back to article How much faster is a quantum computer than your laptop?

I jumped at the chance to interview supercomputing pioneer Bo Ewald and quantum computing whiz kid Murray Thom a few months ago. Although it's been in my “vault of lost content” for a while, the video serves as a good primer for quantum computing and its promise. It turns out that there are three broad categories of problem …

  1. John Deeb
    Boffin

    Quanum leap... in PR magic?

    The only things appearing to be solidly quantum-supercharged are those press releases on QC these latter days.

    The last line of that Techtime link reads interestingly: "The findings of this study are featured in a scientific paper which is yet to be peer-reviewed".

    Compared to e.g http://science.sciencemag.org/content/344/6190/1330.full

    "Quantum or not, controversial computer yields no speedup", Adrian Cho, Science 20 Jun 2014:

    Best bet seems then to not to hold your breath.....

    1. NP-HARD
      FAIL

      Re: Quanum leap... in PR magic?

      "A D Wave system would consider all of the possible solutions at the same time..."

      See headline of Scott Aaronson's "Shtetl-Optimized" blog for why this statement is unadulterated dreck.

      "...It's more complicated than I'm making out, of course..."

      No it's not, it's simply wrong. And the layman still wouldn't understand.

      1. Anonymous Coward
        Anonymous Coward

        Re: Quanum leap... in PR magic?

        Of course it's wrong. Even Canadian Prime Minister Justin Trudeau got it wrong.

        Quantum computers are powered by cats, boxes and boxes of cats. Any fule knows that!

      2. Anonymous Coward
        Anonymous Coward

        Re: Quanum leap... in PR magic?

        "And the layman still wouldn't understand."

        Or indeed any peer reviewers in physics.

      3. Paul Shirley

        Re: Quanum leap... in PR magic?

        Yep, not only does it not 'consider all possible solutions at the same time' it's not even guaranteed to consider 'all solutions' at all. Left wondering if they're even comparing equivalent quality results for those timings.

        1. Anonymous Coward
          Anonymous Coward

          Re: Quanum leap... in PR magic?

          I did wonder when it said "an optimization problem, solving it in mere seconds": how do they know the solution is optimal? Can it provide a proof of this?

          Otherwise, how about they publish the problem and the solution they came up with, and see if anyone else can generate a better solution. If someone can, it proves the "quantum" setup is nothing more than a heuristic that can provide a reasonable (but non-optimal) solution.

      4. Michael Wojcik Silver badge

        Re: Quanum leap... in PR magic?

        No it's not, it's simply wrong

        Indeed. That sentence in the article is a fine description of what a QAC system like the DWave (assuming that machine even does what its vendor claims) does not do.

        But then the whole article is rubbish. A quantum computer, of whatever sort, is only the best choice for a problem if it's economically justifiable. Even if we had QCs today that had useful capacity and operation times better than simulated QCs or conventional algorithms, they wouldn't automatically be the best choice for Monte Carlo simulations or optimization problems - it would depend on the economics of the individual problem. If it's small enough to be amenable to straightforward conventional computation, a QC will never be the better choice. Even if not, conventional algorithms for approximate solutions are good enough for a great many problems.

        And QC for machine learning? Rarely useful, in my opinion, and even more rarely necessary. Current ML systems are pretty much at the point of diminishing returns using conventional hardware. QC won't make them better.

        I haven't watched the interview (because, ugh, video), but if the article is anything to go by, it's not worth the time anyway.

  2. scrubber

    I do not believe these numbers

    This is obvious bullshit.

    If the system works as a quantum computer, and isn't just a regular machine heavily overclocked and set up to excel in quantum-type problems, then I'm sure it is orders of magnitudes better. But we'd have real numbers and not 100 million+ or the sniff test-failing seconds to hours magic number of 3,600.

  3. Anonymous Coward
    Anonymous Coward

    OK, I promise not to get out over my skis. I wouldn't know how.

