Quantum Computers by next Tuesday

[His Divine Shadow]

Quantum computer to debut next week
By Peter Judge, Techworld

Twenty years before most scientists expected it, a commercial company has announceda quantum computer that promises to massively speed up searches and optimisation calculations.

D-Wave of British Columbia has promised to demonstrate a quantum computer next Tuesday, that can carry out 64,000 calculations simultaneously (in parallel "universes"), thanks to a new technique which rethinks the already-uncanny world of quantum computing. But the academic world is taking a wait-and-see approach.
D-Wave is the world's only "commercial" quantum computing company, backed by more than $20 million of venture capital (there are more commercial ventures in the related field of quantum cryptography). Its stated aim is to eventually produce commercially available quantum computers that can be used online or shipped to computer rooms, where they will solve intractable and expensive problems such as financial optimisation.
It has been predicted that quantum computing will make current computer security obsolete, cracking any current cryptography scheme by providing an unlimited amount of simultaneous processing resources. Multiple quantum states exist at the same time, so every quantum bit or "qubit" in such a machine is simultaneously 0 and 1. D-Wave's prototype has only 16 qubits, but systems with hundreds of qubits would be able to process more inputs than there are atoms in the universe.
Scientists in the world's many quantum science departments are looking anxiously at whether the demonstration - linked to a computer museum in Mountain View California, will vindicate their work or cast doubt upon it.
"This is somewhat like claims of cold fusion," said Professor Andrew Steane of Oxford University's Centre for Quantum Computing. "I doubt that this computing method is substantially easier to achieve than any other."
Others are more enthusiastic: "I'll be a bit of a sceptic till I see what they have done," said Professor Seth Lloyd of MIT. "But I'm happy these guys are doing it." Lloyd is one of the scientists who helped develop the "adiabatic" model of quantum computing which D-Wave's system exploits - a method which D-Wave believes will sidestep the problems which have restricted progress in quantum computing so far.

http://www.tech.....agtype=all

[His Divine Shadow]

Well whops. I was going to add that I think its a fake, but hell, we can hope!

[Gustav32Vasa]

Will i get the computer before I buy it?

[The Grim Squeaker]

I'll believe it when I see it and an article about it in Science and/or Time magazine after checking it via a neutral skeptic source.
In other words when pigs step through walls

[Netko]

A somewhat more useful and technical article from Ars:

A detour into the strange world of quantum computing

2/12/2007 12:16:26 PM, by John Timmer

On Tuesday, a startup company called D-Wave is going to publicly demonstrate a quantum computing system that they call Orion. Is this the dawn of a new era of computing or simply an exercise in hype? It's hard to say prior to the announcement, but my guess is a little of both. To fully appreciate why, we'll have to take a detour into the strange world of quantum computing.
Speeding calculations through superposition

In what passes for normal in the quantum world, a properly prepared quantum object kept in isolation from the environment will actually exist in all possible states (be they energy, spin, etc.) at once—it exists as a superposition of states. Depending on the inputs to the system, a single state will be the most probable one occupied; repeated measurements of the same, freshly-prepared system can identify this most probable state. If you think of that object as a bit, it's easy to see how calculations can be done: set it in a ground state, perform a calculation by providing an input that influences its state, and then read it. Repeat the process a few times to get the probabilities sorted, and you've done a 1-bit calculation with a quantum bit, or qubit.

Needless to say, such a computer would badly underperform whatever is currently running your cell phone. But in the quantum world, two objects can be entangled; the state of one is linked to the state of another. Thanks to the combination of superposition and entanglement, this two-particle system will simultaneously exist in 22 states, but with only one probable state, influenced by experimental inputs. As more entangled particles are added, the system can simultaneously explore 2n possible answers, where n is the number of qubits.

This ability to explore a large number of states simultaneously may have applications for a subset of a class of math problems called NP-complete, and D-Wave specifically claims that Orion solves these problems faster than a traditional computer. Those of you wanting to learn more about NP-complete problems can read more at Nobel Intent, but their most important feature is that there are no algorithms that work consistently to find a solution to an NP-complete problem. This leaves brute-force testing of possible answers as the only generalized approach. The difficulty of solving an NP-complete problem is the foundation for the cryptography we all rely on to ensure the security of our transactions on the internet. But a 128-bit quantum computer could, via superposition, potentially test all keys at once, providing a rapid answer to an otherwise unapproachable problem.
Bringing theory to the real world

So why is AMD talking about making one of their cores a GPU and not a quantum coprocessor? For one, allowing any of the quantum objects to interact with their environment will immediately destroy the entanglement, eliminating the system's superposition of states. Avoiding this takes a room full of equipment, much of it dedicated to chilling qubits to temperatures near absolute zero. For another, it's not always clear how to provide the system an input so that a single superposed state becomes probable. A few algorithms have been described (including a quantum database search), but currently only a limited number of problems can be explored by a quantum computer.

