NASA backs quantum computing demonstration

NASA has confirmed that it built a quantum computer under contract for Canadian startup company D-Wave Systems Inc. after D-Wave was accused of faking a recent demonstration to businesses and academia. During a demonstration last month, the company revealed that its 16 qubit (more qubits = more processes) quantum processor had been left back at the company's offices, and therefore the test would be show via an internet link. According to D-Wave CEO Ed Martin, "businesses aren't too fascinated about the details of quantum mechanics": unfortunately, a group of industry experts were interested, and they made their skepticism clear. The result is NASA's confirmation that it did in fact construct D-Wave's quantum processor design under contract; a confirmation that most people will be inclined to accept. The underlying technology shown in the demonstration was the use of a normal digital processor in conjunction with a quantum chip, which D-Wave hopes will enable commercial applications of quantum computing. D-Wave is planning to up its design from 16 qubits to 1,024 qubits by the end of 2008, a rather controversial aim for some that think practical quantum computing is still up to a decade away. There's gotta be something about quantum computing that puts scientists on edge: perhaps it's the platform's potential to make all current forms of encryption obsolete ...

[Via Slashdot]

[paralipsis @ Mar 11th 2007 4:03PM]

Still not quite a guarantee that it actually works as advertised, but it does add a little more veracity to the claim.

If it does work, Moore's law is in some trouble...

[Aaron @ Mar 11th 2007 4:10PM]

Won't this destroy the whole binary system?

[Faisal @ Mar 11th 2007 4:28PM]

No, this won't destroy binary. It's interesting how many people think that quantum computing introduces some kind of ternary number system; quantum computing doesn't introduce a third state for logic, or any such thing that's already been possible (base 10 computers were actually tried for a while and eventually discarded).

Basically, the big difference with quantum computing is that you can compute solutions with your system in an indeterminate state, allowing you to explore the entire solution space quickly. This makes previously so-called NP-hard problems much easier (if not trivial), because the known methods for solving them (mostly brute force) become feasible when huge swaths of the problem space can be computed concurrently and practically instantly.

This may be slightly inaccurate, as I'm only a CS major in training...

[surge @ Mar 11th 2007 7:14PM]

u are right sir, the skeptics of d-wave are mainly jealous of what this group outside of the traditional academic circles have been able to accomplish...biz people just don't understand it

[mike10010100 @ Mar 11th 2007 5:59PM]

Regardless, i'm waiting for the day when a quantum computer will run my cpu and gpu. lol

[yoinkers @ Mar 11th 2007 7:36PM]

Yeah you're right. All the academics who pioneered this work and actually built proof of principle devices are just jealous. I mean, industry would have totally done this on their own, they didn't need the help of hundreds of academics who pioneered quantum mechanics.

Oh, and everyone knows that NASA never wastes money on bogus science and technology. Nope, never.

I remain skeptical. The onus is on them to prove without a doubt that they've really built a quantum computer on a chip. It still looks dubious right now.

[Don @ Mar 11th 2007 7:48PM]

Quantum computing will, if it works as advertised, be able to compute ALL permutations of an equation instantly, which is its appeal in applications like cryptography, protein folding analysis, large-number factoring, n-complete operations (like finding the most efficient route between many destinations, for example), physics modeling and the like. If you had a 512-qubit quantum processor, you would be capable of breaking a 512-bit encryption instantly. the number of calculations per cycle that can be achieved goes up exponentially with each new bit to the number of qbits raised by that same # of qbits. A 512 qbit proc could therefore solve for 1.4 e1387 calcs every cycle.

Even a 16 qubit proc is amazing, performing 18,446,744,073,709,551,616 (that's 18.5 quintillion) operations in one go. The new p4 duo would take
***10 million years*** to perform the same calculation in the standard manner.

Crazy. I for one welcome our superpositioned, indeterminate overlords.

[Craig @ Mar 12th 2007 12:33AM]

Yeesh...and we thought women were complex.

[James @ Mar 13th 2007 9:31PM]

Let's debunk some of this.

"Quantum computing will, if it works as advertised, be able to compute ALL permutations of an equation instantly"
Totally false. This is one of the common misconceptions of quantum computing. Certainly you can encode a linear combination of all "permutations" as you say. But your result will be a linear combination of ALL the results. That is you won't be able to distinguish one answer from the next with only one run.

