Performance of quantum computer no better than ordinary PC, say analysts

http://www.csmonitor.com/Science/2014/0619/Performance-of-quantum-computer-no-better-than-ordinary-PC-say-analysts

by aasarava • 12 years ago

46 42 12 years ago

mpthrapp2• 12 years ago

The title is slightly misleading. This article is only talking about the performance of one specific quantum computer. (The D-Wave Two).

VonGuard4• 12 years ago

Are there others that you can buy? As far as I know, D-Wave is the only one out of the lab. I'm totally a D-Wave skeptic, but I do have to state that the scientists who tested this thing sound like idiots, unless they really did try quantum annealing. If the D-Wave guys say they tried the wrong set of equations for the device, I'm inclined to believe them.

Even though I don't even believe this thing is really a quantum computer, there are much more specific and defined ways to poke holes in D-Wave than just by running some tests and comparing speed. D-Wave has, very specifically, targeted this machine at one function: http://en.wikipedia.org/wiki/Quantum_annealing

Unless you're doing quantum annealing on the D-Wave, it might as well be a toaster, an Android phone, or a Windows PC. It's made to do only quantum annealing.

Did D-Wave 2 somehow expand its capabilities? I'm basing this on D-Wave 1 knowledge, but I cannot imagine anything really changed except the number of Qbits inside the thing.

All this having been said, the D-Wave has yet to prove that it is even doing Quantum Annealing, let alone that it's faster than a traditional computer at this problem set. No one has shown it to be faster than traditional outside of D-Wave.

mpthrapp2• 12 years ago

I'm not as up on Quantum Computing as I'd like to be, but the title seems to imply that QC as a concept is no faster than regular computing. Which may be true for general purpose computing, but that's never what QC was intended for.

shoyer• 12 years ago

Yes, of course they did what D-wave calls quantum annealing. As you say, that's all the D-Wave machines can do.

The D-Wave executive's claim that the problems used in the study are "not at all the right choice for probing a quantum speedup" is in reference to the specific subtypes of quantum annealing problems that this study looked at, not whether they were doing quantum annealing at all.

Xcelerate2• 12 years ago

The title of the article "Performance of quantum computer no better than ordinary PC" is just flat out wrong. The real issue that physicists are concerned about is whether the D-wave machine is a quantum computer or not. The issue is not whether quantum computers are "better" than classical ones.

Assuming the correctness of quantum mechanics, a quantum computer WILL be faster than classical computers for specific classes of problems, exponentially faster in some cases. This has been mathematically proven. (EDIT: No it hasn't apparently. See Michael Nielsen's response below.)

michael_nielsen3• 12 years ago

Your first paragraph is quite right. But the second paragraph is wrong: for (non-oracle) problems quantum computers are merely suspected to be faster in principle than conventional classical computers. It's never been proven.

Xcelerate2• 12 years ago

I guess I'm confused then. I thought that quantum algorithms had been proven to be faster than the best known classical algorithms (for instance, Shor's factorization), assuming the correctness and completeness of the postulates of QM. In other words, the order of the algorithm is derivable from the postulates. Is that not the case?

I think there's a few people (t'Hooft if I remember correctly?) that believe a quantum computer will never be able to experimentally demonstrate this speed-up because some aspect of quantum mechanics is incomplete/incorrect.

Edit: Here's the thread I was thinking of where t'Hooft and Shor discuss this http://physics.stackexchange.com/questions/34217/why-do-peop...

tzs• 12 years ago

What about searching an unsorted database? Quantum can do it in O(N^0.5). Hasn't it been proven that classical computers need O(N)?

colanderman• 12 years ago

Calm down dude. "Quantum computer" in the title refers specifically to the D-wave. Headlines leave out specifiers all the time.

ex.: "Man bites dog" doesn't mean "all of mankind bit a dog", it means "this man bit a dog".

EpicEng3• 12 years ago

Except no one, ever, would take "dog bites man" as "one dog bit all of mankind". The title is misleading and, at least one person (myself), took it to mean "quantum computing in general".

P.S. why did you go with "man bites dog"?

baddox2• 12 years ago

It's just a slight grammatical ambiguity in the title, worsened by the "headline grammar" and the fact that it's a newspaper with a general audience. The headline is intended to be interpreted as "the performance of a particular computer that is marketed as a quantum computer is no better than the performance of an ordinary PC."

But it's pretty easy to quickly read the headline and interpret it as "the performance of ideal quantum computers is no better than the performance of ordinary PCs."

Assuming the correctness of quantum mechanics, a quantum computer WILL be faster than classical computers for specific classes of problems, exponentially faster in some cases.

I'm not sure about this part either. Some smart CS guys apparently think that quantum computing isn't even possible. Leonid Levin (yes, of the Cook-Levin theorem seems to be one of them: http://www.cs.bu.edu/fac/lnd/expo/qc.htm.

throwwit• 12 years ago

For a performance reference similar to Moore's law, check out the quantum version (Rose's law): https://2.bp.blogspot.com/-AZFZhBUmwds/UHT2yF39AHI/AAAAAAAAI...

dm25• 12 years ago

Here is another article today bashing the D-Wave Two computers: http://www.wired.com/2013/06/d-wave-quantum-computer-usc/

These machines are testing a completely new computing concept. We're trying to learn how they work so that we can apply that to building machines that can actually outperform traditional computers, and once that's achieved then the potential benefit will be enormous.

