Under the Hood of Quantum Computing

nanotrends writes "Gordie Rose, the CTO of Dwave Systems, the venture funded company that plans to offer paid use of a superconducting quantum computer starting in 2007, reveals secrets of his quantum computer construction. It is based on nobium superconducting 'circuits of atoms' and is not RSFQ. (Rapid Single Flux quantum)."

Comments

Advantages?

[Zouden]

I read the article, but it didn't make it very clear - what will be the advantages of paid use of their quantum computer? Unless it's going to be faster than other supercomputers, I can't see the point. Is there some other advantage I'm not aware of?

I'd be very suprised if their quantum computer will be faster than conventional computers by next year. 20 years away, maybe.

Re:Advantages?

[QuantumG]

I don't think anyone can assess the capabilities of his systems from that article. I also don't think that was unintentional.

Re:Advantages?

[Mathinker]

Uhm, from the article, nobody can even assess whether it really is a quantum computer.

Re:Advantages?

[Kjella]

I read the article, but it didn't make it very clear - what will be the advantages of paid use of their quantum computer? Unless it's going to be faster than other supercomputers, I can't see the point.

Well, it's a quantum computer. Given the problem it might be like trying to make your CPU compete against a GeForce or ATI. If you try to do it all with CPU emulation, there's not much doubt who'll win. That said, I got the impression that current quantum computers have a so limited number of qbits (the computing power pretty much grows to 2^n with n bits), that it's faster and cheaper to just cycle through all 2^n possibilties one at a time. Currently the largest I've seen is a 12 qbit computer. Now 2^12 = 4096 states at once is a nice curiosity but nothing that makes my encryption keys worry. Basicly it's man vs Deep Blue at computer again - the quantum computer is great at testing many solutions at once but the sheer computing power of traditional computers takes home the victory. Now, if they can get hundreds of qbits together things will change massively. But the difficulty in keeping all those in a cohesive quantum state also raise drastically, so I think we're far off from a usable quantum computer.

Re:Advantages?

[RKBA]

"Now, if they can get hundreds of qbits together things will change massively."

I think the point of the article is that D-Wave Corp claims to be able to create qbits from "large" objects (ie; large enough to be fabricated using standard IC fabrication techniques), but with niobium rather than silicon. This enables them to create a quantum computer without all the hassle of having to manipulate individual atoms as the present research lab quantum computers do. From the article:

Superconductors are the only type of material that we know of where big lithographically defined devices (like really big. Like centimeter on a side big.) can be built that behave just like they were atomic-sized.

Since supercooling is required, it's highly unlikely that you or I will be able to afford one of these things any time soon (assuming it's not all marketing hype in the first place), but you can be assured the NSA and other government "intelligence" agencies will be able to afford as many as they want because of all the tribute they demand from us on pain of imprisonment, in the form of exorbitant taxation.

Re:Advantages?

[lgw]

As far as I know, only RSA-style cryptograophy is affected by quantum computing. There are other ways to do public key encryption, such as elliptical curve cryptography that should be unaffected, as they depend on a different class of problem being hard, and of course quantum computing won't help with symmetric key crypto at all.

The NSA has been advising the security community against using RSA-style encryption for some time now - it's not like they're trying to keep the weakness a secret for some nefarious reason.

Re:Advantages?

[mickwd]

Or, following the principles of Heisenburger's Uncertain Cat, we can determine if it really is a live quantum computer, but we can't know where it is.

Re:Advantages?

[ZombieWomble]

I do believe you're mistaken. Quantum bits are exactly like regular bits in their possible observable states - that is, they are either "on" or "off" when observed. The interesting part of quantum computing comes from the fact that, when they're not being observed, they exist in a superposition of both "on" and "off" states. Now, if you put 8 of these bits together, you have a 'qbyte' which, while when it's observed it can only represent the same range as a regular byte, can be used in calculations representing every single possible permutation of the data at once - i.e. every number from 0 through 255. Each bit you add doubles the number of states you can simultaneously test using this superposition property - this is what the GP meant when he said that quantum computing scales as 2^n.

Re:Advantages?

