back to article Large Hadron Collider gets 4,500 more data-crunching GPUs

Swiss super Piz Daint is getting an upgrade with 4,500 Nvidia Pascal GPUs, replacing 5,200 existing K20x accelerators. The iron, operated by the Swiss National Supercomputer Centre CSCS in Lugano, will have its 7.8 petaflop current performance doubled to nearly 16 petaflops. Announced at GTC16, the upgrade will also include …

  1. Anonymous Coward
    Anonymous Coward

    They shoot particles together and make shrapnel, then they search among the shrapnel for a piece that fits the theory they are trying to prove. Analyzing with a bigger computer doesn't fix the issue there.

    What's needed there is a control.

    Because if the current fundamental particle model is incorrect, (61 fundamental particles), and those particles are not fundamental but themselves made of smaller particles, then you'll find a piece of shrapnel to fit your theory if you try long enough among the shrapnel you're creating.

    There's also the why question. 61 particles, would lend themselves to 61 properties of larger particles (61 at least). Yet we have only 2 properties of larger particles: +ve and -ve charge. So you'd expect a model to be based on just 2 fundamental particles with the only difference between the two being charge! Why 61?

    1. Anonymous Coward
      Anonymous Coward

      Take two particles, one positive, one negative and one force electrostatics.

      The smallest stable particle would be a dipole. If it spins it carries energy. Faster spin, wider circle, more energy.

      Take two dipoles, the force between them is the clumping force. i.e. likes repel, opposites attract, so the dipoles organize themselves to spin in unison, to have a weak attraction force. (model it, if you're unconvinced, at macro scale its well known).

      So, we've taken just three steps up from nothing, and already we have a force outside the standard model.

      So, 61 particles, including neutral particles, yet no macro scale neutral particles (neutrons have a dipole moment and this are made of charged particles). An all pervading field with none zero component that gives mass, yet standard model particles without mass that exist in this field. Forces in the model that selectively apply to some particles not others. Lots and lots of forces, but gravity not covered.

      Mod me down if you like, but that is a summary of where we stand with the standard model.

    2. Anonymous Coward
      Anonymous Coward

      @AC

      What do you suggest for a 'control', a second universe having different properties we can also test against to eliminate false positives?

      Just because they are searching through "shrapnel" doesn't mean they'll find a piece that fits if they look long enough. Regardless of whether our current understanding of particles is correct, or they are made up of still smaller particles, they can only be combined in certain ways which are governed by physical laws. Using theories about those physical laws we make predictions that a certain type of particle at a specific mass (or range of masses) should be found and decay in defined ways. If those theories are wrong the predicted particle won't be found, because this isn't like sticking pieces of gum together where anything will stick to anything in any way you want!

      And no, we aren't limited to just two properties of larger particles being +/- electrical charge. You are forgetting radioactivity and gravity. The only properties of elementary (or what we currently believe are elementary) particles that don't translate in any way to larger particles made up from them are spin and color. Furthermore there are other emergent properties of larger groups of particles like ability to resist, conduct, semiconduct or superconduct, magnetism, ability to conduct or resist flow of heat and so forth - all those properties we use when making computer chips, solar cells, piezoelectric speakers and on and on.

      Is it possible the particles we are familiar with are made up of smaller particles? Yes, but that does not change the validity of these experiments, these experiments would be how we would be able to eventually learn that. Just as we used to not know that atoms contained protons and neutrons, and then didn't know that protons and neutrons were made up from quarks.

      1. allthecoolshortnamesweretaken
        Pint

        Re: @DougS

        Doug, I really like the idea of having a second universe as a control group!

        As it's the weekend cheers!

  2. Crisp

    4,500 Nvidia Pascal GPUs

    Sod how many petaflops they're getting! What's their frame rate like?

    (inb4 But can it run Crysis. (It obviously can.))

  3. Cthugha

    GTA IV

    I still bet GTA IV will refuse to go faster than 30fps.

  4. bon_the_one

    Rift

    At least their Oculus Rift will run nice and smoothly...

    1. Anonymous Coward
      Trollface

      Re: Rift

      and Oculus can helpfully upload, process and store all the data for you as well.

      http://www.theregister.co.uk/2016/04/07/oculus_vr_franken/

      1. allthecoolshortnamesweretaken

        Re: Rift

        Heh, nice thought, route all the data from all the experiments in the NSA's data centres and see what happens... or did somebody already do this and that's why Utah kept having meltdowns?

POST COMMENT House rules

Not a member of The Register? Create a new account here.

  • Enter your comment

  • Add an icon

Anonymous cowards cannot choose their icon

Other stories you might like