    1. gerdesj Silver badge

      "OK, I promise not to get out over my skis. I wouldn't know how."

      Me too. I've been a skier for over three decades now (not all the time) and can't quite work out what that means. From the context I think: "Let's not get ahead of ourselves" - a far more logical and understandable phrase!

  4. scrubber
    Boffin

    Quantum computers

    Assuming Quantum computers are thousands of times faster than the best supercomputer for problem X, how do you go about checking the machine got the right answer if it takes a several hours to run the program? It would take years on a regular supercomputer. Do we check the simple ones then take its word on anything that takes over a few minutes? Seems like a great way to sell a modern Mechanical Turk.

    1. Anonymous Coward
      Anonymous Coward

      Re: Quantum computers

      @"how do you go about checking the machine got the right answer"

      DWAVE is an analogue computer running a sort of annealing algorithm configured using circuits and magnetic coils.

      Lots of problems, that can be described as annealing problems, already have faster digital algorithms to solve them. So you can run the algorithm in a DWAVE as an annealing problem and in a computer using a different method, and see if the computer result is more optimal (i.e. better) than the DWAVE. So you know the DWAVE didn't return the optimal result.

      IBM did this to diss DWAVE, pointing out that actually digital computers running digital algorithms are faster than DWAVE running annealing. Whereas DWAVE promoters generally compare DWAVE annealing to simulated annealing on a digital computer to claim the speedup. But that would only be faster for problems where no suitable digital algorithm exists.

      The other way its done is to run an annealing algorithm in DWAVE and simulated annealing in a computer. And simply run the computer for longer till its result is better than the DWAVE. The computer doesn't need to get a perfect result, because we don't know the perfect solution. It only needs to get a *better* solution than the DWAVE to know the DWAVE didn't find the perfect solution.

      Feynman, the physicist who coined the phrase Quantum Computer, did his work in the 50s and 60s, where computers were often analogue computers. And he imagined such a computer that could work instantaneously (using entanglement) and always get the optimal solution (using superposition). A sort of brute force that tries every possible solution.

      DWAVE isn't that. It's a sort of annealing circuit.

      It's also clearly not a quantum computer in the Feynman sense. No superposition and no entanglement. It's very clever marketing though.

    2. Dave 126 Silver badge

      Re: Quantum computers

      >how do you go about checking the machine got the right answer if it takes a several hours to run the program? It would take years on a regular supercomputer.

      Checking a single answer can be done near instantly on a classical computer, for many types of problem. The issue is, you have 10^daft possible solutions to choose from - and it is this area that true quantum computers will excel (if someone ever manages to build one).

      If you are still doubtful about your quantum algorithms, you could test them with problems to which you already know that answer (because you have constructed the problem for this purpose, from the answer and working backwards)

  5. Richard Wharram

    Wrong question

    What we really need to know is whether it can run Crysis.

    1. Thoguht Silver badge

      Re: Wrong question

      Of course, quantum computers are fantastic for gaming. No more agonizing about which path to take or which weapon to use, you can make all possible choices at the same time and then select the best outcome.

  6. James Hughes 1

    My, what a lot of cynical people we have with us today.

    Can people not accept that sometimes, something does turn up that really is as good as people think/hope, rather than sitting in their armchair and pontification on something they are not experts in.

    1. FrogsAndChips Silver badge

      Google a bit about D-Wave, and you'll understand why a lot of people are sceptical about their claims. So far they've shown no strong evidence of achieving quantum properties in their computers, and their test results have been outperformed by optimized software running on classical computers.

      1. Paul Shirley

        The unqualified use of 'quantum computer' plays on expectations of what the device actually does. 'Adiabatic quantum annealer' risks too many noticing it's not a general purpose quantum logic device, which is what most tend to think 'quantum computer' means.

        A quantum annealer would be a useful device in theory for the problems it can solve, there's little evidence it's actually any faster in real life so far.

    2. DropBear Silver badge

      "My, what a lot of cynical people we have with us today."

      And it's a bloody good thing too.