[PatriceVignon]

One of the most interesting categories contains problems that are called NP-complete. These all have the feature that in order to solve the problem all possible solutions must be tried, and the number of possible solutions grows exponentially with the problem size. An example is the Travelling Salesman Problem, although there are literally thousands of them. This category is a particularly interesting target from a commercial perspective because most real-life business problems are in it. ... Quantum computers can be used to get approximate solutions to large NP-complete optimization problems much more quickly than the best known methods running on any supercomputer.

Sorry, but that is simply not true. If you have a classical NP complete problem (e.g. Travelling Salesman), you can solve it by trying out exponentially many steps, 2^n, and most people believe that you cannot find faster (classical) algorithms. With a quantum you can improve this to 2^(n/2) by the so-called Grover search algorithm. This is not nearly enough to make these problems tractable in practice. And to make things worse, this "speed-up" will most likely be eaten up by the necessary error correction.
Lance Fortnow posted a very nice summary of this on his blog:

But I'm not a physicist or an engineer and suppose we can overcome these obstacles and actually build a working machine. Then I can imagine the following conversation in 2025:
Quantum People: We now have a working quantum computer.
Public: Yes after 30 years and 50 billion dollars in total funding. What can it do?
Q: It can simulate quantum systems.
P: I'm happy for you. What can it do for the rest of us?
Q: It can factor numbers quickly.
P: Yes, I know. We've had to change all of our security protocols because of your machine. Does factoring have any purpose other than to break codes?
Q: Not really. But we can use Grover's algorithm for general search problems.
P: That sounds great! So we can really solve traveling salesperson and scheduling problems much faster than before?
Q: Not exactly. The quadratic speed-up we get from Grover's algorithm isn't enough the offset the slow-down by using a quantum computer combined with error correction. But we can solve Pell's equation, approximate the Jones polynomial and a few other things very quickly.
P: Are those algorithms particularly useful?
Q: No.
P: They why did you build a quantum computer?
Q: Because we could.

"Quantum" computer is misleading

[kestasjk]

What D-Wave has done is begun with the standard approaches to building metal-based processors and modified them in such a way that these processors use quantum mechanics in order to accelerate computation.

Wow, they use quantum mechanics? Every chemical reaction in our universe uses quantum mechanics; they couldn't be more vague if they tried. They're clearly trying to capitalize on the 'quantum computer' buzz.

Re:"Quantum" computer is misleading

[slashdotmsiriv]

From dwave's site: "There are many potential ways to build quantum computers (QCs). Of these, four types have emerged as being most likely to succeed. These are based on (A) assemblies of individual atoms trapped by lasers; (B) optical circuits, for example using photonic crystals; (C) semiconductor-based designs, usually including atomic-scale control of dopant atom distribution or quantum dots; and (D) superconducting electronics. D-Wave focuses exclusively on superconducting electronics. This is because superconductors have the unique property that very large structures can be built out of them that behave according to the rules of quantum mechanics. Because of this, design of superconducting QCs does not require new technology development. This is in contrast to the other three types of QCs, in which information is stored using atoms or individual photons (particles of light), and controlling and manipulating this information requires technologies that do not yet exist. The two superconductors used to build QCs are aluminum and niobium. At room temperature these materials are metals. When they are cooled down close to absolute zero, the electrons in the metals pair to form particles called Cooper pairs. These particles carry charge in the superconductor. Cooper pairs are very different from electrons. One key difference is that Cooper pairs are what physicists call bosons, while electrons are fermions. Bosons are allowed to occupy the same quantum state, while fermions are not. In a superconductor, all the Cooper pairs can (and do) exist in exactly the same state. This means that all of the charge carriers in the superconductor are fundamentally linked. They directly inherit their behavior from the scale of a single Cooper pair. One way to think of this is that a chunk of superconductor amplifies the quantum effects that exist at the level of extremely tiny individual particles up to the scale of the whole chunk, even if the chunk is very large. This amplification of quantum effects is responsible for the well-known properties of superconductors, such as zero resistance to current flow and exclusion of magnetic field. It is also extremely useful for building components of QCs. Superconductors naturally shield themselves from external noise, creating a safe haven for quantum effects. This ability to build large things that behave like small things overcomes many practical problems in building real QCs."

Re:RTFA, WTF?

[kfg]

... What have I missed here?

