Stories
Slash Boxes
Comments
Search

News for nerds, stuff that matters

Slashdot Log In

Log In

Create Account  |  Retrieve Password

DARPA Awards HPC Contracts To IBM, Cray, Not Sun

Posted by ScuttleMonkey on Wed Nov 22, 2006 05:02 PM
from the no-soup-for-you dept.
IBM Supercomputing IT
snedecor writes "DARPA has awarded a third round of funding for the next-generation petascale computing system. IBM and Cray roughly split the $494M, while Sun, with little track record, received none. This is in spite of Sun's radical proposal for proximity communication."
+ -
story
This discussion has been archived. No new comments can be posted.
The Fine Print: The following comments are owned by whoever posted them. We are not responsible for them in any way.

DARPA Awards HPC Contracts To IBM, Cray, Not Sun 50 Comments More /

 Full
 Abbreviated
 Hidden
More
Loading... please wait.

  • UCAN (Score:5, Funny)

    by Stanistani (808333) on Wednesday November 22 2006, @05:05PM (#16958008) Homepage Journal
    You can pet a dog, you can pet a cat, but you can't petascale.

    • ...you can't petascale.

      Maybe you could hug it with your nuclear arms.

      • First of all, my personal vanity domain is petascale.org. Really.

        Second: it's nice to see supercomputing covered in Slashdot :)

        Finally: these systems are going to be both revolutionary and evolutionary. The main thing you can predict, is that you can't predict what impact and uses they will have.

        • And in 10 years we'll be like, "They paid $500 million for that?? But it can't even run Halo 10!"
  • This is in spite of Sun's radical proposal for proximity communication."

    More likely, that was because Sun's radical proposal for proximity communication.

    • Re: (Score:2, Informative)

      by Anonymous Coward
      IBM's radical idea was to release whatever product they'll have ready in 2010, but with enough processors to reach petascale. Go Blue! Had Sun offered up Niagara2 @ petascale, then they might have had a chance to win as well.

      • Re:more likely... (Score:4, Informative)

        by georgewilliamherbert (211790) on Wednesday November 22 2006, @05:58PM (#16958690)
        Niagra and Niagra 2 have lousy floating point performance (1 FPU for entire chip, shared by all the cores). Given that the DARPA project is for FLOPS, Nigra is just about the worst processor one could propose for the project.

        I love the Niagra design; for 90% of what I need done, it does great. It's just terrible at floating point.

        Sometime down the line, past Niagra 2, one could posit a version of such a chip with enough floating point units that it's efficient in FLOPS; it's an obvious upgrade of the current chips. However, that also is not optimal for FLOPS in the HPC regime. HPC is all about hiring enough computer scientists and physicists to micro-optimize the code so that you make close to theoretical maximum efficiency in utilizing the CPU cores. Niagra is all about keeping enough contexts on the chip that you can productively use the time that normal programs spend wasted, waiting for main memory accesses and so on. HPC by definition spends the CS time and effort to avoid that already.

        • Re: (Score:3, Interesting)

          by SnowZero (92219)
          Niagra 1's FP indeed sucks, but I thought they were fixing that for Niagra 2. I heard about an FPU-per-core design being done (maybe that's a later follow-on?).
        • I thought Niagara 2 is supposed to have more FPUs.
        • It's Niagara

          Niagara II (T2) has one floating point unit per core...so for a T2 outfit with all cores, eight FPUs.
          • It's Niagara

            Maybe he meant Viagra.

            I wonder if I could work Sun into this someow, something about putting it where the Sun don't shine....
        • You might find the history of the company that is now called Cray to be interesting. They used to be called TERA before they bought the remains of Cray from SGI. TERA had a CPU that did something like a thread context switch on every memory access and had no data or instruction cache either. I think they called MTA for multi-threaded architecture.
          • Cray is actually just about to launch a next generation MTA, the XMT [cray.com]. Interestingly, the processors plug into Opteron sockets and except for the processors themselves all the other hw is from the XT4 (Cray's Opteron-based supercomputer).

        • Re: (Score:3, Informative)

          by davecb (6526) *

          Sun's HPC contribution is in optical chip interconnects, described (somewaht fluffily) at http://research.sun.com/spotlight/2006/2006-04-07_ Sun_on_HPCS.html [sun.com]

    • Nuclear Simulation Need Complex Cores (Score:1, Interesting)

      by Anonymous Coward
      The fact of the matter is that Sun's "strength" is in building multicore processors that combine hordes of simple, weak cores. These processors are suited for processing, say, web-based commercial transactions. A web billing server typically runs hundreds of thousands of threads; each thread services one customer among the hundreds of thousands of customers submitting their credit-card information to buy the services sold by, for example, Priceline.com.

