars. That makes it 192 times more expencive. So you can build 192 VT Clusters. And then in theory get. 3.2PFlops for the same amount of money. If you detract performance for cable lenght etc. - You will most definitly get around 1PFlops.

    So, you supercomputerusers out there - build a 1PFLOPS cluster NOW!
    • Yeah, but you forget between these 2 computers lie nearly 2 years of moores law, and if i remember correctly ES was more like 360 mio$.

      And 5m Bucks for the cluster? 2000 machines a 2.500$.
      SOmewhere something doesnt fit there. Sure, they have gigabit ethernet on board, but the switching architectuer alone should cost at least a few millions.
      Not to mention the building.
      • And 5m Bucks for the cluster? 2000 machines a 2.500$.

        First of all, the cluster only has 1100 nodes, not 2000. Mind you, the price for the default config is $2,999 which gives us $3,298,900.

        On the other hand, the education price is 2,699, giving us $2,968,900.

        Mind you, they probably don't need mice, keyboards, DVD-R drives, or even graphics cards on most of them. In fact they only really need motherboard, CPU, and RAM (although a lot more than the default config), case and a small hard disk. Factori

        • All the machines have 2gb of ram, which is much more expensive than the default config if apple is selling you the ram. I'm sure they got a fat educational discount on top of some other deals for the sheer visibility of such a powerful cluster.
  • Virginia Tech's new 1100-node dual 2.0 GHz G5 Terascale Cluster. The report quotes the performance as 17.6 Tflops.

    Finally I'll be able to play SimCity 4!!

  • by DeeKay (263782) on Sunday October 12, 2003 @02:15PM (#7195047)
    Simple equation: 4 FPops/cycle (IBM-PPCs) * 2GHz * 1100 G5s * 2 CPUs/G5 = 17.6 TFLOPS!

    No *real* Rmax linpack scores are known yet, and from what i figured the submissions on Oct 1st are just for *inclusion* in the list, real Linpack scores can be submitted till shortly before (or even on!) the conference mid-November..

    This article is BS and should be removed...

    P.S.: 4 FPops/cycle per clock with 2 FPUs i hear you scream - Impossible! - That's due the Multiply/Add FMAC thing that counts as 2 FPops!
  • The only skepticism I have regarding this figure (and I have this skepticism as a current G3 owner who's waiting for a dual G5 that had #$%#% better ship by Tuesday like they said it would! ;) is that in the past, Apple has published figures based on single-precision floating point performance, and TOP500 rankings use a benchmark that requires double-precision.

    I had this pointed out to me in May of 2002 when Apple introduced the Xserve... in this Slashdot thread [slashdot.org].

    Apple's page about the G5's execution co [apple.com]

    • You did forgot a factor of 2. First you say

      Each processor can do 2 fpu instructions/clock. So each computer can do 4 fpu instructions/clock.
      • I don't think I missed that factor of two. But let's take it from what you just said:

        Each computer can do 4 fpu instructions/clock.
        (and those are double-precision 64-bit flops).

        There are 1,100 computers. Times 4, that gives us 4,400 flops per clock.

        (Note: given 1 DP flop per clock per FPU, this ties in nicely with my use of 4,400 as the number of FPU's, in my comment.)

        4,400 flops per clock * 2 billion clock cycles per second = 8.8 trillion flops per second.

        I'm still getting the same number

  • I suppose I should mention that the pageid for this thing is '666' :P

This place just isn't big enough for all of us. We've got to find a way off this planet.

Working...