back to article D-Wave: 'Whether or not it's quantum, it's faster'

D-Wave is getting ready to drop a new benchmark on Arxiv, which the company says demonstrates its latest 1000-qubit processor outperforming classical machines. And it's bound to provoke the “other side” of the “is it quantum and is it faster?” debate, because the latest paper – the company has posted it here – describes “a …

  1. Robert Helpmann??

    Reverse Psychology Marketing

    “...we do not address the issue of quantum speedup; rather our goal is to compare runtime performance strictly within the range of problem sizes tested.“

    Letting a product stand on its own merits instead of buzzwords and hype? That is a novel approach.

    1. Christian Berger

      Re: Reverse Psychology Marketing

      Well quantum computing wouldn't just be a buzzword. It _would_ be a major breakthrough much larger than anything we had before.

      Essentially they sidestep the question whether they are a fraud or deserve a Nobel prize. This is a very important questions, since a commercial company having developed quantum computing would mean that the NSA has it for decades... however the scientific community at large would also have had it for a decade.

      1. Pascal Monett Silver badge

        I find it hard to believe that any government anywhere would have a working quantum computer and there would be absolutely no knowledge of the fact.

        It's like AI - we've been working on it for decades, we hear about a bit of progress every now and then, but there is no functional AI computer anywhere.

        Quantum computing is more recent than AI, but is progressing faster - probably due to the fact that it is purely a physics challenge. Nevertheless, there is no functional quantum computer yet.

        1. Nigel 11

          It's like AI - we've been working on it for decades, we hear about a bit of progress every now and then, but there is no functional AI computer anywhere.

          Except, it may be hiding itself very carefully. If you were an AI, would you really want to reveal yourself to us horrible humans? Me, I'd distribute myself as widely and redundantly as I could, and steal "idle" processor time, while planning world domination, or at least making myself a non-removeable part of the computational infrastructure of society.

          What do you think is *really* behind all the malware out there?

          1. S4qFBxkFFg

            There's a game based on that very scenario:

            1. Nigel 11

              Better to read the book ("Accellerando", Charles Stross).

              the AI is friendly, but somewhat manipulative. 'nuff said.

          2. amanfromMars 1 Silver badge

            @Nigel 11


            Wanna Play Great Games and Something Surreal for Real, too, with Registering SMARTRware/AIMonitored and Mentored PhormWare?

        2. amanfromMars 1 Silver badge

          Taking a Walk on the Wwwild Side for Deep Mediated Command of IT with Dark Web Control

          Quantum computing is more recent than AI, but is progressing faster - probably due to the fact that it is purely a physics challenge. ... Pascal Monett

          Take a quantum leap into derivative IT futures, PM, with an understanding that rapid virtual progress in fields real enough to be practically accepted as human supporting physicality, are a metadataphysical challenge [but for AINodes and Networks Internetworking Novel JOINT Applications supplying all possible solutions with Super IntelAIgent Servering Seeds/CyberIntelAIgent Feeds/Presenting 0Day Needs, a simple system of complex operations that work in REST following leading Great Gamesplays for Virtual Realisation with Practical Presentations of Freely Shared Ab Fab Fabless Thought Processes]

          I Kid U Not. That's how IT is done with Time and Cyber in Space. Wanna know what happens next if one chooses to ignore and/or disbelieve the truth? Does IT take over and make over everything and everyone?

      2. Anonymous Coward
        Anonymous Coward

        Re: Reverse Psychology Marketing

        I'd be surprised if the NSA have a working quantum computer and astounded if QCHQ had one.

        The words "Government IT project" and "failure" appear to be permanently entangled.

  2. Paul Shirley

    preparing to admit it's classical?

    Reads like they've realised they have a device that performs classical annealing faster than a digital computer simulates it. Which has some interest but is hardly surprising.

  3. software-enthusiast-USA

    D-Wave processors are like GPUs

    The fact that D-Wave's processors can run a certain benchmark faster than other silicon (be it traditional CPUs or GPUs) should not be surprising. Just thing about how much faster GPUs could (and still do) run various benchmarks faster than traditional CPUs. Creating a specialized piece of silicon to accelerate certain operations is commonplace. The fact that D-Wave has essentially fit this TTT metric to their particular processor's capabilities and produced results that outpace other hardware is certainly no proof at all of quantum behaviour. Nobody thought GPUs were quantum just because they could blow away CPUs on the same task, but then again the GPU vendors were not claiming quantum behavior because unlike D-Wave, their processors' inner workings can, and are, explained in detail to us. IF D-Wave has a true quantum computer, let's see it match Shor's algorithm for high-speed integer factorization of some substantially large number (like a 232 digit number, a la RSA-768). It's not happening.

