Toward On-Chip Quantum Computing

Darum writes "Researchers are working to create devices built on the rules of quantum mechanics. These would allow quantum computers which can do certain problems such as prime number factorization for decryption and simulation of complex systems (such as protein folding) in a tiny fraction of the time required on classical computers. Two papers appearing in this week's Nature raise the possibility of developing such quantum devices by manipulating light signals by semiconductor quantum dots. One of the approaches bases on photonic crystals, which seem pretty ideal for on-chip integration of a full set of computation components. One of the study's authors put up a good background story of this work on CVitae. The author discusses the potential simplicity and microchip scalability of these two quantum-dot 'light switch' systems. This could be good news for quantum information processing and ultra-secure long-distance communication applications. It could also allow all-optical signal processing, which has long been a holy grail for optical communications and could allow extremely fast and low-power optical interconnects."

Re:

[looseSpark]

This post is both in the state of being the first and not the first post until I hit the submit button and decide the outcome by refreshing the page. That's quantum computing.

I just saw your first post. Now it's the LAST post!

Run!

[mstahl]

Quantum trolling! It has begun....

Re:

[account_deleted]

Comment removed based on user account deletion

Questions of SW developer

[should_be_linear]

I don't quite understand how Quantum thing will change SW development as we know today, so maybe someone can bring insight beyond factorization, which is not all that interesting (since everybody is informed about factorization speedup many times). For example, does it mean traditional programming languages will be modified, handling of variables will be different, maybe new operators? Or, only some external library will be introduced for set of quantum operations? Will it be able to solve Traveling Salesma

Re:

[QuickFox]

On the contrary, Slashdot users are from everywhere around the globe. Not everyone has English as their first language. And yet they're definitely welcome.

And when I say welcome, obviously I'm thinking of everyone except you. Asshole.

Re:Questions of SW developer

[exp(pi*sqrt(163))]

I would never hold it against someone that they know nothing about quantum computers or nothing about games. But you have chosen to hold forth on these subjects in response to someone's questions without taking any note of the fact that you haven't the faintest clue what you are talking about.

> A variable will still be a variable

This is completely incorrect. The concept of a variable radically changes in a quantum computer because you are allowed superposed states.

> I'm not aware of quantum mechanics introducing any new operators

What in heaven's name are you imagining? Of course quantum mechanics introduces new operators. It completely turns classical mechanics on its head and introduces concepts that make no sense in a classical framework. Here's an example [americanscientist.org] of a specifically quantum operator.

TSP is NP-hard, and quantum computers don't, as far as we know, make NP-hard problems solvable in polynomial time. Grover's algorithm [wikipedia.org], however, does allow you to search a database of N items in time sqrt(N) so it could provide many speedups to familiar algorithms.

> Chess, aside from being Zero Sum

Are you *trying* to look like an ignoramus? Zero-sumness has absolutely nothing to do with chess. Zero-sumness is about the payoff you get from game of incomplete information. It has nothing to do with the strategy you should use in a game of complete information like chess. I guess you just want to sound smart by throwing around technical terms you don't grasp.

> seriously doubt there is one unbeatable strategy, since a player cannot control the first piece the other player moves.

Woah! Where are you getting this stuff from? Are you just making stuff up as you write it? It's incredible. Whether or not a game has a winning strategy has nothing to do with whether you can control the other player's first move.

As I say, there's nothing wrong with not knowing stuff. But spouting garbage in response to someone's genuinely inquiring questions is nothing short of obnoxious and just serves to lower the signal to noise ratio on Slashdot.

Re:

[exp(pi*sqrt(163))]

My penis is microscopic and doesn't even work most days, not that I've ever had an opportunity to use it properly. Does me saying that make you feel better? Does having an average sized penis make you feel better about showing off your ignorance?

Re:

[rucs_hack]

I managed to rile you about something so insignificant as a post on a web page, so yes, silly person....

Re:

[Poromenos1]

Wikipedia disagrees [wikipedia.org] with your definition of zero-sum, and states that chess is, in fact, a zero-sum game.

Re:

[exp(pi*sqrt(163))]

That Wikipedia article is very misleading, chess is completely the wrong type of game and that's the worst example to start with. There is no numerical payoff assigned to wining, losing or drawing a game of chess. Pick up any book on game theory and you'll find that any discussion of zero-sumness is entirely separate from discussion of combinatorial games like chess. Zero-sumness refers to games in which there is some kind of payoff (eg. at the end of each round) and the total paid out is zero. You can exte

The concept of a variable

[mangu]

> A variable will still be a variable

This is completely incorrect. The concept of a variable radically changes in a quantum computer because you are allowed superposed states.

