Slashdot Mirror


$24.5 Million Linux Supercomputer

An anonymous reader wrote in to say "Pacific Northwest National Laboratory (US DOE) signed a $24.5 million dollar contract with HP for a Linux supercomputer. This will be one of the top ten fastest computers in the world. Some cool features: 8.3 Trillion Floating Point Operations per Second, 1.8 Terabytes of RAM, 170 Terabytes of disk, (including a 53 TB SAN), and 1400 Intel McKinley and Madison Processors. Nice quote: 'Today's announcement shows how HP has worked to help accelerate the shift from proprietary platforms to open architectures, which provide increased scalability, speed and functionality at a lower cost,' said Rich DeMillo, vice president and chief technology officer at HP. Read Details of the announcement here or here."

211 of 376 comments (clear)

  1. Other OSes by frizz · · Score: 5, Interesting

    What OSes do the other top 10 supercomputers run?

    1. Re:Other OSes by hawkstone · · Score: 5, Informative

      1. IBM ASCI White,SP Power3 375 MHz
      Lawrence Livermore National Laboratory

      It runs AIX.

      2. Compaq AlphaServer SC ES45/1GHz
      Pittsburgh Supercomputing Center

      Haven't used it, but I'm guessing Tru64.

      3. IBM SP Power3 375 MHz 16 way
      NERSC/LBNL

      Once again, AIX.

      4. Intel ASCI Red
      Sandia National Labs

      A poor home-grown OS (no offence) called Cougar or TFlops which doesn't even support X11 or sockets.

      5. IBM ASCI Blue-Pacific SST,IBM SP 604e
      Lawrence Livermore National Laboratory

      Can you say AIX?

      6. Compaq AlphaServer SC ES45/1GHz
      Los Alamos National Laboratory

      I assume Tru64.

      7. Hitachi SR8000/MPP
      University of Tokyo

      No idea. Sorry.

      8. SGI ASCI Blue Mountain
      Los Alamos National Laboratory

      IRIX.

      9. IB SP Power3 375 MHz
      Naval Oceanographic Office

      Don't know for sure, but you can bet it's AIX.

      10. IBM SP Power3 375 MHz 16 way
      Deutscher Wetterdienst

      Again, I'm sure it's AIX.

      All Unix. No, no linux on there yet, but Pacific Northwest will be right up there near the top, and Lawrence Livermore is also probably getting a linux cluster of almost that size pretty soon. That will make two in the top few slots.

      No Windows on these puppies! ;)

    2. Re:Other OSes by qurob · · Score: 1, Funny

      My Mac G4 runs OS 9, y0!

    3. Re:Other OSes by YoDave · · Score: 1

      PNNL will run MSC.Linux.

    4. Re:Other OSes by rutledjw · · Score: 2, Interesting
      I may be ignorant, but I is a college graduate. Doesn't Hitachi compete (to a degree) with IBM in the Big Iron class of machines? Wouldn't that suggest an OS/390-like OS? Just guessing.

      Another thing that I just thought about, maybe someone can answer for me. What about OS/390? I thought that was their big mainframe OS. Is this a speed issue with the OS, clustering limitations (certianly not) or maybe ease of use (people would rather deal with *nix than a 'frame OS)?

      Any input?

      --

      Computer Science is Applied Philosophy
    5. Re:Other OSes by DNS-and-BIND · · Score: 3, Funny
      4. Intel ASCI Red Sandia National Labs

      A poor home-grown OS (no offence) called Cougar or TFlops which doesn't even support X11 or sockets.

      Yeah, everybody knows any computer that can't support netris or even plain old tetris is poor indeed.

      --
      Shutting down free speech with violence isn't fighting fascism. It IS fascism!
    6. Re:Other OSes by JordanH · · Score: 2
      You are correct on the Compaq entries all being Tru64. The AlphaServers will run Linux and Compaq will sell you one, but no buyers yet among these big machines.

      Are any Search engines running Windows yet? I would assume the msn.com search engine runs Windows, but I don't know for sure... If so, I'd believe it's the only one.

      There are some facts that speak so loudly that MS Marketing can't overcome no matter how hard they try.

    7. Re:Other OSes by hawkstone · · Score: 1

      I'll agree with that, but it's poor for other reasons as well. First, having used it, I'm convinced of the futility of synthetic benchmarks, since I don't think it belongs in the #4 spot. Pure processor count is the only way it's up that high. IO performance is also pitiful.

      And if you want to look at the data output from one of their large simulations, you ideally would like to do it in parallel on the machine it was run on (i.e. Red itself). It doesn't support X11 so you can't run a graphical app at all and get a window to display to your desktop. It doesn't support sockets, so you can't even run a parallel visualization app and send data over sockets to a workstation with graphics hardware. The only way to get data off that machine is by sucking the terabytes of data you created off their slow filesystem.

      And that is why I think it is a poor OS.

    8. Re:Other OSes by hawkstone · · Score: 1

      The AlphaServers will run Linux and Compaq will sell you one, but no buyers yet among these big machines.

      I'm pretty sure there are smaller ES45 (or ES50 or whatever) linux clusters around, in the low hundreds of nodes instead of thousands. Also, I thought Sandia's CPlant was running linux on some of these guys and it's sitting at #30. The TOP500 list doesn't say, and I've even used it and still don't remember.... :)

    9. Re:Other OSes by afidel · · Score: 1

      Nope it's yet another flavor of unix HI-UX.

      --
      There are 4 boxes to use in the defense of liberty: soap, ballot, jury, ammo. Use in that order. Starting now.
    10. Re:Other OSes by charmer · · Score: 2, Insightful


      4. Intel ASCI Red
      Sandia National Labs

      A poor home-grown OS (no offence) called Cougar or TFlops which doesn't even support X11 or sockets.


      Why does a parallel machine need X11 or poor (slow) communication primitives? Why should a full OS run on all the processors ? The OS really needs to get out of the way of the computations where every microsecond counts.

      charmer

    11. Re:Other OSes by JordanH · · Score: 1
        • The AlphaServers will run Linux and Compaq will sell you one, but no buyers yet among these big machines.
        I'm pretty sure there are smaller ES45 (or ES50 or whatever) linux clusters around, in the low hundreds of nodes instead of thousands. Also, I thought Sandia's CPlant was running linux on some of these guys and it's sitting at #30. The TOP500 list doesn't say, and I've even used it and still don't remember.... :)

      I believe you are right. There are ES45/40s fielded with Linux, but when I said "big machines" I was referring to the really big machines of your list.

    12. Re:Other OSes by Captain+Large+Face · · Score: 2

      Teoma is using ASP, which I assume is running on top of IIS...

    13. Re:Other OSes by markmoss · · Score: 5, Insightful

      What about OS/390? I thought that was their big mainframe OS.

      Supercomputer != Mainframe

      Supercomputers are just for calculations on massive arrays. Mainframe OS's are designed for government & large corporation databases, etc. They are heavily loaded with "frills" that are unneeded on a pure number-cruncher; they improve database reliability and do many other useful things in the data-processing environment, but they're just wasted cycles on a supercomputer.

    14. Re:Other OSes by jeffy124 · · Score: 1


      2. Compaq AlphaServer SC ES45/1GHz
      Pittsburgh Supercomputing Center

      Haven't used it, but I'm guessing Tru64.
      (emphasis me)

      are you saying you've used the machines on that list? or did you just r00t them?

      --
      The One Rule Of Chess You'll Ever Need: Don't play someone who carries a kit in their bookbag.
    15. Re:Other OSes by PurpleFloyd · · Score: 2

      Why would you suck cycles that could be used for computation for data visualization? (no, it's not a rhetorical question, I have no supercomputer experience and I'm curious). It seems to me that there would be little need for a graphical system on a pure number cruncher; ideally, you would never do much with it outside of sending it data and programs, and pulling out the results. Still, the lack of a socket system seems rather odd. What is the point of a supercomputer without a decent way to get information in and out?

      --

      That's it. I'm no longer part of Team Sanity.
    16. Re:Other OSes by greenrd · · Score: 1
      Why would you suck cycles that could be used for computation for data visualization?

      Obvious answer: You typically run code on a supercomputer only if it would be too slow when run on a slower computer. (Don't know if that's the case here - might be just a convenience thing.)

    17. Re:Other OSes by rapidweather · · Score: 1

      and...is that your father's car? (looks sideways at his chick in the passengers seat) Nooo...!

    18. Re:Other OSes by gregraven · · Score: 1

      My calculations must be off. It looks to me as though they could have bought $24.5 million worth of dual-processor Mac G4s with full RAM and maxxed-out SCSI dual hard drives, they'd have a lot more CPUs and everything else -- and they could still run Unix in the form of Darwin.

      --
      Greg Raven
      As long as there's any left, I'll take mine first.
  2. Wow... good thing they chose linux... by FortKnox · · Score: 5, Funny

    ... Cause if they put WinXP Pro on it, the project would cost:
    $24,500,399.98
    Which was juuust over budget!

    BTW - Can you put in code during the "post slashdot story" to automatically close the <I> tags? I don't think that would be too difficult to add...

    --
    Good quote, too many chars. Seriously, the slashdot 120 char limit sucks!
    1. Re:Wow... good thing they chose linux... by mike_scheck · · Score: 5, Funny

      Not to mention, they would have 1368 wasted CPU's......