Because of these limitations, quantum computers have rarely escaped the world of academic research labs. But D-Wave has attracted some academic talent, as well as both private and government funding. That combination of minds and money has gotten them to a point where they're ready to demonstrate a 16-bit quantum computer on Tuesday at the Computer History Museum. A repeat of the demo will occur on their home turf on Thursday; Ars will have a reporter in attendance.

The computer itself won't be traveling to the demo; it will be accessed remotely. It operates at 5 milliKelvin, just a hair above absolute zero, which requires some specialized equipment. The qubits are laid out on a four-by-four grid, and each is entangled with up to eight of its neighbors. It is designed specifically to solve NP-complete problems by reading the free energy state of the qubits, putting it in the class of computers called "adiabatic." There is some doubt within the quantum computing community regarding the utility of adiabatic systems compared to some of the alternatives being explored; a detailed examination of these issues appears at Nobel Intent.

It appears that there may be a number of additional reasons to question whether Orion will perform as promised. These go well beyond the Ars staff's ability to analyze the arguments; it's worth noting, however, that our physics expert (Chris Lee) was unable to find a single paper on this sort of system published by anyone at D-Wave. Although D-Wave's head scientist supplies links to some relevant publications coming out of the academic community, not all of these appear to have survived peer review.
Is D-Wave's demo all about their market position?

D-Wave's demos will show Orion crunching through a database search and handling a seating plan with a large number of constraints, such as you might find at a wedding reception. It's probably safe to say that, despite any of the doubts about the company, Orion will almost certainly show a major speed advance when applied to these two problems. But, given the limitations and questions, and the fact that D-Wave doesn't appear to currently be hurting for money, why are they bothering?

I think Jon Stokes' recent analysis of the extension of x86 into new markets is probably informative here. Even when an x86 processor isn't necessarily the "best" solution (in terms of price or power requirements), the extensive set of software tools, support chips, and human expertise may make up for any deficiencies. I think D-Wave's announcement may be their attempt to try to become the x86 of the quantum world.

As best as anyone is aware (and as evidenced by the lack of a competing set of press releases), nobody else is close to putting together a usable quantum computer. The adiabatic system they're using may not wind up being the "best" in terms of its practicality or ability to solve a wide range of problems (the academic world is primarily working with other systems) but it's apparently ready for some sort of usage now. And that may be all D-Wave needs. If they can convince others to work with their system—code additional algorithms, provide high-level interfaces to Orion, work around its limitations—their approach may not have to be the "best" in any sense.

In order to induce developers to begin building up an Orion-specific toolkit, D-Wave plans to provide free access to Orion to anyone interested in porting software to it over the next few months. If this lure works out, it can ensure two things. First, that D-Wave will have customers interested in paying for access as more sophisticated successors to Orion come online. Second, that by the time other quantum system comes online, it will face a competitor with a variety of advantages in terms of tools and expertise.

In other words, while interesting and with very useful applications, this won't be replacing your computer any time soon, if ever.

[Drooling Iguana]

Unfortunately, now that we know when it's coming no one can find the damn thing.

[drachefly]

h-bar/(1 day) is a rather small distance. I don't think we'll have any trouble.

[drachefly]

DOH... that's what I get for replying without thinking clearly. h-bar over a time range would be an energy.

[White Haven]

Perhaps for sufficiently large values of Tuesday...

[Admiral Valdemar]

If true, it'd be hilarious to see this tiny company make a computer the size of a desktop unit that beats Intel's new parallel processing chips in the 45 nm range like a Deep Blue versus a pocket calculator.

Silicon's days are numbered anyway, it's just a question of when, not if now.

[Molyneux]

It's likely a hoax, or at least vastly unimpressive...but one can hope.

[CaptainChewbacca]

I am eager to see a demonstration which will simultaneously succeed and fail.

[Netko]

If true, it'd be hilarious to see this tiny company make a computer the size of a desktop unit that beats Intel's new parallel processing chips in the 45 nm range like a Deep Blue versus a pocket calculator.