      This kind of processor is not suited for the high-p

      • By contrast, IBM is one of the 3 remaining American companies that still makes general-purpose, complex, and powerful cores for crunching scientific applications. The other two companies are AMD and Intel.

        You forgot Freescale [frescale.com], Sun Microsystems [sun.com], Texas Instruments [ti.com], Hewlett Packard [hp.com], geesh, there are actually quite a few others...

        Once you start searching for US chip design and manufacturing firms, you realize that there are tons of them that produce silicon that is general purpose. You only listed the thr

  • FTA:
    DARPA, which funds computing and technological projects for the military
    ...
    These computers will be used to simulate global climate changes or the spread of hypothetical epidemics.

    Now, of the two stated applications, which do you think is more interesting to the military? I suppose one could argue defense against bioterror, but it's still kinda scary.
    • Well, being that global climate changes could trigger large crop failures, mass human migrations (with their attendant civil wars), changes in sea levels, ports and river navigation and all-around Bad Stuff (TM), I'm guessing the first choice might have more real impact than the second.
      • I can see how global climate changes affect military activities / create military situations, but I still feel the military is much more interested in the shorter-term (relatively speaking) effects of bioattacks (whether from the offensive or defensive perspective). The former would certainly have more real impact in the longer term, but I don't think we as a country are good at long-term thinking. :)
    • Defense? Who said anything about defense? Pre-emptive strikes are where it's, baby!

      This bio-attack brought to you by IBM.

    • You might be surprised but the military is VERY interested in climate as well.

      That said, the HPCS program - if continued until completion: whole new COngress coming up here - will produce an HPC platform that will probably end up being used by most HPC sites. Not just DOD ones.

    • Don't forget restricted climate change is a scary attack too. There were rumors that US tried this in Vietnam (colossal downpour).
      --Ram

  • Loosing may be a good thing (Score:4, Interesting)

    by renau (123225) on Wednesday November 22 2006, @05:25PM (#16958250) Homepage
    The HPCA program is a "cover" (not so cover) funding to companies. The problem is that it is not so clear that it is even good for them. The reason is that "lots" of additional resources from these companies are also diverted for these projects. Since these machines have a "doubtful" application besides the DARPA contract, I think that it may be better for these companies to invest on research more related to their product or may-be products.

    For example, Sun Labs was in charge of the DARPA project at Sun. They have "invested" 3 years on that. My question is "what do they have to show?".
    They do not have publications on any top computer architecture conference, they do not seem to have anything that may save Sun ass. (At least from
    an architecture point of view)

    This is not such a strange comment, I have head it from people at IBM research itself. Some people there is not sure that winning is the best thing either.

    • Re:Loosing may be a good thing (Score:4, Interesting)

      by lowoddnumber (814033) on Wednesday November 22 2006, @05:53PM (#16958612)

      Working at Sun I've heard the higher-ups discuss these grants and awards and from what I gather, despite the award money, it is still very expensive for the company. The award money is not enough to fund a competitive entry. IBM definitely has a lot more money to put into the effort than Sun does. Not defending them though. It's a bummer for the company that they didn't get the grant, but I probably agree that losing could be a good thing. Winning would be great for bragging rights and image.

      I really don't know anything.

    • Interesting notion. For Cray it's a no brainer. They only exist in the upper-end of the HPC market.

      For IBM, I'd still say it's a win. If it weren't for some of the early work with blue gene, would there have ever been a cell processor? If their project is more related to power6/7, well those processors are used for the low end of HPC, and for high-end database servers too. Even if the DARPA system requires CPU modifications (VIVA) that don't help the general business user, any advances in memory technology,

    • Re: (Score:3, Insightful)

      by guaigean (867316)
      Since these machines have a "doubtful" application besides the DARPA contract, I think that it may be better for these companies to invest on research more related to their product or may-be products.

      You may want to rethink that. The number of products developed out of DARPA initiatives which have become mainstream are astounding. For now, yes, they may be specialized devices, but the research driven by these funding sources is responsible for home technology 10 years from now. Just because you can't
    • For example, Sun Labs was in charge of the DARPA project at Sun. They have "invested" 3 years on that. My question is "what do they have to show?".

      erh.... $50 million in cash from the last round of funding?

      http://www.businessweek.com/magazine/content/04_23 /b3886007.htm [businessweek.com]

  • I used to work at Cray (SGI vintage) as a contractor. It would be nice to think that some jobs are opening up there.

  • A win for Linux (Score:1, Funny)

    by Anonymous Coward
    A win for IBM is a win for Linux. Too bad for Sun, but congrats to Linux.