    2. Re:Wow... good thing they chose linux... by SuperDaveTX · · Score: 1

      Is that using the per processor or per user licensing scheme? Of course this is all moot since W2K DataCenter can only handle 32 processors. (what a hoss!!) ;-)

    3. Re:Wow... good thing they chose linux... by Jugalator · · Score: 1

      And if they actually weren't wasted, the Product Activation system would probably instead freak out from the masses of hardware it's required to hash a key from... :-P

      --
      Beware: In C++, your friends can see your privates!
    4. Re:Wow... good thing they chose linux... by liquidsin · · Score: 2

      I've heard a lot of arguments against windows, but this is the first time I've heard "it can only handle 32 processors". I don't think I've ever heard *only* 32 processors as a valid argument against an OS...nor did I ever expect to...

      --
      do not read this line twice.
    5. Re:Wow... good thing they chose linux... by n9hmg · · Score: 1

      1368 wasted CPU's

      Are you kidding? With the possible exception of Cray and maybe the Hitachi(can't find info on that one... they seem to be out of the supercomputer business), nobody builds single-unit supercomputers any more. The scalability with clustering and shared memory over high speed networking overcomes the contention problems with massively parallel processors, though the Numa-Q may let us put more cpus in a single box. Check out the AlphaServer SC, the RS/6000 SP, both supercomputers in the top 10 (with the SP dominating... awesome switch for the shared mem), and the Beowulf, or (no shared memory here), the Mosix, clustering. If I were building a linux supercomputer, I think i'd rather go with a pile of dual-cpu units anyway, cutting down on resource contention.

      With distributed processing like these systems, adding another unit adds about another unit worth of processing power, whereas adding cpus to a SMP configuration gives diminishing returns. As long as you build the infrastructure along with the growth (add switch capacity), the sky's the limit.

    6. Re:Wow... good thing they chose linux... by SuperDaveTX · · Score: 1

      Remember that this thread is about a computer that has *1400* (!!) processors. Nothing against windows (no, really) but it isn't even in the ballpark for consideration on this platform.

    7. Re:Wow... good thing they chose linux... by diablovision · · Score: 1

      No no...Microsoft would probably make them buy a separate license for every processor or something ridiculous like that.

      --
      120 characters isn't enough to explain it.
    8. Re:Wow... good thing they chose linux... by ackthpt · · Score: 2, Funny

      Meanwhile, in a lair in Redmond, WA, a figure with glasses and a bad haircut jabs another pin into stuffed tux and mutters, "but... we have the way out..."

      --

      A feeling of having made the same mistake before: Deja Foobar
    9. Re:Wow... good thing they chose linux... by mike_scheck · · Score: 1

      I was simply stating that with windows XP, it can only recognize 32 CPU's. I have never heard of HPC cluster of XP boxes, and because of this all the other CPU's (1368) would be wasted. If you can tell me where a XP cluster is set up that uses multiple machines for parallel processing, let us know.

    10. Re:Wow... good thing they chose linux... by ArsonSmith · · Score: 2

      I think your math there is a little off. Sense windows XP only supports 32 processors, the best you could hope for would be 1400/32 = 43.75 or 44 licenses. at $399.98 each this would be $17,599.12. Of course it wouldn't be practical to set this up with clusters of 32 proc systems and would most likly be in the 2x ($279,986) to 8x ($69996.50) so this can be pretty expensive just in the windows licenses.

      --
      Paying taxes to buy civilization is like paying a hooker to buy love.
    11. Re:Wow... good thing they chose linux... by sharkey · · Score: 2

      Not to mention, they would have 1368 wasted CPU's......

      No, they'd have 1400 wasted CPUs, because, after all, a CPU (and memory, and disk, and power, and...) running XP is wasted.

      --

      --
      "Outlook not so good." That magic 8-ball knows everything! I'll ask about Exchange Server next.
  3. Sigh... by buckeyeguy · · Score: 4, Funny

    all that capability and all I can think about is how much power the dang thing would consume... it'll take one big, big UPS/power conditioner.

    --
    I'd have a personalized plate on my car, but "toxic bachelor" won't fit into 7 letters.
    1. Re:Sigh... by u01000101 · · Score: 1

      all I can think about is how much power the dang thing would consume

      I really hope they turn it off during the nights, and run distributed.net clients on their personal computers at home... ;)

      --
      if you use a good enough junk-filter, slashdot.org will display a single, *blank*, page
    2. Re:Sigh... by teslatug · · Score: 1

      Look at it from the other side...it will be nice and toasty in the winter in the computer room :)

    3. Re:Sigh... by ryanwright · · Score: 2

      I really hope they turn it off during the nights

      Clusters such as this take hours to power up & down. They are rarely turned off.

      --
      -Ryan, with the unoriginal sig
    4. Re:Sigh... by Klatma · · Score: 1

      That is a great quote from The Simpsons.

    5. Re:Sigh... by warpSpeed · · Score: 1

      Not to mention it would be cheaper to run them at night when the demand for power is less.

  4. is it just me... by mraymer · · Score: 1

    or did the author forget to end an italic tag in this story?

    --

    "To confine our attention to terrestrial matters would be to limit the human spirit." -Stephen Hawking

  5. Uhmm... by qurob · · Score: 5, Funny


    Scheduled to be fully operational in early 2003...


    Won't it be obsolete by then?

  6. can i haggle the sticker price?? by dryueh · · Score: 1
    increased scalability, speed and functionality at a lower cost

    A lower cost? Hell...maybe I'll pick one up after work today. With a price tag of only 24.5 million, you're actually making money with this purchase (or, as cases dictate, losing money by not taking advantage of this offer)!

    Sheesh. I think 'reduced' cost is more appropriate.

  7. Cool by jhines0042 · · Score: 2, Funny

    All I can say is:

    "I have GOT to get me one of these!"
    -- Will Smith, "Independence Day"

    (42 Karma, don't mod me)

    --
    42 - So long and thanks for all the fish.
  8. in the immortal words of the old man... by Em+Emalb · · Score: 2

    painting the football field...

    great good googly moogly.

    awesome, lets just hope it functions as it is designed to, could be a huge publicity boost for Linux....

    --
    Sent from your iPad.
    1. Re:in the immortal words of the old man... by FortKnox · · Score: 1

      Good job, but who are the "chefs"?

      --
      Good quote, too many chars. Seriously, the slashdot 120 char limit sucks!
  9. 11 responses... by evacuate_the_bull · · Score: 1, Redundant

    and already 4 comments about a beowolf cluster of these things. cmon slashdot, you can do better than that

    --
    Satanists get good grades too...suspiciously good grades
  10. Ahhhh... I see... by psxndc · · Score: 5, Funny
    So THIS is what we'll need to run PerlBox. :-)

    psxndc

    --

    The emacs religion: to be saved, control excess.

  11. 1,800 intel processors? by qurob · · Score: 1


    Let's see the story when they make one with 1,800 AMD processors!

    1. Re:1,800 intel processors? by Slynkie · · Score: 2

      Nonono, don't you remember? AMD is evil this week.

    2. Re:1,800 intel processors? by pstreck · · Score: 1

      lol, amd is evil, intel is the borg.. what about 1800 POWER3-II procs... mmmm more power. thats it i have ram envy

      --

      Later,
      Phil
    3. Re:1,800 intel processors? by RazorJ_2000 · · Score: 1

      1,800 Intel processors, eh! Damn, I can't think of a MS O/S that will run on it. Maybe it's because MS doesn't have a truly scalable O/S. Yah, yah, cheap shot.



      --
      pi=sigma{n:0-infinity}[(1/16)^n][(4/(8n+1))-(2/(8n +4))-(1/ (8n+5))-(1/(8n+6))]
    4. Re:1,800 intel processors? by archen · · Score: 1

      And the story would be "people for a 100 mile radius blinded by ball of fire created by 1,800 AMD processor computer"

    5. Re:1,800 intel processors? by Indras · · Score: 5, Funny

      Let's see the story when they make one with 1,800 AMD processors!

      Palo Alto, CA: In the news today, 26 researchers, who had been constructing a new super computer for the government running on 1,800 AMD processors, were killed today when they fired up the machine for a test run. Apparently, they had forgotten to turn on the water pumps for the computer's cooling system before starting up the computer. Thousands of megawatts of electricity were instantly turned into heat energy, resulting in a contained explosion that vaporized all the researchers instantly, and turned the building into a pile of melted plastic, metal, and concrete.

      One local, who wishes to remain unknown, said when interviewed, "It was crazy! I mean, the whole building just melted. The heat waves coming out of the building were staggering, it was all I could do just to run into the nearest air-conditioned Starbucks and catch my breath."

      --
      The speed of time is one second per second.
    6. Re:1,800 intel processors? by Tower · · Score: 1

      Nah - 1800 POWER IV procs... now *that* would be some power (pun intended).

      Kinda like this one. And that one is still pretty small (only 24 TB of disk, 4 TFLOPS... but *much* better memory/inter-proc bandwidth...

      --
      "It's tough to be bilingual when you get hit in the head."
    7. Re:1,800 intel processors? by diablovision · · Score: 1

      Well...considering the Itanium (predecessor to McKinley) dissipates 130W of power at 800mhz and the fastest Athlons probably dissipate around 70-80w, I think their choice of the McKinley processors is more likely to create a nuclear furnace than any AMD chip...