> I'm not aware of quantum mechanics introducing any new operators

What in heaven's name are you imagining? Of course quantum mechanics introduces new operators. It completely turns classical mechanics on its head and introduces concepts that make no sense in a classical framework.

Let me guess, would those new operators be somethi

Re:

[exp(pi*sqrt(163))]

> Let me guess, would those new operators be something like this?

No. Quantum computers are not just parallel computers and the new operations that quantum computing introduces are not simply parallel operations on arrays. (Is that what you were getting at?) If that were the case, Grover's algorithm would run in time O(1), not O(sqrt(N)). None of the nice quantum algorithms out there (eg. Grover's, Shor's) work simply because they do stuff in parallel (though doing stuff in parallel is an essential ingr

Re:

[blueg3]

Neither the superposed q-bit nor the Hadamard operator that is the fundamental quantum operator have classical computational equivalents.

Re:

[maxwell demon]

Saying that the Hadamard operator is the fundamental quantum operator is wrong. Indeed, there's not much you can do with only a Hadamard operator (applying it twice recovers your original state). And even if you combine Hadamard with the quantum versions of classical operators (like CNOT), you still don't get all possible quantum operations, not even approximately. You'll have to add another operation (like a phase shifter gate).

Re:

[blueg3]

That's correct, you still need other gates to perform interesting functions. Yet the Hadamard is still the most essential and fundamental quantum operator, and one that has no classical equivalent.

Re:

[maxwell demon]

Yet the Hadamard is still the most essential and fundamental quantum operator, and one that has no classical equivalent.

How exactly do you propose to define the "fundamentalness" of a quantum operator? Basically an universal quantum computer needs to be able to do (at least approximately) any one-qubit unitary transformation (which can be done with Hadamard + phase shift, but equally well e.g. with square root of NOT and phase shift) and some non-trivial two-qubit operation (like CNOT). I don't see why the

Re:

[blueg3]

Quantum computing is very different. The details are of course very different (such as the operators, and the need for bit-level error checking), but more important to the software developer, the algorithms are fundamentally different. You approach a problem with an entirely foreign set of tools. With quantum computers, it's not a matter of the quantum computer just being "better" -- it has access to a way of doing things that is more powerful (in the mathematical sense) than classical computing.

I don't rem

Re:

[maxwell demon]

With quantum computers, it's not a matter of the quantum computer just being "better" -- it has access to a way of doing things that is more powerful (in the mathematical sense) than classical computing.

That depends on your definition of "powerful". If "more powerful" means "can solve more problems", then no, the quantum computer cannot solve any problem which a classical computer cannot solve. If "more powerful" means "can solve problems in (asymptotically) less time", then yes, quantum computers can be mo

Re:

[blueg3]

Yeah, you're right re: the speed of Grover's algorithm.

Clearly in the above I was using the definition of powerful that makes sense, not the other typical definition ("capable of solving a larger body of problems").

Re:

[kryten_nl]

Is it able to solve chess game?.

"Solving" of the chess game like American Checkers was solved is not really a problem of computational speed (I mean it could be done by our civilization if we decided it would be worth it), but more of storage. There are more possible chess positions then there are atoms on Earth. http://en.wikipedia.org/wiki/Shannon_number [wikipedia.org] http://pages.prodigy.net/jhonig/bignum/qaearth.html [prodigy.net]

Re:

[Jeff DeMaagd]

Assuming it works, I'd think it would be more of a functional unit within a chip or system that does math for you. Before that, it might even be an add-in device, conventional software gives the device commands and it will return the results when it's done. I'm not sure whether there are consumer uses for the technology. But I think it would probably still be largely controlled with conventional software.

Looks like /. stories have quantum states, too.

[Enleth]

A story can be in a duped or non-duped state at the same time, as we can't be sure if a yesterday's story is there again under a new title until we open the Slashdot main page, making the state collapse into either extreme with a roughly equal distribution.

This time, the cat is de... Er, the story is duped. Well, OK, not really an exact dupe, but looks like it references the same information, just from different sources...

Re:

[FST]

You know, most people just say "DUPE!!!!" and get over with it.