"If you had a 512-qubit quantum processor, you would be capable of breaking a 512-bit encryption instantly."
Close, but you're still way off. You need 512-qbits to encode the number you want to factor but you need some extra memory to do the computation. Also "instantly" is meaningless. You cannot get it in one shot, you'll need to run the algorithm several times since there is just over a 40% chance you'll get each bit correct in your answer.

"the number of calculations per cycle that can be achieved goes up exponentially with each new bit"
Astoundingly absurd. There is only a performace increase when there exists a quantum algorithm more efficent than any known classical algorithms.
A QUANTUM COMPUTER RUNNING A CLASSICAL ALGORITHM HAS NO INCREASE IN PERFORMACE.

"Even a 16 qubit proc is amazing, performing 18,446,744,073,709,551,616 (that's 18.5 quintillion) operations in one go."
Site your source?

[Genome @ Mar 12th 2007 7:56AM]

So if a quantum computer already exists, what are the chances one is already being used to find keys? Considering NASA built one for D-wave, surely they did a "Contact-style" trick on D-wave and built another one for themselves?

[Don @ Mar 12th 2007 3:16PM]

Why build one when you can build TWO at twice the price?

[vainey @ Mar 12th 2007 6:55PM]

I attended the event in Mountain View and documented the interviews they screened that day, by Colin Williams of JPL and Alan Aspuru-Guzik of Harvard. As far as NASA's involvement, the chips were built by JPL, a division of NASA's R&D. JPL is a scientific outfit and has a semi-conductor fab, among other things, that produces many kinds of chips. Saying the computer was made by NASA is sort of like saying Chevron made the gas in my car. True, but irrelevant from the conspiracy angle. It's not as interesting to me to question whether or not quantum functions are indeed occuring, as to observe the trajectory of the company and see what they come up with. If the technology becomes robust and they begin doing the kinds of calculations they say they will be able to do, that will be proof enough at least to business people that a quantum computer is at hand. The scientific community is another story. They have a different goal in the end. But please note, D-Wave was founded by a group of theoretical physicists, not business people. The business end of things is a much more recent development. Also fyi, quantum computers will not be an exclusive entity. If some qubits ever end up in your home PC, they will be running right along side your intels. Qubits are not good at everything. Nor do they provide instant answers. Nor are they the holy grail of processing. Digital is still faster at many things. Skepticism is healthy. Dismissing the announcement out of hand is foolish. Dr. Rose was quite forthcoming and sincere in his presentation. The whole thing was hardly a sales pitch, but simply an assessment as to what may be around the corner. btw, the entire event was videotaped and posted to youtube not a day later, if anyone is interested in seeing the whole thing.

[Markarian @ Mar 12th 2007 10:41PM]

While I don't know much about quantum mechanics, I can say that this computer requires an enormous amount of cooling and is so delicate, it can't even be moved or even witnessed in person. Therefore, I'd say we've still got a long way to go. A usable quantum computer in a year is unrealistic.

And also, since when did a government agency devoted to aerospace research start working under contract for private industry. Isn't it supposed to be the other way around?? That's like BMG hiring the US Marshals to be concert security, or paying the NIH or DHHS to test your new drug.

[don @ Mar 13th 2007 7:47PM]

I believe the idea is for them to rent time on the machine. The average person doesn't have problems that would need qubits to solve, so they are aiming this at people that need serious number-crunching done. And don't confuse "commercially packaged" with "working". Even if there was only one of these things in existence, it would still be an amazing device - assuming they have some sort of software written for it.

Put another way, even if problem execution speed were as slow as one calculation a month (since the equation needs to be programmed in, a process which I have no idea how long it takes them since it's been mostly theory without experiment up till now), it would STILL be faster than even current distributed solutions by a factor of hundreds of thousands. Think of the SETI and Folding @Home units that we could get done!

This tech won't be driving video cards or playing Quake 4. It's really a different solution for a very different set of problems.

[mike @ Mar 16th 2007 10:23PM]

all of this is meaningless if it won't speed up my boot time