      --
      120 characters isn't enough to explain it.
    8. Re:1,800 intel processors? by compwiz3688 · · Score: 1

      Nice.... now if the heated water can be used for fish farming, like those close to the nuclear plants...

    9. Re:1,800 intel processors? by shren · · Score: 2

      And if we built it out of old Cyrix chips, we'd melt a hole to the center of the earth. Cyrix chips run hotter than anything short of a V8 engine. When the fan on one of mine broke, it melted the foam between the cpu and the fan in seconds.

      --
      Maybe the state's highest function is to grind out insoluble problems. (Zelazny, Hall of Mirrors)
    10. Re:1,800 intel processors? by Indras · · Score: 2

      it melted the foam between the cpu and the fan in seconds.

      May I make a suggestion then? Next time, put a metal (aluminum, silver, copper, etc) heatsink in between the fan and CPU, it might hold out better than foam.

      --
      The speed of time is one second per second.
    11. Re:1,800 intel processors? by shren · · Score: 2

      I have no idea why it was the way it was. It came to me that way from the salesman.

      --
      Maybe the state's highest function is to grind out insoluble problems. (Zelazny, Hall of Mirrors)
  12. Re:Those CPUs aren't available yet.... by buckeyeguy · · Score: 1

    Since HP is an insider on that Itanium development, I'd imagine they'll have 'em in hand before the rest of us do.

    --
    I'd have a personalized plate on my car, but "toxic bachelor" won't fit into 7 letters.
  13. Supercomputer Operating Systems by Anonymous Coward · · Score: 1, Funny
    The verdict from major market research firms is in: they unanimously confirm that the following supercomputer operating systems are DYING:

    • AIX is dying.
    • AmigaOS is dying.
    • BSD is dying.
    • BeOS is dying.
    • CPM is dying.
    • DOS is dying.
    • FreeBSD is dying.
    • GNU Hurd is dying.
    • HP-UX is dying.
    • IRIX is dying.
    • Inferno is dying.
    • Linux is dying.
    • LynxOS is dying.
    • MINIX is dying.
    • MacOS is dying.
    • Mach is dying.
    • MicroC/OS is dying.
    • NachOS is dying.
    • NeXT is dying.
    • Nemesis is dying.
    • NetBSD is dying.
    • NetWare is dying.
    • OS-400 is dying.
    • OS-9 is dying.
    • OS/2 is dying.
    • Oberon is dying.
    • OpenBSD is dying.
    • Palm OS is dying.
    • Plan 9 is dying.
    • pSOS is dying.
    • QNX is dying.
    • RTEMS is dying.
    • SCO is dying.
    • Solaris is dying.
    • SunOS is dying.
    • TRON is dying.
    • ThreadX is dying.
    • TinyOS is dying.
    • Unix is dying.
    • VMS is dying.
    • VxWorks is dying.
    • Windows 2000 is dying.
    • Windows 3.11 is dying.
    • Windows 95 is dying.
    • Windows 98 is dying.
    • Windows CE is dying.
    • Windows ME is dying.
    • Windows NT is dying.
    • Windows XP is dying.
    The Free On-Line Dictionary of Computing defines an operating system as: "The low-level software which handles the interface to peripheral hardware, schedules tasks, allocates storage, and presents a default interface to the user when no application program is running. The OS may be split into a kernel which is always present and various system programs which use facilities provided by the kernel to perform higher-level house-keeping tasks, often acting as servers in a client-server relationship. Some would include a graphical user interface and window system as part of the OS, others would not.

    The operating system loader, BIOS, or other firmware required at boot time or when installing the operating system would generally not be considered part of the operating system, though this distinction is unclear in the case of a rommable operating system such as RISC OS. The facilities an operating system provides and its general design philosophy exert an extremely strong influence on programming style and on the technical cultures that grow up around the machines on which it runs.

    1. Re:Supercomputer Operating Systems by krujos · · Score: 1

      Whew DGUX is safe

  14. Sweet by ImaLamer · · Score: 3, Funny


    That answers my question of what I would have done if I won the Powerball last night

  15. Re:And this would be even faster by inkblot69 · · Score: 1

    You mean Microsoft's Xenix?

  16. Wow - that is a big swap space! by gosand · · Score: 5, Funny
    1.8 Terabytes of RAM

    So does that mean it has 3.6 Terabytes of swap space?

    --

    My beliefs do not require that you agree with them.

    1. Re:Wow - that is a big swap space! by petej · · Score: 2

      Some guy at Exabyte need to change his pants now.

    2. Re:Wow - that is a big swap space! by maraist · · Score: 2

      So does that mean it has 3.6 Terabytes of swap space?

      Cute, but swap is optional as far as I know.. When you have enough memory, or when you want real-time (at least in solaris or qnx), then simply leave it out.

      -Michael

      --
      -Michael
    3. Re:Wow - that is a big swap space! by lw54 · · Score: 1
      Swap is still quite handy to have around for troubleshooting when the box core dumps.

      Besides, dont' you know you can't run Windows 2000 without swap? ;-)

    4. Re:Wow - that is a big swap space! by AftanGustur · · Score: 2
      So does that mean it has 3.6 Terabytes of swap space?

      No it does not mean that, you could use just 1MB and still be able to make full use of it.

      --
      echo '[q]sa[ln0=aln80~Psnlbx]16isb572CCB9AE9DB03273snlbxq' |dc
    5. Re:Wow - that is a big swap space! by Gameshow+Bob · · Score: 1

      It's Linux, not HP-UX :)

      --

      You Like Science?
      You Like bottomquark.
    6. Re:Wow - that is a big swap space! by jelle · · Score: 1

      For 2.4, Torvalds recommends SWAP= 2xRAM

      Has to do with swap fragmentation.

      --
      --- Hindsight is 20/20, but walking backwards is not the answer.
    7. Re:Wow - that is a big swap space! by Fjord · · Score: 1

      Only when running emacs.

      ba dum tss

      --
      -no broken link
  17. Insanely expensive by Jeff+Knox · · Score: 4, Interesting

    Wow, do the math thats $17,500 per processor (node). Thats 24.5 million divided by 1400. Whats the deal with that? Even with top of the line components, the fastest interconnects available (Dolphin or Myrinet or whatever), thats a 7 million dollar computer at most (5 grand a machine, with SCI could even build much faster then a 8Teraflop box, hell a dual Athlon or Intel based system would be cheaper and whale on that). Software? Nothing, althought they are probably going to use Scyld or something and pay the bucks. Im willing to bet that half that cost pure adminstrative and contract over head and support.

    --
    Jeff Knox
    1. Re:Insanely expensive by Raleel · · Score: 5, Interesting

      AFAIK, 1/2 of the cost of each node is the interconnect, which has 1-3microsecond latency and gigabit bandwidth. The 24.5 million figure also includes a huge storage array on fibre channel (like 150 terabytes, I believe). And note, each node has 12 gig of ram.

      --
      -- Who is the bigger fool? The fool or the fool who follows him? --
    2. Re:Insanely expensive by Quixote · · Score: 2

      And note, each node has 12 gig of ram.

      Sorry, chief, that should be 1.2GB of RAM.

      What all of your calculations don't include is the standard 200% markup that companies apply when supplying something custom-built to the gummit. Anyone remember the $7000.00 screwdriver ?

    3. Re:Insanely expensive by BurritoJ · · Score: 1

      What all of your calculations don't include is the standard 200% markup that companies apply when supplying something custom-built to the gummit. Anyone remember the $7000.00 screwdriver ?

      There are a lot of factors that lead to high prices for government purchased items.

      First, the product must comply to whatever specifications they give you and you must test and document the compliance. That process alone can add thousands of dollars to the cost of a part.

      Second, custom, one-of-a-kind items always cost many times more than their mass produced counterparts. It doesn't have anything to do with a '200% markup.' It's merely economics of scale. If you have to produce 1 million widgets/year you buy a plant and dedicated, specialized tooling and make them for a penny apiece, sell them for a dollar and make a huge profit. If you have to make three, ever, you hire a machinist and make them for $300 apiece. Of course, with only three parts the engineering development time will only be split 3 ways, resulting in an even higher actual cost.

      It is safe to say that the contractor selling the $7000 screwdriver is making far less profit, no matter how you define it, than manufacturer cranking out millions of consumer grade tools.

    4. Re:Insanely expensive by Stonehand · · Score: 2

      It also had to do with how the contractors accounted for overhead; at least at one time, they divided it up pretty much evenly among all items instead of using a more detailed system.

      Of course, the "journalist" who first picked up and ran with the non-story apparently never bothered to check... or, perhaps, did check but decided that it'd hurt the story.

      --
      Only the dead have seen the end of war.
    5. Re:Insanely expensive by flatrock · · Score: 2

      There are lots of options that can get you 1-3 microseconds of latency and 1 Gbit of bandwidth. Of course you'll likely have some serious software issues. Most programmers like using the good old IP stack to send and receive data. Trying to send and receive 100+ MB/s through an IP stack will pretty much hog the entire CPU. From my expericece with IP over fibrechannel, the throughput is CPU limited because of the IP stack. THere's no use in building a supercomputer if you're going to spend all the processing power just sending information around the interconnect. This means using Lightweight networking protocols, and writing a lot more custom software. Custom written software is very expensive since all the development costs aren't spread out across selling multiple coppies. This is likely one of those cases where a lot of it can be built with off the shelf hardware, but putting it together as a system is more than 90% of the cost. Fibre Channel is actually a good choice for this because you can run multiple protocols over the same network. Actually the company I work at has LINUX drivers that will do IP, a Lightprotocol, and IP simultaneously. We even have them for both X86 and PPC LINUX. We also have better multiple LUN and RAID support than other LINUX drivers I've seen. I wonder if we're involved in this. Maybe I'll go talk to our Fibre Channel Product Manager.