      Cynic (noun):

      - "a person who believes that people are motivated purely by self-interest rather than acting for honorable or unselfish reasons." So basically anyone with actual functioning eyeballs.

      - "a person who questions whether something will happen or whether it is worthwhile." I certainly hope so - when exactly did questioning things become unfashionable outside countries with "beloved leaders"...?

      1. DougS Silver badge
        Boffin

        If it isn't any good

        Why are Google and the US government using them? I know they aren't really quantum computers in the proper sense, but if they can solve a certain class of problem much faster than a regular computer does it really matter.

        BTW, it was Google who claimed the 10,000 times faster on a 500 qubit problem, based on their own measurements. Either the DWave can do what they claim, or they are running a scam so good they are able to fool Google's army of PhDs. Either way, the DWave folks should have demonstrated to anyone's satisfaction that they are pretty damn smart!

        1. Paul Shirley

          Re: If it isn't any good

          @DougS: they do it because a real quantum annealer could be a genuine improvement. Problem is D-Wave keep demonstrating significant speed improvements over classical implementations only to see the classical version improved to beat them, usually on hardware at 1% of the cost.

          At some point they'll either put together enough usable qbits to beat classical hardware on speed, cost or both. OR they'll hit a noise wall and fall out of the race. We're a long way from finding out which.

        2. Michael Wojcik Silver badge

          Re: If it isn't any good

          Why are Google and the US government using them?

          They have money to burn.

          it was Google who claimed the 10,000 times blah blah whatever.

          It was some group of people who work for Google. It's not like the entire technical workforce looked closely at what happened in that experiment. And, in any case, as any number of people have pointed out, the "DWave beats X" results are usually followed shortly thereafter by "X' beats DWave".

          the DWave folks should have demonstrated to anyone's satisfaction that they are pretty damn smart

          Only for a rather credulous audience.

  7. Anonymous Coward
    Anonymous Coward

    I'm uncertain about that but it's great for pictures of cats.

  8. Efros

    But...

    Will it play Crysis?

    1. Adam 1

      Re: But...

      Yes

      and No

  9. Mingy

    Get back to me if, as, and when, Dwave solves a commercially relevant problem

    Researchers using Dwave have a remarkable capacity of testing the machine against non-optimized competition.

    Let us know know if they ever solve a commercially relevant problem several orders of magnitude faster than the same dollar value of traditional computing running an optimized algorithm.

    1. Anonymous Coward
      Pirate

      Re: Get back to me if, as, and when, Dwave solves a commercially relevant problem

      "Get back to me if, as, and when, Dwave solves a commercially relevant problem"

      Methinks the "article" proves that Dwave has solved a commercially relevant problem: How to advertise its exorbitantly expensive white elephants.

  10. JimmyPage Silver badge
    Boffin

    Maybe look to nature ?

    There was a pretty good (by current BBC standards*) documentary about quantum effects and how nature may have jumped the shark on this one ...

    http://www.bbc.co.uk/programmes/b04v85cj

    The way that plants extract energy from sunlight in a quantum fashion is intriguing to say the least.

    *I was watching "Horizon" in the 1970s about quarks and the like, without the science-lite approach the modern age demands.

  11. JeffyPoooh Silver badge
    Pint

    Time to solve one of those "10^77 years" IT Security Cracking problems

    Given that the crackers can crack these sorts of "10^77 years" IT Security thingies in about five weeks, a good quantum computing system should be able to crack it in an hour.

    Past time for DARPA to set the rules for a formal competition. Clear away the smoke and mirrors.

    1. Dave 126 Silver badge

      Re: Time to solve one of those "10^77 years" IT Security Cracking problems

      >Past time for DARPA to set the rules for a formal competition.

      No, no it isn't.

      Breaking conventional cryptography requires a class of quantum computer that Shor's algorithm - which D-Wave has never claimed to be capable of.