For starters; a link to the company's website instead of somebody's "See Spot run" blog post:

http://www.dwavesys.com/quantumcomputing.php

KFG

Woo Woo science

[Valacosa]

A functional quantum computer? Really?

I used to be a undergrad lab assistant. I never worked in quantum computing, but our neighbours were some of these guys. I picked up a few things, one of those things being that quantum computing is hard.

Classical computers use the laws of classical physics to operate. Classical physics is deterministic, and that's how we want our classical computers to behave. As the chip and die sizes get smaller and smaller (what are we at now, 65nm?) CPUs are more likely to suffer from quantum effects, but AFAIK there's circutry in there to compensate for that. Error checking.

A quantum computer is just a machine that uses the laws of quantum mechanics rather than the laws of classical mechanics to operate. The advantage is that some algorithms, when implemented on a quantum computer, are 2n instead of n^2. I never really understood this, maybe a better physicist will come along and explain it. Anyway, to build a quantum computer one needs two things:
- (a) You need some Quantum bits (qbits) to store data
- (b) You need to get those bits to interact with each other in some fashion

There are many approaches to building a quantum computer. One guy (Raymond Laflamme) has a bunch of different atoms that are different elements all in the same molecule, those interact with each other but he has only developed the ability to read / write to about 5 different qbits. I read about another guy on Slashdot here who made a giant array of qbits using atoms in a laser trap. That gets you a lot of qbits, but they don't interact at all. There are many approaches.

Anyway, the reason I think Dwave Systems is full of bullshit is that any approach thus far is good at (a) or (b), but not both. Someone who got a powerful quantum computer up and running would most assuredly win a Nobel Prize. Also, why the hell would he need to woo venture capital? I know I'm up in Canada, but I'm sure most governments are throwing scads and scads of research money at Quantum computing. Answer? Venture capitalists are more naive.

If there's anything I learned from here, it's that a lot of Con artists use buzzwords to try and justify their woo-woo science. "Quantum" is one of them.

Smart money on this guy being a fraud.

Re:Woo Woo science

[Iwanowitch]

any approach thus far is good at (a) or (b), but not both.

Ooh, Heisenberg-approaches!

Re:Woo Woo science

[Valacosa]

You're half right. I had forgotten about the quantum properties of transistors.

Though a transister does use Quantum Mechanics to function, it is a discrete unit (a "black box" if you will) with a preidctable outcome. A quantum computer, on the other hand, uses a property of QM known as "superposition of states". A qbit in a quantum computer isn't 0 or 1, but some combination of 0 and 1 at the same time. It's only when the qbit is "observed" (read) that it becomes a 0 or 1.

If we can get these qbits to interact with each other without reading them (or "collapsing the wavefunction", in quantum mechanics lingo) then we can have various superpositions of 0s and 1s interacting with each other within an algoritm. Essentially the algorithm run by the quantum computer is acting in parallel with itself. When we observe the qbits when the algoritm is finished, we see the desired result. I know that sounds like magic, but I've probably explained it poorly. I've explained it better in the past.

Incidentially, someone who is uneducated (not stupid, mind you, just uneducated) may have difficulty distinguishing between the BS in the original article and the more scientifically accepted BS I've spouted. See? That's how these con artists are allowed to succeed!

It just sounds a lot like a RSFQ chip.

[AWeishaupt]

From what has been described on the blog and website, i'm not convinced that what they're working on is much more than simply a superconducting RSFQ - Rapid Single Flux Quantum - chip, which although can concievably run at a breakneck speed compared to todays Silicon CPU's, is not a Quantum Computer in the normal sense. This thing isn't going to run Shor's Algorithm. Also, i'm surprised to notice that there are people here who still consider QCs as science fiction - they're not. Quantum Computing has been practical in the lab since the 90's - and, for example, composite numbers have been factorised in polynomial time. The challenge faced by QCT research groups around the world at present is mainly building the things with a large number of qubits, and still maintaining successful operation. With regards to solid state devices such as the Kane QC model, one of the approaches being investigated involves building multiple small QCs and interconnecting them via conventional microelectronics - perhaps SETs, RSFQs, spintronics or maybe even plain old silicon microelectronics - to create a useful, many-qubit, computer.