  18. big deal is by Ilgaz · · Score: 1, Interesting

    so, what will that supercomputer will be used for? Arms? Petrol? Investigation? What?

    I wouldn't be happy for such thing happen...

    1. Re:big deal is by ryanwright · · Score: 2

      so, what will that supercomputer will be used for?

      Read the link. It's being installed in the Molecular Science Computing Facility, located within the William R. Wiley Environmental Molecular Sciences Laboratory.

      Quote: "The Molecular Science Computing Facility provides the advanced computing capability needed by staff and users of the William R. Wiley Environmental Molecular Sciences Laboratory (EMSL) to address "Grand Challenge" scale environmental research problems."

      --
      -Ryan, with the unoriginal sig
    2. Re:big deal is by Captain+Large+Face · · Score: 2

      Well, it's being installed in the Molecular Sciences Laboratory, so...

  19. Surprised it's intel based... by pstreck · · Score: 1

    I'm kinda surprised to see that this computer is actually running on intel proccessors. Typically when this kind of horse power is required while a risc processor gets chosen. I'm not real familiar with the itanium other than i know it's a 64bit proc, and the workstations are REAL expensive :) The question is how does the itanium compare to it's oppenents like the POWER3-II copper based chip?

    --

    Later,
    Phil
    1. Re:Surprised it's intel based... by 5.25"+Floppy · · Score: 1

      Actually, (although your mileage may vary!) the Intel processors are often faster. Having run some CFD applications on a POWER3-II (375Mhz) and on an Athlon MP 1600+, the Athlon outperformed the POWER3-II by a factor of 2.3!.

      The advantage of the POWER3-II based systems lies in the fact that you can have large SMP systems, large memory access, etc. (It's a 64-bit chip)

      Now, the latest RISC from IBM is actually the POWER4... and it boasts some damn impressive SPEC marks. I haven't run on one yet, but from what I hear, I can expect it to be slightly faster than my Athlon 1600+... but with the Athlon MP 2000+ out, and the price difference between that and the POWER4 - well, I like my Athlons.

      Ciao!

      (PS. Oh, and unfortunately, I have no data on running on an Itanium. But if the commodity chips work so damn well, why bother?)

    2. Re:Surprised it's intel based... by tanveer1979 · · Score: 1

      you are actually getting confused here. The processor is not the standard pentium. its the Madison and McKinlay.
      Actually the Madison is a successor to the McKinlay, and these are 64 bit monsters, and they are good :-)

      --
      My Aurora : http://www.youtube.com/watch?v=o91ZsGwJYyg
      FB : https://www.facebook.com/TanveersPhotography
    3. Re:Surprised it's intel based... by Tower · · Score: 1

      For single processor applications, the Athlon can keep pace in many tasks with the POWER-IV, but the Athlon can't match the internal bandwidth of the POWER IV nodes (4/8 proc cores on a module, higher L2 and memory bandwidth, etc). That does, of course, show up in that price tag :-) The other advantage is (as you mentioned) the larger SMP ability of the POWER systems as compared with the Atlhons, but hey - you really need some work to keep 24 nodes of 32 procs each busy.... Fun!.

      --
      "It's tough to be bilingual when you get hit in the head."
    4. Re:Surprised it's intel based... by Derkec · · Score: 2

      This is something I've found interesting too. A couple of the recent super computers have been Itanic. I think Intel has done something that allows them to scale pretty well into the several hundred or thousand processor range. Meanwhile, most of the RISC server processors are focused on configurations of 2-128 processors. I'm speculating here, but that's my guess.

    5. Re:Surprised it's intel based... by YeeHarr · · Score: 1

      I expect McKinley to be the best price/performance FLOATING POINT chip out there when it is shipped.

      In the scientific market integer performance isn't the problem it is floating point.

      IPF (Itanium Product Family) is specifically targetted to do very good floating point. The first generation Itanium in fact had terrible integer performance but very good floating point performance.

      There will be no comparison between a PIV and McKinley when doing things like floating point matrix multiplies etc etc. This is the important measurement when doing scientific computing.

      AMD will not be able to touch the floating point performance of McKinley - even hammer. They are aimed at different markets.

      So ... there are two reasons for McKinley - 64bit and excellent floating point performance.

  20. *cough* bullshit *cough* by evilpaul13 · · Score: 1, Troll

    "Open architecture"? How many other "open architectures" are only and will only be manufactured by one company (Intel)? Itanium is as closed as they come. I've gottta go out now, and buy that "Open Windows" Microsoft just started selling.

    McKinley isn't out yet, and Madison is its successor. Why don't they just build it using Pentium 8s and Athlon XP 52s?

    Sure, mod me a troll for not eating up the nonsensical marketing garbage that is being passed off as a news story. Did HP start advertising on /.?

    1. Re:*cough* bullshit *cough* by rutledjw · · Score: 1
      Not troll, but flamebait. You should feel lucky there isn't a "moron" mod!

      BTW, it could have been Athlon, SPARC, Alpha or PowerPC chips. So much for are only and will only be manufactured by one company.

      --

      Computer Science is Applied Philosophy
    2. Re:*cough* bullshit *cough* by Artana+Niveus+Corvum · · Score: 1

      ...key one word here...

      -->CLUSTER

      the bus differences mean less than nothing
      ------
      sorry, moderators, just couldn't let the...
      ehm... ignorance, yeah, that sounds good...
      slide.

      --
      -----------------------------------------
      Remove the Greed which plagues mankind.
  21. Just in time... by HiQ · · Score: 1

    I think that this will be one of the few machines in the world to run the Perl desktop on a reasonable speed ;)

  22. where's the hurd? by squidinkcalligraphy · · Score: 1

    Geez, ud think they might just invest a little R&D into the hurd and have an even more scalable OS.

    --
    "I think it would be a good idea" Gandhi, on Western Civilisation
    1. Re:where's the hurd? by sketerpot · · Score: 1
      Let's consider some of the advantages and disadvantages of Linux vs. The HURD on this.

      HURD

      • Built for great scalability, with things normally built into the kernel running in userland as daemons.
      • Has a doubly recursive acronym.

      Linux

      • More developed, with lots of good features, like POSIX thread support.
      • Has a very great rate of development, with lots of people contributing.
      • Is easy to get, and has the LSE project working to make it more scalable.

      I have nothing against the hurd, I just think that the development money might be better spent making Linux more scalable.

  23. Figures for the layman by popeyethesailor · · Score: 4, Funny

    1) 8.4 TFLOPS lets you find the sum of 4.2+4.2, 168 trillion times a second.
    2) 170 TB can hold 42.5 thousand times the contents of the entire Library of Congress books .(+ all the MP3s you downloaded )
    3) 1 TB of RAM may let you run as many as 13 Windows applications simultaneously.

    1. Re:Figures for the layman by SuiteSisterMary · · Score: 5, Funny

      In other OTHER terms, on a machine like this, the lastest build of Mozilla is actually smooth and responsive.

      --
      Vintage computer games and RPG books available. Email me if you're interested.
    2. Re:Figures for the layman by Hoo00 · · Score: 1

      Unless you run Mozilla with Sun Java plugin.

    3. Re:Figures for the layman by compwiz3688 · · Score: 1

      1) 8.4 TFLOPS lets you find the sum of 4.2+4.2, 168 trillion times a second.
      And from the kernel compiler's PoV, it only takes a second to compile the latest 2.4? (I haven't compiled one in a long time now, so I don't know)

      2) 170 TB can hold 42.5 thousand times the contents of the entire Library of Congress books .(+ all the MP3s you downloaded )
      The whole Internet!! Download the whole Intern... (I'm pretty sure this is getting old :)

      3) 1 TB of RAM may let you run as many as 13 Windows applications simultaneously.
      And if I set that up as a RAM drive, I get to try several games in RAM without ever touching the HD! (ok, maybe the registry) If I like it, I'll really install it on the HD.

    4. Re:Figures for the layman by damien_kane · · Score: 1

      2) 170 TB can hold 42.5 thousand times the contents of the entire Library of Congress books .(+ all the MP3s you downloaded )

      But how many copies of the Human Genome is this?

    5. Re:Figures for the layman by sean23007 · · Score: 2

      No, not yet, we'll have to wait for the next version...

      --

      Lack of eloquence does not denote lack of intelligence, though they often coincide.
  24. Effect on linux ? by nilsj · · Score: 3, Interesting

    How will this affect linux ?
    Will HP come up with something revolutionary in linux development while constructing this system or is the tech used conventional - just on a bigger scale ?

    1. Re:Effect on linux ? by djbentle · · Score: 2, Interesting

      The leader of the ASCI Blue program spoke at a graduate seminar I attended. He mentioned that when they first got AIX up and running on the over 6000 processors they found once a year bugs (a bug that in a normal implementation would only appear roughly once a year) at a rate of one every few days for quite a while. At the very least it ought to help flush out a lot of very rare bugs.