Silicon's days are numbered anyway, it's just a question of when, not if now.

Erm... No. Read the article I quoted. While an very interesting development, with useful applications, this is in no way an end to non-quantum computers, nor is it necessarily going to be in the future.

[SirNitram]

Quantum computing will eventually replace silicon, because let's face it, machine's are constantly sucking up more processing power. I'm just wondering if their design takes advantage of some of the more esoteric discoveries made lately(Like it working better when.. Turned off.).

[General Zod]

If true, it'd be hilarious to see this tiny company make a computer the size of a desktop unit that beats Intel's new parallel processing chips in the 45 nm range like a Deep Blue versus a pocket calculator.

Silicon's days are numbered anyway, it's just a question of when, not if now.

Erm... No. Read the article I quoted. While an very interesting development, with useful applications, this is in no way an end to non-quantum computers, nor is it necessarily going to be in the future.

Wasn't the same thing said about vacuum tube based computing, the floppy disk and the 8-track?

[Spin Echo]

It has been predicted that quantum computing will make current computer security obsolete, cracking any current cryptography scheme by providing an unlimited amount of simultaneous processing resources. Multiple quantum states exist at the same time, so every quantum bit or "qubit" in such a machine is simultaneously 0 and 1. D-Wave's prototype has only 16 qubits, but systems with hundreds of qubits would be able to process more inputs than there are atoms in the universe.

I don't necessarily think it's a hoax, but I don't see how revolunary it is either. The prof I worked for as an undergrad did quantum computing work, and they had managed to get 7 qubits working together around 4 years ago. One of my officemates expected by the time she finished her thesis (around now), they'd have 15.

As far as I can tell from the article, they don't actually have a way to get the hundreds of qubits working together yet.

[Admiral Valdemar]

Erm... No. Read the article I quoted. While an very interesting development, with useful applications, this is in no way an end to non-quantum computers, nor is it necessarily going to be in the future.

1. Way to miss the joke.

2. Silicon's days are numbered, dummkopf. Or do you think we can bend the laws of physics indefinitely by making finer and finer transistors despite the already mounting task in getting to 45 nm gaps? All you can do at that point is start adding more cores and chips, which is essentially the limit of silicon. The industry needs new routes now, so they can be ready for when the real stagnation comes. These potential future markets may or may not include quantum computing elements, though photonic is looking the most likely right now.

[phongn]

Quantum computing is not going to replace silicon. Quantum computing isn't even useful for the majority of everyday tasks people use computers for. Furthermore, quantum computers are highly unlikely to be able to totally defeat modern encryption schemes; they'll essentually reduce the complexity to its square root (IOW, halving the keysize.)

[Admiral Valdemar]

Quantum computing is not going to replace silicon. Quantum computing isn't even useful for the majority of everyday tasks people use computers for.

I'm glad you have the gift of foresight, because I've heard similar things from world leaders on technology and they couldn't be more wrong. Additionally, what timeline are you working on? I didn't specify, and as laughable as a home computer was in 1945 is about as laughable today as how anyone could hope to have a quantum computer of some type within their home. Besides, I already mentioned one other venue of research yielding an end to silicon exclusive chips, namely photonic inclusion and there is no reason quantum based calculations can't be used in everyday business because, let's face it, anyone expecting to run Supreme Commander on their home computer with a "quantum GPU" is in for a shock, but the military, big business and the like will benefit from the sheer number crunching ability to no end. No one said silicon would disappear, but it sure as hell won't be top dog for much longer in the coming decades.

Furthermore, quantum computers are highly unlikely to be able to totally defeat modern encryption schemes; they'll essentually reduce the complexity to its square root (IOW, halving the keysize.)

Given modern encryption schemes includes quantum encryption anyway, it doesn't matter. That which cannot be brute forced at all doesn't factor into whether a computer can throw more processing power at the job or not.

[phongn]

I'm glad you have the gift of foresight, because I've heard similar things from world leaders on technology and they couldn't be more wrong. Additionally, what timeline are you working on? I didn't specify, and as laughable as a home computer was in 1945 is about as laughable today as how anyone could hope to have a quantum computer of some type within their home.

The set of problems that quantum computers solve well is a relatively restricted set compared to their classical brethren. They are simply not general-purpose systems. They will have their use in various niches but even if the very real issues with construction are somehow overcome, I don't see them ever replacing the classical computer for most people.