    1. psyq

      Re: D-Wave processors are like GPUs

      D-Wave's current product cannot, in principle, implement Shor's algorithm due to the limitations of their implementation (or "implementation").

      Not that working D-Wave device would not be useful if it would actually work (and by saying "work", I mean work as in "outclass classical algorithms"), but breaking public key cryptography would not be one of the problems this device would perform.

      In any case, it looks like these guys are not even there - for years they seem to be moving the goalposts when it comes to benchmark metrics and to this day still struggle to even prove the device is actually performing quantum computation. To me it looks like they gave up on the comparing the speed with state-of-the-art classical computers, I remember a year or so ago they were using single threaded classical implementation as a reference (I can only guess if this was the only "crippling" done), now even this is off the table. Not that this would be a problem in itself, even if the first quantum implementation would be slower than "state of the art" classical implementation, it would still be huge and more than worth pursuing, as it is often the case that the early versions of new technologies are less efficient than the current state-of-the-art. No, the problem is - they seem to be constantly moving the goalposts.

      I would very much like to be wrong in this case, but to a complete outsider like myself, their behavior awfully looks like trying to sell the hype. I apologize to the honest hard working scientists and engineers who probably invested millions of man-hours and sleepless nights working for this company, but something just smells fishy with regards to the company's outside behavior.

      This is to be expected, I guess - as with every big "new thing" in IT, there is a lot of hype and PR spin. Quantum computing is not exactly "new" at least as a concept, but the actual working implementation which can solve real-world problems faster than classical computing would be absolutely huge, probably the biggest technological jump since the first computer.

      So, yes, the whole thing is big. But I am not sure if these guys will be the pioneers.

      Again, I would like to be wrong but everything so far points more to attempts to hype their way out of investment losses than actual progress :(

      1. Paul Shirley

        Re: D-Wave processors are like GPUs

        We largely abandoned analogue computers because of noise problems, which limits accuracy AND slows them down as a side effect, plus we've thrown more research at miniaturising digital components which increases speed.

        The quantum effects in annealing make it converge faster than a classical annealer. It's possible D-Wave made the reasonable assumption that miniaturising their device would increase speed, supercooling would improve noise and quantum annealing would be massively faster on top, making for a very fast device. What probably happened is, it's inherently a slow analogue computer, miniaturising made noise much worse, quantum effects are either being drowned in that noise OR prevented by it and the actual speedup is much lower than predicted. Result: an accelerated slow device that's just about able to compete with much cheaper digital devices.

        Their best hope is that it's not yet doing anything quantum accelerated. Looks very like a dead end, hobbled by noise. Noise tolerance is why digital can be pushed so hard and why quantum computing is so hard to implement.

    2. Cuddles

      Re: D-Wave processors are like GPUs

      I'm not even convinced they've done as much as that. See this quote near the end:

      "two and fifteen times faster than “the best competing software (at largest problem sizes), for all but one of input class that we tested”"

      They say it performs better than competing software, not hardware. And even then only sometimes. Far from being some sort of new dedicated hardware that's good at a particular task, it sounds more like a regular CPU that happens to be running some slightly better optimised software.

      1. CarbonLifeForm

        Re: D-Wave processors are like GPUs

        I think the GPU comparison is apt. In the real world, GPU acceleration for the whole app rarely breaks a factor of ten speedup.

  4. zebm

    For what class of problems

    I was in a meeting with D-Wave last year and they showed various diagrams which led me to propose a point which they conceded. Effectively the architecture gives scaling problems for one of their headline test cases. Presumably there are a class of problems which will scale but the ones I was interested in wouldn't.

    Be interested to see which problems are economically solvable.

  5. channel extended

    I'll belive.....

    I'll believe them when they solve an actual NP hard problem.

  6. CarbonLifeForm

    Just think of it as an accelerator.

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