      David

  25. Re:And this would be even faster by RazorJ_2000 · · Score: 2, Funny

    Yah, because they need the horsepower to run solitaire...


    --
    pi=sigma{n:0-infinity}[(1/16)^n][(4/(8n+1))-(2/(8n +4))-(1/ (8n+5))-(1/(8n+6))]
  26. Banking heavily on McKinley not tanking. by Christopher+Thomas · · Score: 5, Interesting

    They're awfully confident of McKinley not following in the footsteps of Merced if they've placed this order.

    This raises an interesting question, though. If you want to build a high-performance compute cluster nowadays... what do you build it out of? The old answer, Alpha, doesn't really apply any more.

    Sun is optimized for communications bandwidth, not FLOPS, and I'm not sure if SGI even _offers_ machines that huge. HP is betting on IA64. And x86 is competely unsuitable, for memory space reasons if nothing else.

    What am I missing?

    1. Re:Banking heavily on McKinley not tanking. by SuiteSisterMary · · Score: 2

      NUMA. Go look it up. :-)

      --
      Vintage computer games and RPG books available. Email me if you're interested.
    2. Re:Banking heavily on McKinley not tanking. by afidel · · Score: 2, Informative

      IBM RS/6000's with Power4 cores? Since the Power3 is the dominant chip in the current top10, hell top20 I would guess it's bigger and better brother will be there soon. Also what about Sledgehammer could we see another mega cluster like ASCI Red(which is still #4 despite being based on PPro 150's)?

      --
      There are 4 boxes to use in the defense of liberty: soap, ballot, jury, ammo. Use in that order. Starting now.
    3. Re:Banking heavily on McKinley not tanking. by Christopher+Thomas · · Score: 2

      NUMA. Go look it up. :-)

      NUMA doesn't touch the address space problem, or the processor-type problem. It's just a way of arranging the memory hierarchy.

    4. Re:Banking heavily on McKinley not tanking. by Christopher+Thomas · · Score: 1

      IBM RS/6000's with Power4 cores?

      Whoops :). Knew I'd forgotten something.

    5. Re:Banking heavily on McKinley not tanking. by SpinyNorman · · Score: 2

      McKinley was designed by HP rather than Intel as was Merced (aka Itanic), so it may actually perform well, considering HP's PA-RISC work.

    6. Re:Banking heavily on McKinley not tanking. by macinslak · · Score: 2

      I believe that IBM's Power4 is still faster than anything else for serial execution by quite a large margin.

      Also, WRT address space, I would think memory access on these things is quite heavily abstracted for any userland tasks. When you reach outside of any one machine on the cluster, conventional memory access methods probably go out the window anyway. The ASCI Red was just a bunch of P6's soldered together after all, and it doesn't seem to be having too many problems.

    7. Re:Banking heavily on McKinley not tanking. by Christopher+Thomas · · Score: 2

      Also, WRT address space, I would think memory access on these things is quite heavily abstracted for any userland tasks. When you reach outside of any one machine on the cluster, conventional memory access methods probably go out the window anyway. The ASCI Red was just a bunch of P6's soldered together after all, and it doesn't seem to be having too many problems.

      I seriously doubt that any x86 cluster uses a unified address space from any given task's point of view.

      Abstracting memory accesses could let you make your real address space a window into a larger one, but that would have some pretty nasty overhead.

    8. Re:Banking heavily on McKinley not tanking. by Christopher+Thomas · · Score: 1

      SGI markets 1024 machines "out of the box" and there is a 6000+ CPU machine on the top 10. They are currently working on the 2nd generation which will allow 33,000 CPUs (15 bits of CPU's). It'll be cool.

      My mistake re. SGI; thanks for making me aware of this (another poster pointed it out too).

    9. Re:Banking heavily on McKinley not tanking. by ender81b · · Score: 2

      If you wnat to learn more about 64-bit processors I suggest you read this article:

      Extreme Tech article

    10. Re:Banking heavily on McKinley not tanking. by cameldrv · · Score: 1

      ASCI Red I believe originally had Pentium Pro 200s in it. All of the processors were replaced in it a few years back with Pentium II 300 Overdrive chips, which is a drop-in replacement for the PPro. Incidentally, the PII Overdrive has a totally different cache structure than the regular PII, I believe it has a 256k full speed L2 cache.

  27. compilers? by afidel · · Score: 1

    Are there good optimizing compilers for Fortran95 and the other major research languages for the IA64 yet?

    --
    There are 4 boxes to use in the defense of liberty: soap, ballot, jury, ammo. Use in that order. Starting now.
  28. At a lower cost??? by ramdac · · Score: 1

    "Today's announcement shows how HP has worked to help accelerate the shift from proprietary platforms to open architectures, which provide increased scalability, speed and functionality at a lower cost,"

    24.5 Million Dollars?... Lower??

  29. Re:It's official: Slashdot is dying by Arsewiper · · Score: 1

    For someone with such stong opinions it'a pity you weren't brave enough to own them. Poor effort Microsoft. Slashdot will probably outlast you once your shareholders wake up.

  30. Supercomputer(s) by totallygeek · · Score: 3, Insightful
    The problem I have with calling these huge clusters supercomputers is that they really don't seem to fit the mold of the term. I prefer to call them supercomputing networks. When I think of a supercomputer, I am thinking of one entity that is hugely multi-processor or multi-boxed in an enclosure. These systems usually have matrixed processing technology and perform a specialized task for the hardware wrapped around them.

    I am impressed, however, with any of these clusters, and am amazed at the cost savings. But, you have other concerns with a huge cluster: redundancy, heat, energy usage, space requirements, etc.

    1. Re:Supercomputer(s) by afidel · · Score: 2, Informative

      Since almost none of the top500 supercomputers are true old style supercomputer's I doubt many people would agree anymore. Vector supercomputers are expensive (hmm we'll have at least as much r&d as intel but spread it across a couple 10's of thousands of cpu instead of millions), They are finite in their expansion, and they are only extremely usefull for a handfull of problems (although they are the ones that most bother "normal" computers). Clusters of comodity systems using high performance interconnects has been the way HPC has been moving since the mid 90's, There are currently no traditional supercomputers in the top10, and there are only 4 in the top25. Redundancy is taken care of using scheduling algorithms that can handle a few lost nodes, heat, hehe well since the old Crays used liquid nitrogen cooling I doubt a cluster is any worse than any other solution, and yes space is a consideration, but most institutions that have HPC's will either build them their own building or will be removing an old supercomputer that does considerably less and takes up more room (remember computers keep getting more dense in terms of flops/sq ft).

      --
      There are 4 boxes to use in the defense of liberty: soap, ballot, jury, ammo. Use in that order. Starting now.
    2. Re:Supercomputer(s) by Boone^ · · Score: 2

      The only problem with these "high performance interconnects" is that they're too generic to get clustered boxes working on any serious HPC codes. The dedicated high bandwidth/low latency interconnect in a Cray T3E (1994) still smokes any Myrinet (high bandwidth, high latency) serial interconnect linking commodity processors today.

      When you need your "supercomputer" to have more synergy than simply sharing the same power supply, people still go with traditional big iron like cray, sgi, ibm, and nec. These codes require tight coupling and sharing of data between nodes and processors and can't afford to spin NOP cycles as the latency over Myrinet kills their performance.

      Yes, clusters do run some codes extremely well. These are the ones that don't require much node-to-node communication and only use the interconnect to setup the executables locally. RC5 and SETI would be a great example of these. But what if a RC5 key depended on the answer to the results of a key from its nearest 8 neighbors? Suddenly the clustered interconnects are swamped and REAL performance becomes a low percentage of peak.

    3. Re:Supercomputer(s) by tedgyz · · Score: 1

      That was my first thought. We have to redefine the term supercomputer. It used to be that these massively parallel systems were connected through high-bandwidth memory.

      According to the detailed article, it is made up of 700 nodes connected to a SAN. Hence, the term supercomputer seems to cover a network of computers.

      Given that broader definition, one could argue that metaprocessors like United Devices, SETI@Home, distributed.net are supercomputers. Although their connectivity is poor, they have a lot more than 1400 processors.

      I had the fortunate experience of attending a supercomputing conference in 1995 while working for HP. This was back when Thinking Machines, etc. were in their heyday. At the end of the conference, we talked about future supercomputers. Everyone conceded that the internet was the supercomputer of the future, where people would lease out compute power.

      I think what we are seeing now is the nascent stages of this technology. Idealistic research projects may give way to capitalistic projects. Imagine getting paid to share your computer. Take the $24.5M and pay people $50 a system. You could get 490,000 systems at your disposal. :-)

      --
      "No matter where you go, there you are." -- Buckaroo Banzai
    4. Re:Supercomputer(s) by nr · · Score: 1

      Yes, like weather forecasting codes requires thightly coupled processors with highbandwidth. Its sad to see the decline for vector supers, the only true vector ones left that I know of is Crays SV1, Hitachi SR8000 and Fujitsu VPP5000. The Sony Playstation 2 also have a vector processor inside. It requires special assembler instructions to take advance of vector calculations so not many games push the envelope and use the computing power to 100%.

  31. I wonder if there'll be a firesale by the_consumer · · Score: 1

    for the ibm this thing is replacing. it'll only be 6 years old in 2003...

