Dev.SN ♥ developers
Sections
Dev.SN
from the CQluaasnstiucmal-Superposition dept.
AnonTechie points us towards updates on the evaluation of D-Wave's annealing devices.
From Phys.org's reporting on the latest tests:
With cutting-edge technology, sometimes the first step scientists face is just making sure it actually works as intended. The USC Viterbi School of Engineering is home to the USC-Lockheed Martin Quantum Computing Center (QCC), a super-cooled, magnetically shielded facility specially built to house the first commercially available quantum computing processors; devices so advanced that there are only two in use outside the Canadian lab where they were built. The first one went to USC and Lockheed Martin, and the second to NASA and Google. Since USC's facility opened in October 2011, a key task for researchers has been to determine whether D-Wave processors operate as hoped using the special laws of quantum mechanics to offer potentially higher-speed processing, instead of operating in a classical, traditional way.
(Score: 2, Interesting) by acapulco on Thursday March 06 2014, @01:56AM
So how does a company like this make such advanced machines without proving it's inner workings through published papers? I know there's most likely a lot of industrial secrets and patents involved, but they are far into the edge of technology that they don't even seem to have any competition. How do they validate they work besides saying "we were able to sell 3 of them!"?
Do you "just" put a lot of resources into R&D and hope your lab boys can come up with technological advancement without confirming with peers? What I find out of the ordinary (but I'm by no means a researcher, so it could be something common happening here) is that as far as I know stuff that has to do with scientific research so advanced and fundamental as quantum mechanics, definitely need confirmation to avoid biases, willing or unwilling mistakes, etc.
Could anyone more informed than me, please explain this in simpler terms: how could they produce an actual product for something so complex without external assistance/confirmation, i.e. scientific research as usual?
(Score: 2, Informative) by SlimmPickens on Thursday March 06 2014, @03:07AM
I have no idea what the truth is however Geordie Rose (the founder of D-Wave) claims in this interview [youtube.com] that there is so much peer reviewed research out there that anyone could build a D-Wave computer. The interview goes for two hours but it's good to play in the background while you play poker or something.
(Score: 2, Insightful) by glyph on Thursday March 06 2014, @03:28AM
Think of something like deep space probes or even the Large Hadron Collider. These are not "products" in the typical consumer sense. You won't know if they are "fit for purpose" until after you test them.
(Score: 1) by acapulco on Thursday March 06 2014, @12:24PM
I think the LHC example is good analogy for this. Precisely my point is that, as far as I know a lot of published research was done before the LHC was approved to be built. Of course no one could be sure if it could work or even if they could actually build it, however there was a lot of research on the matter and thus the people involved could say that it was very likely that it could be both built and it would work.
With D-Wave's product there wasn't (to my knowledge) that much of published research, and that's why I had the doubt.
(Score: 0) by Anonymous Coward on Thursday March 06 2014, @06:15AM
"I know there's most likely a lot of industrial secrets and patents involved"
Patents are supposed to be the antithesis of 'industrial secrets' but I suspect they will do absolutely nothing to reveal any of those secrets.
It's also interesting to note that Canada, and not the U.S., is the ones building this technology. It seems like everyone else is building the new innovation of the future while the U.S. is simply buying it all and falling behind in our knowledge of how it works and our ability to build and document it.
(Score: 1) by egcagrac0 on Thursday March 06 2014, @06:35AM
It doesn't matter what's inside the black box, or how it does what it does.
If the machine does what's expected - producing useful output from input at a desirable rate - it has value.
Unlike my sliderule, I may not understand all the goings-on inside my pocket calculator, but the thing gives me the right answers to the problems, and it seems to be faster to use.
(Score: 5, Informative) by dl6125 on Thursday March 06 2014, @07:05AM
The D-Wave device uses qubits based on superconducting Josephson junctions. The physics behind them has been well understood for some time. D-Wave's advances have been more in engineering than fundamental physics. For example, a significant problem they had to solve was how to reliably manufacture large number of these qubits to within the required tolerance.
There has so far been no rigorous validation that the devices really do give a quantum speedup. This recent preprint [arxiv.org] claims to reproduce the device's behaviour using a classical model. Also, D-Wave's device is only designed to solve a specific optimization problem. It is not capable of universal computation or factoring numbers using Shor's algorithm.
(Score: 1) by acapulco on Thursday March 06 2014, @12:27PM
Ah, I see. Being more of an engineering advance sound reasonable. Interesting. Thanks for the insight. I really wasn't aware (no surprise.. I'm no physicist) that the underlying theory was well known.