    • Re: (Score:3, Interesting)

      by PCM2 (4486)
      This was modded funny, but it's an interesting point. I've been able to tour a couple big HPC type facilities and I've spoken to some academic types who were involved in other projects. From what I can tell, the HPC community is all over Linux. Linux is pretty much the de facto platform of choice, for all sorts of reasons. I understand that a lot of people in the life sciences were enamored of Apple's later-model PowerPC hardware for HPC applications, but even they would tend to reformat and install Yellow
      • OpenSolaris is nearly as free (in every sense) these days as Linux, and it has some great technology in it, including a battle-hardened networking stack.

        This statement implies that Linux's networking stack is not "battle-hardened" or at least not AS "battle-hardened". WTF? Linux is used on routers for a reason - and it's not because of the freecell game!

        Trying to imply that Linux's networking stack is anything but quite mature is just idiocy.
        • It is only implying that Solaris has a battle-hardened networking stack and is free also. Linux is a great kernel, but Solaris has a longer track record in the industry. Linux has come a long way fast, if Solaris was to GPL I think that both Linux and Solaris would benefit.
      • In our HPC project it has been quite essential that the OS was (a flavour) Linux. We rewrote several drivers and replaced the ones from the manufacturer to boost performance for our application, in some cases getting an almost 10 fold increase. (I didn't do this myself, I work more on the user applications)

        This would have been a lot harder with a closed OS.

  • Where was HP? (Score:1, Interesting)

    by Anonymous Coward
    So, IBM and Cray took the prize... Sun lost. And where is the other major US HPC vendor HP? Did they even enter the competition? Do they have anything new to say?
    • H-P and SGI were knocked-out at the end of round one; Cray, IBM & Sun were selected
      for round two.

    • Re: (Score:3, Funny)

      by Anonymous Coward
      HP couldn't figure out how to sell printer ink for a supercomputer. Thus, they have no way of making any money on it.

  • What about Fortress? (Score:3, Interesting)

    by andrew cooke (6522) <andrew@acooke.org> on Wednesday November 22 2006, @09:43PM (#16960778) Homepage
    Fortress, the language being developed by a bunch of people led by Guy Steele, was funded as part of the HPCS effort. This means that DARPA is going with IBM or Cray's language (X10 for IBM, Chapel from Cray). According to a press release quoted at http://www.eweek.com/article2/0,1895,2063043,00.as p [eweek.com] (but not available at http://www.sun.com/smi/Press/sunflash/index.xml [sun.com]) the work will continue, but how likely is it to succeed?

    Guy Steele gave an excellent talk at OOPSLA on Fortress - the slides are at http://research.sun.com/projects/plrg/PLDITutorial Slides9Jun2006.pdf [sun.com] - I thought it was pretty impressive.

    The groups's site is at http://research.sun.com/projects/plrg/ [sun.com]
  • IBM has been a large governmental contractor from the dark ages... No suprise it isnt continuing. ( and no, that wasnt a slam... there are advantages to working with the same companies you are used to doing business with )
  • DARPA must be simulating giant enemy crabs (or even more appropriately, real-time weapon changing)!

    • Re: (Score:3, Interesting)

      by Boone^ (151057)
      Google's HPC "innovations" are surely software-based, are they not? Not sure they've done anything in the hardware arena that warrants press releases.

      • Re: (Score:3, Insightful)

        by flaming-opus (8186)
        Google's advances are very far from traditional HPC applications like fluid dynamics, weather forcasting, solid body simulations, waveguides, thermal reactions, particle dispersion, oil discovery, etc. Google does data mining, and transactional processing. The very problem that the darpa HPCS program addresses, is that the bulk of the HPC systems sold in the US are just clusters of off-the-sheld database/web-optimized servers. It turns out that these clusters don't deliver very high levels of efficiency, ei

    • Re: (Score:3, Interesting)

      by SnowZero (92219)
      Google are experts at scaling out, not up. Running applications to serve 10000 users at the same time is different from executing one massive program. The main way in which Google and HPC (High Performance Computing, i.e. Supercomputers) overlap is in enormous file systems.

    • Re: (Score:3, Interesting)

      by flaming-opus (8186)
      Someone else will probably license Sun's proximity communication technology at some point, but it might be the graphics card makers. The proximity communication stuff lets you hook together multiple chips, almost as if they were part of the same die, using a bunch of capacitor-like plates in the chips. This could be very useful for putting some amount of memory (almost) on the cpu die, and putting a very wide bus between that memory and the processor. Both IBM and Cray currently use very expensive ceramic m
      • I am not asking for the right to compete for those defense contracts, and I was wondering where did I said that. I am telling that this HPC contracts are a smart way to shove billions of dollar of subside for american computer companies. Basically, the US taxpayer is paying for private companies research under the cover of a very broad and vague "Defense Contract" for HPC computing. So, basically, I am claiming for the right to over-tax those products when I import them, because of that subside. It's just i

All trademarks and copyrights on this page are owned by their respective owners. Comments are owned by the Poster. The Rest © 1997-2009 Geeknet, Inc.