    --
    "If you're thinking what I'm thinking, you're right." -
  32. Support is one reason. by TheViffer · · Score: 2

    Its for the package, not just the hardware. It could even include tax and shipping.

    --
    -- Knowing too much can get you killed, but knowing who knows too much can make you rich.
    1. Re:Support is one reason. by Skater · · Score: 1

      Other ideas:

      Employees to install it...
      Renovations to the room where it'll be that are required...
      Lots of power cables...
      --RJ

    2. Re:Support is one reason. by rgmoore · · Score: 1
      Its for the package, not just the hardware. It could even include tax and shipping.

      It probably also includes the cost of the engineers to install it, the building it's going to be put in, the power conditioning and backup, and a service contract.

      --

      There's no point in questioning authority if you aren't going to listen to the answers.

  33. Don't listen to this! by BitwizeGHC · · Score: 2

    ACK!!!

    On my system, at least, this would make the system unusable!!! 'exec' is a shell builtin that calls execve() to replace the shell process with another process. 'true' just returns a true value to a shell script, and does nothing really.

    Be careful of this troll.

    --
    N4st0r, trixx0r h0bb1tz0rz! Th3y st0l3 0ur pr3c10uzz!
  34. Oh wow... by Ixe · · Score: 1

    Just think how fast I could crash!

    --
    Sigs pose an operational security risk and help the baddies aggregate data. I guess commenting does too, oops.
  35. Bets are up on the uptime! by mastropiero · · Score: 1

    With all that exotic hardware, how do you think it will react?

    The guys on kernel dev are going to have to work a good deal to support that. Anyway, I wonder what its average uptime will be.

  36. In Other News... by Komarosu · · Score: 5, Funny

    "8.3 Trillion Floating Point Operations per Second, 1.8 Terabytes of RAM, 170 Terabytes of disk, (including a 53 TB SAN), and 1400 Intel McKinley and Madison Processors."


    Microsoft finally release the baseline specifications for there next generation operating system...

    --

    "What do you mean you have no ice? Do you expect me to drink this coffee hot?" - Random Customer, Clerks
    1. Re:In Other News... by zbuffered · · Score: 3, Funny

      Of course they claim that it will run with 1.8 TB of RAM, but everybody will tell you that you'll barely be able to boot with anything less than 3 TB.

      --
      Synergy is your friend
    2. Re:In Other News... by grytpype · · Score: 2

      > there next generation

      THEIR next generation.

      Don't they teach their/there/they're in school anymore? Must be dumbing down the curriculum.

      --

      - Have a picture

    3. Re:In Other News... by damien_kane · · Score: 1

      Don't forget... The 170TB install does not include IE, since they've finally decided not to bundle them.
      To get IE running on the system you need an extra 400TB storage and 17TB RAM

      Oh yeah... and it's gonna cost you your first and second child if you want support when you hit back

    4. Re:In Other News... by Komarosu · · Score: 1

      yah, so i cant spell...i work in IT u know :)

      --

      "What do you mean you have no ice? Do you expect me to drink this coffee hot?" - Random Customer, Clerks
  37. IBM? by mgw1181 · · Score: 1

    Didn't IBM build a few large clusters in the last year or two?

    1. Re:IBM? by larien · · Score: 2

      If you look at the current list on the TOP500 web site, a large portion of the systems are SP2 clusters, so I'd say yes.

  38. Here's another supercomputer running Linux by qurob · · Score: 3, Informative


    What about this one?

    3:00 a.m. March 22, 2000 PST
    The University of New Mexico and IBM are teaming up to build the world's fastest Linux-based supercomputer.

    Named "LosLobos", the new supercomputer is scheduled to be fully operational by the summer


    Whats the current status?

    1. Re:Here's another supercomputer running Linux by BigJimSlade · · Score: 4, Funny


      What about this one?

      Named "LosLobos", the new supercomputer is scheduled to be fully operational by the summer


      After getting stuck in an infinite loop playing "La Bamba", IBM Engineers smashed it into little bits.

    2. Re:Here's another supercomputer running Linux by FredGray · · Score: 2

      Here's yet another one: the "Platinum" IA-32 Linux cluster at NCSA in Urbana, Illinois. I'm lucky enough to have an account on this monster.

  39. Begs the question... by Rogerborg · · Score: 2

    What's "a computer" (singular)? The "details" links are a little short. 1,400 processors, wow. How many kernels? 1? 1,400? What's the topography? Will it use resources completely dynamically, or can you split it into fixed side sub-units? If you can hot swap parts, can you turn off e.g. half of it and still feed the other half problems? Are various parts of it drawing from independend power sources? Is there a single point of control, or are there multiple master processes?

    What I'm getting at is: at what point does a multiple processor "supercomputer" start to be indistinguishable from a "distributed computing network". Imagine a Beowulf cluster of SETI@home networks, for example. ;-)

    --
    If you were blocking sigs, you wouldn't have to read this.
    1. Re:Begs the question... by joib · · Score: 2

      700 kernels (it says 700 dual processor systems in the article). Communicating with MPI, I'd guess, which is the standard these days.

  40. Is it big enough to run Doom 3? by Mercodus · · Score: 1

    I think not

    --
    Do not meddle in the affairs of dragons, because you are crunchy and taste good with ketchup.

    --
    All alcoholics quit. Some while they are still alive.
  41. $24 million Linux cluster by pohzer · · Score: 1

    Yeah, it runs Linux, but I bet they didn't make any profit on the sale.

  42. but... by Prong_Thunder · · Score: 1

    will it run nethack?

  43. SGI by mapnjd · · Score: 2, Informative

    SGI certainly do sell machines with more processors than this: SGI ASCI Blue Mountain has 6144 CPUs

    Re: your less-than-insightful comment on x86: Intel's ASCI Red has 9472 x86 CPUs. Guess what - they don't share 4GB memory...

    Like the other poster said: look up NUMA.

    nic

    --
    Bus error in your favour. Collect 200kB
    1. Re:SGI by Christopher+Thomas · · Score: 2

      Re: your less-than-insightful comment on x86: Intel's ASCI Red has 9472 x86 CPUs. Guess what - they don't share 4GB memory...

      *sigh*.

      If you're dealing with problems where you don't need to have the entire data set visible to all processors, great; use x86.

      If you need to map the entire address space, you need more than the 36 or so bits that x86 offers you.

    2. Re:SGI by afidel · · Score: 1

      1GB of ram is commonplace on a webserver, now this thing has almost 2 TERA-bytes, order of magnitude bigger, and not adressable by x86-32 even with the 36bit extensions that Intel hacked onto the architecture starting with the ppro.

      --
      There are 4 boxes to use in the defense of liberty: soap, ballot, jury, ammo. Use in that order. Starting now.
    3. Re:SGI by oliverthered · · Score: 1

      Maybe that's why there making 64bit processors, or it could just be to increase the size of the machine word and slow things down a little.

      --
      thank God the internet isn't a human right.
  44. Re:Someone's going to say it sooner or later... by Lizard_King · · Score: 2

    If this doesn't show the power of linux scalability, nothing EVER will
    I would expect these over-generalized, broad assumptions from Mr. Katz... not from Slashdot readers. There are folks out there currently researching many dimensions of the scalability of Linux that delve deaper than the challenges creating of a supercomputer.

    it's nice to see these companies working together to further common platforms.
    Don't be so quick to buy into the Mr. DeMillo's corporate rhetoric. Pacific Northwest made an educated business decision (*that's* why this is a good thing). For Linux to be truly embraced in the business world, organizations must realize the business value that this OS can provide for their company. Companies do not undertake large expenditures such as this to 'further common platforms'.

    --
    "My mother never saw the irony in calling me a son-of-a-bitch." - Jack Nicholson
  45. Not all that surprising... by pgpckt · · Score: 2


    Most supercomputers have been using Unix (and the many varients thereof) for a long time. Unix has always seemed to be able to handle multiple processors efficently. This is just the rich man's version of a beowulf cluster :)

    --
    Lawrence Lessig is my personal hero.
  46. TCO of Linux by prophecyvi · · Score: 1

    This could be well-spun by Linux companies... this machine is rather obviously not using Linux because it's free-as-in-beer, which is still the common perception of the best reason to use it. Microsoft says "when you pay for software, the software is held accountable", well, $24.5 mil is some pretty deep accounting.

  47. Re:Myrinet has done gigabit for ages by afidel · · Score: 1

    No storage is cheap, a netapp with a half terabyte of fibrechannel disk storage is like $60k, so even for what is very much an expensive way to get storage you still only get like $6 million for 53TB of fibrechannel disk space, if you went jbod I can't see it costing much over $3million even with the fibrechannel switches etc.

    --
    There are 4 boxes to use in the defense of liberty: soap, ballot, jury, ammo. Use in that order. Starting now.
  48. 'Open' by Ed+Avis · · Score: 4, Insightful

    'Open architectures'? But it's going to be running Intel's proprietary IA-64 family, where the USPTO has even granted patents on certain CPU instructions. H-P's claim would ring more true if they'd gone with IA-32 (which has two competing suppliers, at least) or SPARC (which you can license from some half-baked consortium).

    Unfortunately there is no fully open hardware platform at the moment, and closed hardware is less of a problem than closed software, but still this sounds like marketspeak.