Re:

[maxwell demon]

But quantum people say "dupe" and "not dupe" at the same time!

free first submission of other article

[Jimminey Cricket]

A free version of the other article (using microdisks instead of photonic crystals) is available on the arXiv:
http://arxiv.org/abs/0707.3311 [arxiv.org]
Reading the two papers careful, it turns out the photonic crystal paper is only at the "onset" of strong coupling (the decay rate is still about 2x faster than the coherent light-matter coupling rate) while the microdisk paper is actually strongly coupled (the coherent coupling rate is faster than any decay or dephasing).

Re:

[ibillin]

Actually having strong coupling is not necessary to see these effects http://arxiv.org/abs/quant-ph/0610172 [arxiv.org] . It may be necessary to reduce losses. Actually both papers operate in strong coupling, because both observe spectral splitting, and the given regime in the second paper also puts it into strong coupling - there is just not a full oscillation of the photon between the cavity and quantum dot.

Now all we need is...

[maciarc]

Now all we need is On-Salsa Graphics Processing and we're there!

Quantum Crystal Computers?

[MikeUW]

Superman already has this in his fortress of solitude...

It's Obvious......I should have waited

[hyades1]

You might notice that I posted basically the same feline reference twice a day ago under "Light-based Quantum Circuit Does Basic Maths".

Please apply one here, along with any obscure reference to quantum physics and/or time you may think appropriate.

Re:

[kryten_nl]

Your previous comments might have been in state of Funny and !Funny at the same time, however when I looked they weren't.

Re:

[hyades1]

!Thanks.

decryption

[cadeon]

One of the topics in the summary, at least, is being able to do much more secure encryption.

As I understand it, encryption gets it's power from the fact that it takes a whole lot of computing power to guess the key, but if you have the key, everything goes well.

If everyone has these much more powerful computers, aren't we back to where we started? I'd think we'd end up at about the security level we are now, just with more overhead. Can quantum computing provide us with a new encryption method, which doesn'

Re:

[ibillin]

Quantum encryption works a little differently - basically the whole point is that you can create one time keys with almost 100% safety, because a quantum system is sensitive to measurement. With classical keys someone can copy it and then attempt to break it. A quantum system, on the other hand, cannot be copied, because to make a copy you have to first measure what it is. Typically the way the algorithm works is that I send say photons that are linearly polarized. I can randomly send them vertically or hor

Exponentially more difficult to build?

[BlueParrot]

I'll admit that my knowledge of quantum computers is , limited, but surely if you entangle a large number of qubits then an interaction which destroys the state of a single quibit will in fact destroy the state of ALL the qubits. Thus even if you can get reduce the probability that any particular qubit will be distrubed during your calculation, once you try to scale this up to a gigabyte or so, you have a problem anyway since the probability that NO qubit was disturbed increases exponentially with the numbe

Re:

[Soleen]

<quote> Being able to factor large integers quickly is not much use if you have to repeat the calculation many times to make sure it was correct.</quote>

Wrong, it is usally much cheaper to check that answer is correct than to find the answer.
For example once quantum computer has factored a large number, it is simple to check the calculation even with regular PC by multiplying those numbers and compare with the orgincal large number.

Re:

[SeekerDarksteel]

Not only has the problem of reliability not been solved, it is the primary barrier to effective quantum calculations. We simply cannot keep qubits in an entangled superpositioned state for long enough to perform calculations. And you are correct, every qubit you add decreases the time to decoherence, so as you increase performance you decrease reliability. This is however less a problem of computational time (i.e. requiring multiple runs of the algorithm) and more a problem of being able to get any answe

Energy required exponential?

[Myria]

Is the energy required to build a quantum computer and keep it coherent exponential in the number of qubits? It would make quantum computing mostly worthless if this were true. It would also be yet another case where nature conspires against those who try to use quantum mechanics to violate the normal laws.

If such things are possible, I hope when the qubit count reaches ~2048 that someone factors the Xbox public key.

Reversible Logic Synthesis

[kEnder242]

http://www.amazon.com/Reversible-Logic-Synthesis-Anas-Al-Rabadi/dp/3540009353 [amazon.com] (November 5, 2003)

All about what you can and can't do with quantum computing (and how to implement it)
If you don't want to wade through everything, skip to Chapter 11.

http://books.google.com/books?id=0e8LbxngITsC&pg=PA229&dq=reversible+logic+synthesis&sig=l1bT9QLXAuEkhqLlmnU8gopwndY [google.com]

All optical signal processing

[Team Zissou]

This animation deals with all optical signal processing on a big picture level: http://cudos.org.au/cudos/education/Animation.php [cudos.org.au]