Slashdot Mirror

← Back to Stories (view on slashdot.org)

Supercomputing, There's an App For That

Posted by samzenpus on from the in-the-palm-over-your-hand dept.

aarondubrow writes "Researchers at MIT have created an experimental system for smart phones that allows engineers to leverage the power of supercomputers for instant computation and analysis. The team performed a series of expensive high-fidelity simulations on the Ranger supercomputer to generate a small "reduced model" which was transferred to a Google Android smart phone. They were then able to solve engineering and fluid flow problems on the phone and visualize the results interactively. The project proved the potential for reduced order methods to perform real-time and reliable simulations for complicated problems on handheld devices."

18 of 66 comments (clear)

  1. Re:appx. by flaming+error · · Score: 4, Insightful

    It sounds like the supercomputer generated an algorithm for the smartphone to run. I guess they can call that "leveraging the power of a supercomputer" but implying the phone app is doing supercomputing stretches things a bit far. I call misleading headline.

  2. PR Bullshit by pigwiggle · · Score: 5, Informative

    The money quote "This is not the first time that model reduction algorithms have been used to ameliorate the complexities of large-scale physical simulations. The advantage of the system designed by Knezevic and his colleagues is its rigorous error bounds, which tell a user the range of possible solutions, and provide a metric of whether an answer is accurate or not. The error bounds are based on mathematical theory developed in Prof. Patera's research group at MIT over a number of years. "

    The research is about error bounds on coarse grained models. The smart phone is just hype.

    --
    46 & 2
    1. Re:PR Bullshit by oldhack · · Score: 3, Funny

      Why are they spamming slashdot anyways? Has the academia gotten so pathetic that they deem any exposure is a good exposure?

      --
      Fuck systemd. Fuck Redhat. Fuck Soylent, too. Wait, scratch the last one.
    2. Re:PR Bullshit by WankersRevenge · · Score: 3, Funny

      I didn't realize you could run algorithms on a smart phone ... these guys are brilliant! ;)

    3. Re:PR Bullshit by blair1q · · Score: 2, Insightful

      Everyone is spamming slashdot, and the people voting on the firehose are generally too lame to understand it. Throw "reduced order methods to perform real-time and reliable simulations" at them and they click the + just to look smart.

    4. Re:PR Bullshit by blair1q · · Score: 5, Funny

      Go into settings and select "scientific mode". That makes a lot more buttons appear.

  3. Not even... by Monkeedude1212 · · Score: 3, Insightful

    The team performed a series of expensive high-fidelity simulations on the Ranger supercomputer to generate a small "reduced model" which was transferred to a Google Android smart phone

    This is like saying that watching Toy Story on your iPhone leveraged the massive renderfarm used by Pixar.

    1. Re:Not even... by Ambitwistor · · Score: 3, Informative

      Not really. In this case, the smart phone isn't simply rendering output of a supercomputer simulation.

    2. Re:Not even... by Ambitwistor · · Score: 2, Interesting

      What the phone is doing is "reduced order modeling", which (if the article is using the term accurately) finding a simple set of equations whose solution provably approximates a far more complex system of equations. It's not just interpolation (lookup tables, or machine learning from a training ensemble). Reduced order models actually have dynamics in them.

      The potential innovation here, as I see it, is that it takes some supercomputing effort to build the reduced order model for a specific problem. So you can input a description of your problem into the phone, communicate it to the supercomputer and let it analyze the problem for a couple hours, and it spits out a custom approximation to your problem that you can run in the field on your phone.

      Now, what I'm not sure that they've developed is some convenient way to specify the problem using the phone's interface. Maybe the problems they're working with are too complex to be easily specified with a few keystrokes. If all they're doing is loading a precomputed reduced model onto the phone, and there's no interaction with the supercomputer to let it handcraft solutions to problems in semi-realtime, then it's not nearly as interesting.

    3. Re:Not even... by Ambitwistor · · Score: 3, Informative

      I just re-read the article. What it sounds like they're doing is having the supercomputer craft a reduced order model which is optimized to a particular range of parameters. That suggests to me that they're constructing a perturbative model about some fixed solution that the supercomputer produces. Perturbative approximations are more accurate the closer they are to the "reference" solution. So the innovation appears to be: the user can specify what set of parameters they want to perturb about, and therefore construct a custom model which is optimized to perform well in the parameter range that user is interested in.