    --
    -- Ed Avis ed@membled.com
  49. Design ? by AftanGustur · · Score: 2

    Cause if they put WinXP Pro on it, the project would cost: $24,500,399.98

    If one of the design goals is raw performance (and it likely is), the number might be a lot higher with XP on it.

    --
    echo '[q]sa[ln0=aln80~Psnlbx]16isb572CCB9AE9DB03273snlbxq' |dc
  50. Finally something to play Doom on! by screenbert · · Score: 1

    Do you think it has a joystick port? for $24 million it better have a freaking joystick port and a sound card. And if the freaking thing doesn't have edlin loaded on it just forget it.

    ... If you hold your HD to your ear can you hear the c:?

  51. expensive by sdflkgfljdqshgjkqsfg · · Score: 1

    170 Terabytes of disk thank god it's not being shipped to Canada. Imagine the price tag on that thing!

    --
    how does one change his /. id?
  52. How is an IA64 machine "open architecture"? by sawdey · · Score: 1
    I assume they must be crowing about the software architecture because IA64 is about as closed as you can get. Intel/HP have patents that will pretty much prevent anyone else from cloning it (assuming anyone would want to).

    But if they are talking about the software on the machine, why the emphasis on HP? Did HP fund the IA64 linux port? Perhaps they are referring to HP's willingness to sell a big pile of boxes with Linux as the operating system.

    1. Re:How is an IA64 machine "open architecture"? by zwarte077 · · Score: 1
      HP and intel co-developed the McKinley (and I assume Madison) processor. HP developed a chipset (zx1) that features a high memory-bandwith low memory-latency. The 2-way board has 3 PCIX (512 MByte/s) ports. The 4-way has 4 PCIX ports. http://www.hp.com/products1/itanium/chipset/index. html

      Since HP helped Dreamworks move to Linux (Shrek was rendered on a linux farm), it appears they are realizing the market for Linux services. In January, they announced they were open sourcing their printer drivers. Why emphasis HP? They are providing the services, the hardware, and chipset. Intel is only providing the CPU's (and only half if HP co-developed them).

  53. Yeah... much lower cost... by __aahlyu4518 · · Score: 1

    "Today's announcement shows how HP has worked to help accelerate the shift from proprietary platforms to open architectures, which provide increased scalability, speed and functionality at a lower cost,"

    Yeah... only about $2.5 million per buzzword... Proprietary platforms DO cost only $2 million per buzzword, but they have a lot more of 'em :-)

  54. McKinley and Power Consumption by Space+Coyote · · Score: 1

    OK, does it strike anyone else as being very odd that a processor like McKinley would be used in this manner over a RISC processor? Smells like Intel's close relationship with HP had a hand in the decision making process for this one. What would be the benefit of all that extra silicon crammed into the IA64 chips over a much smaller RISC chip in this kind of computing environment? One would not be remiss in thinking that a similarly equipped RISC-based system from IBM or SGI would probably cost a lot less both up front and in terms of power consumption, etc. than this monster of a space heater they're planning on building. If there is any reason other than marketing that they are going with McKinley on this one I'd genuinely like to know.

    --
    ___
    Cogito cogito, ergo cogito sum.
    1. Re:McKinley and Power Consumption by JayAndSilentBob · · Score: 1

      They plan to build a turbine above the computer and sell the power to *other* supercomputers as "better" power.

      --


      Love,
      Jay and Silent Bob
    2. Re:McKinley and Power Consumption by TrollBurger · · Score: 1

      Your tubby friend loves the cock

  55. I'll be waiting by PhunkyOne · · Score: 1

    I know all that sounds really impressive but I will have to wait and see what the Quake III benchmarks come out at or how fast it encode DiVX :)

  56. Top500 slashdotted... by pibakic · · Score: 2, Funny

    "This will be one of the top ten fastest computers in the world."

    Anyone else find it amusing that the link to the top 10 fastest computers in the world appears to be slashdotted?

    Pib.

    --
    "NEVER, EVER feed your computer beer" - some /.er
  57. I want A beowolf cluster of these by rveno1 · · Score: 2, Insightful

    Imagine a Beowolf cluster of these babies

  58. GOOGLE! by Jagasian · · Score: 5, Informative

    What about Google?!? It should qualify as a Linux supercomputer. For those who don't know, Google, the popular search engine, uses a huge cluster of PCs running Linux.

    1. Re:GOOGLE! by Julian+Plamann · · Score: 5, Interesting

      Yep, Google runs on a cluster of approximately 4,000 1U servers. Each can be pulled and replaced including automatic configuration/loading of the operating system and software configuration within about 20 minutes I believe. Pretty neat setup.

    2. Re:GOOGLE! by UranusReallyHertz · · Score: 1

      Actually, the Google Compute FAQ mentions that Google runs on OVER 10,000 computers. Damn, but thats gotta be lotsa fun to administer. With that many computers, one must fail every couple days I bet, unless they have some quality hardware.

      --
      Smoking is an expensive, slow, and unreliable method of suicide.
    3. Re:GOOGLE! by mikeee · · Score: 2

      From what I hear, they have cheap-junk hardware.

      However, the software is built to handle node failures, so they probably just swap out a couple a day. :)

      Not sure Google qualifies as a supercomputer in that it isn't general-purpose... I don't think you can run a simulation on google.

  59. oohh... yes, yes by mastropiero · · Score: 1

    Kernel compiled in 1 ms, UT starting in 0.5, a vast RAM space to put whatever crap i can think of.

    I got to r00t me one o' these!

    I think I just had an orgasm...

  60. Linux IS Unix by leereyno · · Score: 4, Informative

    AIX is Unix
    BSDI is Unix
    HP-UX is Unix
    Solaris is Unix
    Sun-OS is unix
    Digital Unix...is Unix
    FreeBSD is Unix
    NetBSD is Unix
    OpenBSD is Unix
    A/UX is unix
    Xenix is unix
    Unixware is unix
    SCO Unix is Unix
    NextStep is unix
    Unicos is unix
    Irix is unix
    Ultrix is unix

    and yes, Linux is Unix.

    It may not be Unix(tm), but it certainly is unix, at least as much as any of the above operating systems are. Whether or not an OS has one line of code from Thompson and Ritchie or BSD is irrelevant. What matters is what kind of a system its code implements. The code for Linux, including all of the GNU components and other userland parts, implement an operating system that is at least as similar to any of the above mentioned OS's as they are to one another. I don't know just exactly how compliant Linux is with the various posix standards, but I have heard it referred to as posix compliant, and I know that NO version of unix is completely compliant.

    If it looks like a duck, walks like a duck, and quacks like a duck....its a duck.

    Lee

    --
    Muslim community leaders warn of backlash from tomorrow morning's terrorist attack.
    1. Re:Linux IS Unix by anti-snot · · Score: 1

      Don't forget, with Interix, Win2000 is Unix.

    2. Re:Linux IS Unix by warpSpeed · · Score: 1
      Xenix is unix



      isn't that a bit of a stretch?



      why not add 3b2s while we are ATT it (pun intended)

    3. Re:Linux IS Unix by sean23007 · · Score: 2

      AIX is Unix
      BSDI is Unix
      HP-UX is Unix
      Solaris is Unix
      Sun-OS is unix
      Digital Unix...is Unix
      FreeBSD is Unix
      NetBSD is Unix
      OpenBSD is Unix
      A/UX is unix
      Xenix is unix
      Unixware is unix
      SCO Unix is Unix
      NextStep is unix
      Unicos is unix
      Irix is unix
      Ultrix is unix
      Linux is Unix.

      But just remember: GNU is Not Unix

      --

      Lack of eloquence does not denote lack of intelligence, though they often coincide.
    4. Re:Linux IS Unix by egreB · · Score: 1

      You forgot OS X. It's Unix. And put in LUnix (http://lng.sourceforge.net/) when you're at it.. It's kind of Unix as well. There is a lot of things that are Unix, when I think about it.. My ruler (you know, the long plastic-thing with lines, dots and numbers printed in symmetric intervals often used at school (you know, the place where you're supposed to learn stuff (you know, .. forget it))) is made by "UNIX" for some reason. More?

    5. Re:Linux IS Unix by edrugtrader · · Score: 2

      but remember: XINU Is Not Unix.

      --
      MARIJUANA, SHROOMS, X: ONLINE?! - E
    6. Re:Linux IS Unix by mchappee · · Score: 1

      >>and yes, Linux is Unix.
      >>It may not be Unix(tm), but it certainly is unix

      Actually, since Caldera bought SCO, didn't they purchase the rights to Unix(tm)? If so, wouldn't that make Linux the only "true Unix"?

      Matthew

      --
      /. finds me to be 20% Troll, 80% Funny
    7. Re:Linux IS Unix by gregraven · · Score: 1

      Darwin is Unix, too, isn't it?

      --
      Greg Raven
      As long as there's any left, I'll take mine first.
  61. If you find that interesting... by jbischof · · Score: 2, Interesting

    The worlds largest supercomputer is being built as we speak at various campuses around the world. Its a multipart system with various clusters linked together at the different campuses. If your interested I covered the basics of the system below.

    TeraGrid is the name of the soon to be world's largest computing cluster that will be completed in 2002. It will contain approximately 3,300 Itanium(TM) and McKinley processors on IBM servers running Linux connected through a Qwest fiber-optic network. Once completed the TeraGrid will be capable of a massive 13.6 teraflops and will have access to 450-600 terabytes of data.
    This is a huge step (for Intel at least) in acceptance of the Itanium processor into the server market. Intel is fueling the program by providing optimized compilers and software as well as various customized tools.

