back to article Incomprehensible boffins bring quantum computers a step closer

The long-awaited arrival of quantum computers could be one step closer, as boffins from Oz have for the first time encoded quantum information in silicon. Unlike conventional computers, which store data on transistors and hard drives, quantum computers encode data in states of microscopic objects called qubits. The arrival of …

  1. Christopher Reeve's Horse

    At Last!

    A quantum solution to the NP problem could be just round the corner, and then we can have satellite navigation with routes!

    Oh, hold on...

    1. Anonymous Coward
      Anonymous Coward

      Re: At Last!

      Correct me if I'm wrong, but didn't they already prove that a quantum computer cannot improve an NP-complete problem to any significant degree over conventional computing, given that a quantum computer is still, in the most basic sense, deterministic?

      1. Destroy All Monsters Silver badge
        Holmes

        Re: At Last!

        You can't prove that but it seems to be a law of nature (see most anything by Scott Aaronson and in particular NP-complete Problems and Physical Reality and this blog entry) Indeed, whenever you think up a physical machine with quantum powers to collapse NP to P, it fails through some unphysical assumptions you need.

        As P = NP would actually would mean that things that are easy to check are easy to guess, immense powers of godlike computation would ensue and as this trick is not being performed in nature to all available evidence, the P = NP believers and agnostics have to be firmly placed in the camp of cranks and perpetuum mobilists.

        Technically: BQP is strictly smaller than NP.

        1. phil dude
          Boffin

          Re: At Last!

          but then again there are special cases for graph theory called cliques (used for expression array analysis).

          But generally the proof of P=NP is not going to exist without us solving the many-body problem first...

          As the canonical physical problem we know (according to our best understanding) there is NP complexity behind the Schrodinger equation. The universe might need NP complexity to work, but things such as having a finite "c" seems to suggest something else is going on. This implies some of approximations might help at least reduce the search space....far from a solution but might be useful.

          This is sort of what ab initio MD does; a pile of approximations/constraints that yield incredibly accurate results due to a greatly increased phase space for particles (and more accurate energetic states).

          It really does make you wonder what is further on down there...more turtles?

          P.

          1. Destroy All Monsters Silver badge

            Re: At Last!

            there is NP complexity behind the Schrodinger equation

            That's the first time I have heard that kind of statement. You may want to provide citations.

            The Schrödinger equation is a bog-standard differential equation. It solves nothing. You just let it run. Whether you need infinite precision at the density of reals is a matter of discussion, and I remember reading that the symbolic computation of the values would eschew the need of making physical use of the reals.

            1. phil dude
              Boffin

              Re: At Last!

              To solve the SE it starts to look NP complete pretty rapidly, so we approximate it. No citation necessary, just need to know the results we count on and the fact we only have:

              1) Solved for hydrogen

              2). Sort of solved for Helium (Hartree-Fock etc).

              3) Everything else - See 1) (Slater orbitals, HF, DF etc...).

              BTW "Bog standard differential equation" is the understatement of the week ;-)

              This is not really my field (!) but I have written enough MD simulations to know that there are many approximations required, starting with the 2-body truncation of the many-body expansion. You gain very subtle improvements by expanding the terms, hence ab initio (Carr-Parinello) is useful for systems with known geometry (e.g. crystal structures of enzyme catalytic sites), and really useful (though still approximate) chemistry. When I first saw a visualization of protons hopping around ab initio water molecules, the hairs on my neck stood up....! *Phenomenal accuracy*.

              I stand by my comment that the N-body problem is at the heart of why NP != P.

              It is a fascinating constraint on the universe structure though...

              P.

            2. Michael Dunn
              Joke

              Re: At Last! @Destroy All Monsters

              And you wouldn't need to check the cat!

  2. frank ly

    Silicon-28

    It seems that Silicon-28 is the most abundant (92%) naturally occuring isotope, but that still leaves 8% 'impurities' to get rid of. I wonder how difficult/expensive that is.