      Quantum computers than can run Shor's algorithm have been built in the lab, but only up to a few qubits - far too small to be of any use in factorising large numbers (and thus breaking encryption as we now use it). In the event of researchers making significant progress in this area, DARPA or the NSA would not have to hold a competition - they'll already have been paying for the lab. In any case, it would be trivial for even an amateur mathematician to quickly determined if it worked or not.

      The D-Wave is a different beast. The smoke and mirrors to which you refer are due to the difficulties in finding fair tests. Still, there's enough uncertainty for Google to take a punt on it, as they should if there is even a small chance it can be made to work for them, finding low points in fitness landscapes.

      1. Michael Wojcik Silver badge

        Re: Time to solve one of those "10^77 years" IT Security Cracking problems

        Breaking conventional cryptography requires a class of quantum computer that Shor's algorithm

        Asymmetric cryptography, and that statement (the requirement for Shor's algorithm, i.e. a true QC) hasn't been proven. There may be better conventional algorithms for factoring, for example. Nor is ECC proven to be NP-hard, at least the last I checked.

        For symmetric cryptography, the best QC attack is Grover's algorithm, which effectively cuts the key in half. That's pretty much useless if defenders are using a decent key length to begin with, unless the QC is much faster than a conventional computer at conventional computation as well.

        And then we have post-quantum cryptographic algorithms, which are finally becoming feasible with the RLWE family.

        And then there's the question of how many encrypted messages are worth pointing a large and very expensive QC at. Particularly when there will almost certainly be better attacks available: protocol flaws, implementation flaws, social engineering, suborning one of the keyholders...

        Quantum cryptanalysis is mostly a big fat non-issue.

  12. Captain DaFt

    "How much faster is a quantum computer than your laptop?"

    Hard to tell. Measuring it keeps changing the results, but the real bugger is that it's never where I left it after I measure it. :/

  13. Draco
    WTF?

    I don't claim to understand it, but ...

    The problem I see with quantum computing is that everything I read about it (I have not read any "official" papers though) always sounds the same (in style and substance) with those inventions that claim to be revolutionary in energy use / production. So, I read about it with a very sceptical eye.

    As I understand it (from the popular press), the magic of quantum computing lies in it evaluating all possible solutions simultaneously and popping out the correct answer blazingly fast. Now, for me, the question is "how do you code something so that all possibilities are evaluated simultaneously?" It seems obvious to me that if you are coding all the permutations, then, obviously, you are also coding the answer in. Once the quantum haze (or should that be foam?) dissipates, presto, you have the right answer. But ... it would seem that the quantum computer had no choice but to reveal the correct answer, since all other answers would be "unstable" (so to speak) and thus collapse away.

    As I said, I don't pretend to understand this one bit ... which is why the WTF? icon.

    1. Michael Wojcik Silver badge

      Re: I don't claim to understand it, but ...

      My recommendation is to read articles like this carefully, with the understanding that pretty much every claim in them is wrong.

  14. allthecoolshortnamesweretaken

    Whenever something sounds too good to be true, it usually is. But I'd be happy if Dwave would prove me wrong on this. The potential uses of quantum computers are very interesting, but so far I have yet to see anyone delivering what they promise.

  15. Anonymous Coward
    Anonymous Coward

    What has this got to do with DevOps?

    I thought i was reading something on the reg? Confused...

  16. FrankyBkk

    I do not understand the negative tendency

    Why people are always so negative about things like this?

    It does not matter if QCs fail or dont fail. It does also not matter if they do some work slower or faster.

    Important is to put money in science and development and experiments.

    Failures are part of later success!

    I think that Google do the right thing. Try it.

    The US government had 4,7 trillion us$ to help banks to bailout ... and they are whining when they have to invest 80 billion US$ (also called peanuts) for science.

    Why we look up to the NASA? The tried the impossible and went to the moon. They had bad times and good times.

    But we humans have to try, invest money in science is the best what can happen. This includes investment in computer which are maybe a failure. But it bring us forward.

    Science is the soul of mankind!

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