    It is being funded by the National Science Foundation by a $53million grant. Various researchers will have access to the system to perform a variety of simulations. Possible uses include :

    -Molecular modeling for disease detection
    -Drug discovery
    -Automobile crash simulations
    -Climate and atmospheric simulations
    -any other approved scientific research purposes

    The TeraGrid will be unique because it will link together various computing clusters at different locations rather than host them all at the same location. Globus is providing open-source protocols that will determine how the grids will communicate with each other. These open-source protocols will create a "plug-n-play" type effect where more machines could easily be added to the network.

    The largest section of the TeraGrid will be hosted at the National Center for Supercomputing Applications (NCSA) at the University of Illinois Urbana-Champaign. There will also be portions of the TeraGrid at the University of California San Diego, Argonne National Laboratory, and the California Institute of Technology.

  62. another Linux cluster by tellurian · · Score: 1

    Argonne has a scalable Linux cluster. It's called Chiba City

    --
    The Grid Report

  63. Itanium not proprietary??? by ivan256 · · Score: 2

    shows how HP has worked to help accelerate the shift from proprietary platforms to open architectures

    Last I checked, only intel made itanium architecture chips, chipsets and firmware, and all the machines are intel reference designs. How is this not a proprietary platform again?

    Even Sparc is less proprietary then this. It's unfortunate that intel and HP can blatently lie, and people will eat it up.

  64. How Tought Is It? by Captain+Large+Face · · Score: 2

    It may have fancy hardware, but is it any good in a fight?

  65. Re:Too Funny... by Dstrct0 · · Score: 1

    M$ Licensing fees!

    --
    Build boards not bombs
  66. The pain of early adopters by AndrewCox · · Score: 1

    They're really going to be kicking themselves when they see an ad on TV a month later for a computer that's just as powerful for only $12 million.

    --
    The Red Pill ... all I'm o
  67. Distributed Computing by Captain+Large+Face · · Score: 2

    How long before distributed computing networks such as those used in the projects by United Devices, SETI@Home and KaZaA :-P are included in the supercomputing list?

  68. nntp by Cadre · · Score: 2

    Wow, I bet this thing could handle at least an eight day retention of a full alt.binaries feed.

    --
    All editorial writers ever do is come down from the hill after the battle is over and shoot the wounded.
  69. Casemoding that baby by fabiolrs · · Score: 1

    "8.3 Trillion Floating Point Operations per Second, 1.8 Terabytes of RAM, 170 Terabytes of disk, (including a 53 TB SAN), and 1400 Intel McKinley and Madison Processors."

    in other news: "The geeks that will use the supercomputer, being loyal slashdot readers, are already planning how they will casemod that baby, they already ordered thinkgeek.com 28 window case kits and 1400 water cooling system"

    --
    Fabio - Sumare/Sao Paulo/Brazil/South America/Earth/Solar System/Milky Way/Universe
    http://www.morroida.com.br
  70. Oh, I just couldn't resist by racerx509 · · Score: 1

    Imagine a Beowulf cluster of those!

    --
    13 year old white supremacists are shitty web designers.
  71. Quake 3 Benchmark by Rcoonz · · Score: 1

    You know what they say :
    "It's not a Super Computer if the Quake3 Benchmarks suck!"
    Ok...i'll stop being funny.....

  72. Ah, but the question is... by Salsaman · · Score: 2
    ...will it run KDE or Gnome ???

  73. IABCOT by Hard_Code · · Score: 2

    that's it

    --

    It's 10 PM. Do you know if you're un-American?
  74. Sorry + retraction by mapnjd · · Score: 1

    Please accept my apologies - it was clearly I who wasn't thinking. I feel like a real twat now. Sorry.

    I retract my previous comment as it is very obviously wrong to anyone with a brain.

    nic (hanging head in shame)

    --
    Bus error in your favour. Collect 200kB
  75. Build your own top 500 by giberti · · Score: 1

    All you would need to do is put together 10 powerMac G4's and you would be at about 150GFlops and bam, your on the list!

    --

    AF-Design, web development.
  76. If you imagine a beowolf cluster of these.... by lumpenprole · · Score: 2


    I'll smash your face in.

    --
    Disclaimer: MINAA (Mummy! I'm Not An Animal!)
  77. Palm by discstickers · · Score: 1

    But does it sync with my Palm?

    --
    I have a shitty sig!
  78. Framerate by Soulfader · · Score: 1

    Tuxracer would be absolutely wicked.

  79. Nice...but by SashaM · · Score: 1

    Can't they just wait for 2006 when it'll be available at the local computer shop for $700?

  80. Re:Myrinet has done gigabit for ages by afidel · · Score: 1

    Umm, no I get full support for that $60k ( I damn better since I can build a pretty damn reliable box with over a TB of space for $3k). In fact the software solution and the support are the only reasons I even use the netapps (they share out files over both nfs and SMB as well as do snapshotting and a host of other things, as for support I get 4 hour turnaround on everything, so I doubt thist cluster comes with better support)

    --
    There are 4 boxes to use in the defense of liberty: soap, ballot, jury, ammo. Use in that order. Starting now.
  81. Finally! by kingharrison · · Score: 1

    Maybe now they have a computer that can run a website that can't be slashdotted!!!

  82. good time to buy storeage.... by Miska · · Score: 1

    IBM leaving the storeage business probably means cheaper drives....

    ... I knew there was _something_ good about it...

    _

    --
    -
  83. Imagine a Beowulf Cluster of These by benthesinister · · Score: 1

    Well, it'd be cool, no? Maybe Unreal Tourney wouldn't lag anymore...

  84. Re:GOOGLE switch to FreeBSD! by Jagasian · · Score: 2

    Well, considering that Yahoo uses Google for its searches, Yahoo uses Linux too. Bet you feel like an idiot now. I mean, if FreeBSD is so tuff, then why does Yahoo have to rely on Google?

  85. Imagine... by cerskine · · Score: 2, Insightful

    Can you imagine a beowulf cluster of these babies?!

  86. Yes by SeanAhern · · Score: 1

    He's a co-worker of mine. We both work at Lawrence Livermore National Laboratory, one of the DOE labs where these machines keep appearing. Computer scientists here program on them all the time.

    1. Re:Yes by jeffy124 · · Score: 1

      that makes sense. I just found it kinda humourous to see "havent used it" in the middle of his list. thanks

      --
      The One Rule Of Chess You'll Ever Need: Don't play someone who carries a kit in their bookbag.
  87. Yes, but... by SnarfQuest · · Score: 1

    Everyone's going to be asking them "Does it run Word?"

    --
    Who would win this election: Andrew Weiner vs Andrew Weiner's weiner.
  88. each node has 1.25gigs of ram, not 12 by Jeff+Knox · · Score: 2

    Each node only has 1256megs of ramwould be my guess. 1800gigs of ram today / 1400 processors. That comes out to be ~1.28gigs. The harddrives are also not that much a terabyte of drives is about 8000 grand using high quality scsi (okthey use fibre channel, close enough price bracket). Still doesnt account for 17.5 grand a processor. Thats a prett high cost per gflop rate.

    --
    Jeff Knox
  89. ObSimpsons by sharkey · · Score: 2

    Check out the bowl-job, Marge.
    --Homer

    --

    --
    "Outlook not so good." That magic 8-ball knows everything! I'll ask about Exchange Server next.
  90. You mean they didn't take the way out? by Kymermosst · · Score: 2

    I thought Microsoft and Unisys had all the answers for high-performance computing.

    --
    "Alcohol, Tobacco, Firearms, and Explosives" should be a convenience store, not a government agency.
  91. Licensing by jridley · · Score: 2

    So, will they have to buy 1400 Windows licenses then throw them in the trash, like the rest of us do?

  92. God Damn... by kclick · · Score: 1

    This thing would make Half-Life SMOKE!!!

  93. Yes, that's nice ... by theCyde · · Score: 1

    Yeah, this computer is nice, but imagine what a Beowulf cluster of these could do ...

  94. Re:Myrinet has done gigabit for ages by fferreres · · Score: 2

    Mhh, so what would you built with 25 grandgrands? Will it work? Will you need to push the OS in other ways? Would you need any custom stuff to be accounted for? Will you need to provide warranties that it will work perfectly? Or you are just guessing you could build the array for just the cost of the hardware + some scatered support depending on each vendor, and hope that everything will run smooth and that they won't have any other requirement than that?

    All these questions need to be adressed before you can complain. And this of course doesn't mean i don't agree with you, they can always save a lot but they just like paying more, for some reason. I've seen companies spend 1 million for some task, and another company spend 10k for the same task. Yet the 10k solution was no worst.

    --
    unfinished: (adj.)
  95. not sure what you ment by systemaster · · Score: 1

    but win XP only supports at most 2 proc. an above post does talk about 2k datacenter which does support 32 proc. So i'm not sure if your just out of it, dumb, or trolling. So I'm sure your math is off, you need to price data center, which is not possible because you can't just buy datacenter the only way to get it is with a big computer, and then you get into the debate of how much of the price is the OS and how much the hardware/support cost...either way your math is off.

    --
    LinuxWorx
    Spelling errors are intentional as are gramatical error