According to http://www.dwavesys.com/d-wave-two-system as you scale up qubits the power demands do not increase. That alone has huge potential.

Maybe new materials need to be developed to see the full potential of these computers, lots of research needs to be done. Why are all of these articles so pessimistic?

NASA, Lockheed, and Google purchasing these multi-million dollar machines should indicate that there is at-least some potential and value in the D-Wave computers, even if in the end they just learn what D-Wave did wrong.

https://www.youtube.com/watch?v=CMdHDHEuOUE

"We don't know the best questions to ask the quantum computers, that's what we're trying to find out now."

JoeAltmaier• 12 years ago

So negative because there's still a non-zero probability that its not going to work at all, ever. That there's something about quantum computing we've figured out wrong, and no benefit is possible.

VonGuard4• 12 years ago

They're also negative because no one has actually proved that the D-Wave is, in fact, doing any actual quantum computing. That's been the issue all along with D-Wave: no one can prove that it's really doing anything with those qbits.

dm25• 12 years ago

My point is that even if we can't prove the D-Wave Two is faster at any problem than a traditional computer that the massive amount of knowledge gained about quantum computers will outweigh the costs.

It's far too early to worry about a non-zero probability of it not working, just like fusion power. All it takes is one breakthrough and a revolution in computing will begin. Doing research on quantum computers is a low-risk and high-reward situation. None of these companies are betting everything on quantum computers.

AUmrysh• 12 years ago

We're trying to learn how they work

Is the operation of the D-Wave an unknown? I don't know much about it.

dm25• 12 years ago

No, we don't have a perfect understanding of what's going on.

The qubit processor is like an intermediate step, instead of a traditional processor, the grunt of the work of special problems is handled by the qubit chip.

The data is put in and the results are as expected, but they are not as fast as modern computers, so now the challenge is figuring out if different types of calculations can increase that performance or what can be done to improve the chip or other methods.

https://www.youtube.com/watch?v=HKUZ6IuJyHw

The machine itself seems so simple though from a visible hardware perspective, most of the D-Wave Two is just to keep the main chip cool and shielded.

It's just experimenting right now. The internet was probably dismissed by many because of it's enormous expense to implement, but the value could be envisioned by many and it paid off.

There is potential for future versions of these quantum computers to be able to solve problems that massive amounts of supercomputers would have trouble solving, and the energy usage is enormously less (same for a small calculation as a huge calculation).

Encryption is one thing that could be hurt by this technology. Instead of millions of years to crack encryption it might be solvable instantly. That's all just what I remember reading about the subject, there's plenty of information out there that's probably more reliable that my babble.

ColinWright• 12 years ago

Also:

https://news.ycombinator.com/item?id=7917482

https://news.ycombinator.com/item?id=7917749

icegreentea• 12 years ago

This is the paper in question: http://www.sciencemag.org/content/early/2014/06/18/science.1...

Which appears (I can't tell, paywalled) to be an update or this one: http://arxiv.org/abs/1401.2910

I don't have to time to read the article properly right now, but doesn't look like they did anything ridiculous in their choice of testing algorithm.

robgibbons• 12 years ago

Firstly, this is the first of a new class of computers, so I am not surprised it may be slow at first. Secondly, this machine uses a relatively small number of bits, compared to a classical computer. Lastly, they're throwing classical problems at a quantum processor and wondering why it's not magically faster. I would like to see how an optimized quantum algorithm performs.

owenversteeg• 12 years ago

Original article: http://www.livescience.com/46414-first-quantum-computer-no-f...

MrBuddyCasino• 12 years ago

Previously: https://news.ycombinator.com/item?id=7917749

michaelochurch• 12 years ago

Weird title. If someone built a quantum computer that performed remotely as well as a 2014-era classical machine, that would be huge news. Quantum computers are really hard to build.

yellowapple2• 12 years ago

What's the definition of "ordinary PC" being used? Not finding an answer to that in the article, which makes it difficult to get an idea of what the researchers are comparing this thing to.

Bakkot• 12 years ago

One which doesn't exploit quantum-mechanical principles of computation.

acjohnson55• 12 years ago

Wouldn't that be a classical computer? PC is a weird term to use here.

yellowapple2• 12 years ago

So is that a brand-new Dell XPS machine? Is that my aging-not-too-well Compaq Presario 1210 with a Pentium? Is that a Macbook Air? Is that one of Oracle's $200,000 M-series servers? Is that some beige-box piece of junk that the researchers found in their neighbor's backyard? The category "personal computers which do not exploit quantum-mechanical principles of computation" is so mind-bogglingly vast that such a description alone is hardly useful.

bwy• 12 years ago

A few other comments have pointed it out already, but how does this comparison actually mean anything? For example, multiplication with the first hole-punch computers was no faster than multiplying by hand and paper (probably slower), but look where we are now!

Performance of quantum computer no better than ordinary PC, say analysts

Related Stories