Besides, I already mentioned one other venue of research yielding an end to silicon exclusive chips, namely photonic inclusion. No one said silicon would disappear, but it sure as hell won't be top dog for much longer in the coming decades.

I know about the various alternates to silicon computing.

Given modern encryption schemes includes quantum encryption anyway, it doesn't matter. That which cannot be brute forced at all doesn't factor into whether a computer can throw more processing power at the job or not.

Well, quantum encryption isn't that useful, anyways IMHO. It's just a way of keeping the key secure and there are plenty of well-proven ways of doing so which don't require a dedicated fibre line.

[Admiral Valdemar]

The set of problems that quantum computers solve well is a relatively restricted set compared to their classical brethren. They are simply not general-purpose systems. They will have their use in various niches but even if the very real issues with construction are somehow overcome, I don't see them ever replacing the classical computer for most people.

I'm aware of that, which is why I laugh at people wishing for a QC to play their games better or some such. Now, in the future it may be that you'll have workstations that are much more compact, cheaper and efficient and can be used for media, such as processing images or working on hard physical models such as protein folding. Whatever happens, such machines will find their way into the public eventually, just likely not in the readily visible "wow" factor way (thank you, Microsoft) that they expect a new videocard to give,

I know about the various alternates to silicon computing.

Yeah right. Not like you're a technophile or anything. You Luddite, you.

Well, quantum encryption isn't that useful, anyways IMHO. It's just a way of keeping the key secure and there are plenty of well-proven ways of doing so which don't require a dedicated fibre line.

That is good enough for the current clients, namely, large multinational banking corporations. I have a step-uncle with a Ph.D in physics who works in the City of London, so if I ever get to meet the bastard (and steal his cash), I'm going to ask what his views are and whether the bank he does accounting for has contemplated the system. So far it's pretty tiny in following, but we know it works and anything that can make such transactions more secure in a standard way is good (obviously an inside man will ruin your day, but one hopes when you write cheques for billions to a nation you don't have a guy lurking to steal the change).

And it would seem our friends at The Reg have got the scoop. I await the Ars write-up too.

[phongn]

I'm aware of that, which is why I laugh at people wishing for a QC to play their games better or some such. Now, in the future it may be that you'll have workstations that are much more compact, cheaper and efficient and can be used for media, such as processing images or working on hard physical models such as protein folding. Whatever happens, such machines will find their way into the public eventually, just likely not in the readily visible "wow" factor way (thank you, Microsoft) that they expect a new videocard to give,

Eh, I don't see it. Quantum computers will remain in their niche doing their science or maths workloads while the rest of the computing world stays classical (in whatever form).

That is good enough for the current clients, namely, large multinational banking corporations. I have a step-uncle with a Ph.D in physics who works in the City of London, so if I ever get to meet the bastard (and steal his cash), I'm going to ask what his views are and whether the bank he does accounting for has contemplated the system. So far it's pretty tiny in following, but we know it works and anything that can make such transactions more secure in a standard way is good (obviously an inside man will ruin your day, but one hopes when you write cheques for billions to a nation you don't have a guy lurking to steal the change).

Yeah, but there's also the cost-benefit analysis. The added security in the physical layer is frankly not that important.

[His Divine Shadow]

A somewhat more useful and technical article from Ars:

*snip*

In other words, while interesting and with very useful applications, this won't be replacing your computer any time soon, if ever.

Hmm, yes thats kinda what I figured.

[SyntaxVorlon]

Quantum computers will certainly widen the set of problems that a computer can do, and significantly shorten the time it takes to do numerous types of calculations, but they are currently by no means small these days and they're mostly hacked together by like an advanced quantum physicist's version of a bong made out of a bear can and duct tape.

[drachefly]

I think they might become more widespread, and here's why:

We've been focusing on QC's ability to tackle NP-complete problems. I may have this term wrong, but it's the set of problems where answers are easy to check (i.e. polynomial time), but hard to find (i.e. exponential time). This is as opposed to NP problems, which are hard no matter how you cut it.

But QC could be used to great effect for many less impressive tasks:
The ubiquitous array sort can be done in O(N) time instead of O(N logN), for example.
And of course the ability to solve constraints like the problem they used as their demo would be able to greatly speed the handling of common tasks like, oh, figuring out which widget is responsible for drawing any given pixel on a screen, or tracking back to figure out where to send the mouse click.

Of course, to do any of that we'd need to get one not mounted in a cryo-tank.

;)