    1. This post has been deleted by its author

      1. Anonymous Coward
        Anonymous Coward

        Re: Silicon-28

        Which method is very similar to how you do isotopic separation of ic U-235 if you don't like the gaseous diffusion method. Far safer approach, btw. Just sloooow.

  3. Aristotles slow and dimwitted horse Silver badge
    Go

    Incomprehensible?

    It made perfect sense to me once I'd seen through the sciency fog and considered it in relation to 24 Olympic sized swimming pools, 4 London buses... and a set of Bulgarian airbags.

  4. Little Mouse
    Coat

    but but but

    researchers at New South Wales University placed the qubits inside a thin layer of specially purified silicon

    Presumably the trick was to create a kind of pseudo-pyramid of isometric cubes for the qubits to hop around in?

  5. tmTM

    Quantum Dot

    Whats with the pictures of the TV's??

    1. Jimmy2Cows Silver badge

      Re: Quantum Dot

      Quantum, init

      Can't have an article about quantum jiggery pokery without a picture of something entirely irrelevant but still "quantumy"

  6. launcap Silver badge
    Go

    So..

    .. will this give us ansibles[1]? Imagine - two entangled photons captured in separate wafers and (until they de-entangle) changes to one instantly affect the other..

    I want my flying ansible!

    [1] Tips hat to Ursula K Le Guin who originated the term. As well as NAFAL.

  7. TeeCee Gold badge
    Coat

    Quantum Computing.

    I can see the business applications now.

    The main problem with conventional computing in addressing the needs of business is that whenever you come up with a correct answer that causes some bugger to change the question.

    Hopefully a quantum computer will be able to deliver the right answer and the right answer to the new question at the same time.

  8. Pascal Monett Silver badge

    We are now one step closer

    to the positronic brain.

    But really, influencing atoms with local electric charges, turning the knob by modulating the voltage, this all sounds more Star Trekky than Star Trek itself.

    I just hope nobody's going to reverse the polarity....

    1. Mark 85 Silver badge

      Re: We are now one step closer

      That will only work if you can turn the knob all the way to "11".....

  9. Beau
    Holmes

    '' a small electrode placed above the atom.”

    '' a small electrode placed above the atom.”

    May I put this forward as the best understatement in 2015., well, so far?

    1. Michael Dunn

      Re: '' a small electrode placed above the atom.” @Beau

      I just wonder what the "small" electrode is made of if it can be placed over a single atom? A quark whisker?

  10. Anonymous Coward
    Anonymous Coward

    What's quantum dots got to do with quantum computing?

    What's with the picture? Does it relate to the article at all?

    1. Destroy All Monsters Silver badge

      Re: What's quantum dots got to do with quantum computing?

      No

  11. Destroy All Monsters Silver badge
    Holmes

    Unlike conventional computers, which store data on transistors and hard drives, quantum computers encode data in states of microscopic objects called qubits.

    Well, the "qubit" is the mathematical representation of a superposition of two distinct states, whereas transistors and hard drives are physical machines. Apples and oranges?

    Also, the qubit is not very interesting. Interesting is:

    1) Get a bunch of N qubits together

    2) Wire them up so that some Hamiltonian is being implemented

    3) Let the machinery run, thus moving the "state space vector" in the 2^N dimensional complex space (entanglement!)

    4) After some time, measure the qubits along the projective axes, getting a classical 0..1. string with classical probabilities (collapse the wavefunction!)

    5) If 2) was good, 4) is probably the solution to your problem.

    Here, we are at 1).

    1. Michael Wojcik Silver badge

      Also, the last time I checked, modern conventional computers "store data ... in states of microscopic objects". Certainly my eyesight isn't good enough to distinguish the individual transistors and magnetic domains in my computer's chips and drives.

      That whole "Unlike conventional computers" sentence had rather a bit of the Fry-nature.

      And as an addendum to your item 1: for practical interesting problems, the "bunch" needs to be pretty big. Certainly much bigger than anything anyone's demonstrated so far.

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

Biting the hand that feeds IT © 1998–2019