  4. Semantics, maybe... by mea37 · · Score: 2, Informative

    ...but I'm going to go ahead and argue that they are not "performing supercomputing on a phone", because that kind of marketing doesn't belong in research.

    Yes, it could be very useful; I have no doubt it's just as useful as they claim. And yes, it allows someone in practice to solve a problem "in the field" with a phone, when otherwise a supercomputer might have to be used.

    But the supercomputing was done on a supercomputer in advance, when the reduced model was calculated. Its just that instead of giving one specific answer for one specific input, the supercomputer is returning an algorithm that will approximate the answer within known error bounds for a specified domain of inputs. Executing the algorithm isn't supercomputing (if it were, you couldn't do it in a few seconds on a phone); it's using the fruits of the earlier supercomputing that produced the algorithm.

  5. Put CUDA on a phone, then we can talk by jpapon · · Score: 2, Insightful

    I'll believe they've created mobile supercomputing when someone puts a powerful GPU that is CUDA-ready in a smartphone.

    Of course, you better get some big batteries for your phone, because Teraflops ain't free

    --
    -- Let us endeavor so to live that when we pass even the undertaker shall be sorry. -- M. Twain
    1. Re:Put CUDA on a phone, then we can talk by blair1q · · Score: 2, Funny

      CUDA isn't supercomputing. CUDA is more like super-doopercomputing. And it's a fucking crime that nVidia isn't doing better in the market with it.

    2. Re:Put CUDA on a phone, then we can talk by Entropius · · Score: 2, Interesting

      The lattice QCD people, at least, are porting their code to CUDA just as fast as they can. The bottleneck right now is that there's no good way to get multiple GPU's to communicate (quickly). So, for the largest problems (simulating a 64x64x64x192 lattice), you still need a conventional supercomputer (like Ranger, the one in the article here), because it's just too huge to put on a single GPU and multi-GPU doesn't scale well.

      But for smaller problems (like a 24x24x24x64 lattice), GPU's will be great, and people are developing this capability as fast as they can.

  6. FEMM for android. by Facegarden · · Score: 2, Insightful

    I've been using FEMM lately for some magnetics stuff I've been working on. I would LOVE an android port, or some way to run simulations from my phone.

    I don't *really* need it, but its just funny how something like that is actually possible these days. We probably will have supercomputers in our hands someday. I mean, current phones already are supercomputers by the standards of what...? 30 years ago? 20 years ago?

    Smartphones will become the tricorders of the future, its inevitable.
    -Taylor

    --
    Worldwide Military budgets: $2100 billion. Worldwide Space Exploration budgets: $38 billion. Really, world? Really?
  7. Re:Yes smartphones can display results from by Kurofuneparry · · Score: 2, Informative

    That's an RTFA comment right there.

    This isn't just a UI, it's a reduction of the algorithm provided by a supercomputer. However, I believe that this first set of lines is misleading, inaccurate, and likely an example of the writer not knowing what they're talking about:

    What if you could perform supercomputing calculations in real-time, on your smartphone ... Researchers ... have created an application that does just that.

    It doesn't do supercomputing because it isn't a supercomputer, it just makes an educated guess based on sitting at the supercomputer's knee and playing "monkey see, monkey do". Not a bad trick but the claim's overwrought.

    --
    ...... and idiots rule the world....
  8. Re:Seems inefficient by Bill+Barth · · Score: 2, Insightful

    You don't understand how this works. You do the computation ahead of time on the supercomputer to build your reduced order model which you download onto your phone and take out into the field. Once you've downloaded the model, you don't need the supercomputer any more. You can use the phone to do computations using the reduced model as much as you like. If you get into a regime where the predicted error from the reduced order model is too high, you can go back to the supercomputer and update the model. If that happens, then you'll probably have to wait in queue again, but that's not such a big deal.

    --
    Yes...I am a rocket scientist.
  9. Re:marketing by jwpeterson · · Score: 2, Informative

    Furthermore the article has no details on how the error bounds are calculated.

    Good point, if you are interested in the details of the error bounds, please check out our preprints below, and the references cited therein.

    http://www.tacc.utexas.edu/~peterson/articles/2010_rboomit_cmame_preprint.pdf
    http://www.tacc.utexas.edu/~peterson/articles/2010_hafs.pdf