Slashdot Mirror
Quantum Computer Possible From Silicon Fab
Cash Mitchell writes: "This article from the EE Times says 'Researchers at the University of Wisconsin in Madison claim to have created the world's first successful simulation of a quantum-computer architecture that uses existing silicon fabrication techniques.... With existing fabrication techniques, the team estimates that a million-quantum-dot computer (1,024 x 1,024 array) could be built today and operated in the megahertz range.'"
...unless of course you try to look at the results.
How many cats will be sacrificed to test a 1024x1024 quantum array I wonder?
What are practical, everyday use? (besides breaking incredibly big and long keys to steal identities) These things operate at room temperature and are small and cheap enough for everyone to have.
A personal weather forecaster, fluid dynamic calculating, realtime, 3d cellphone with a cute ring tone? Or a wash machine that can predict el nino's?
Help me here...
Uh, hate to burst the bubble of your little security apocalypse, but encryption schemes which will baffle quantum computers have been worked out for a while now, well in advance of the hardware's availability. Of course, for all I know it may *take* a quantum computer to implement these schemes (otherwise it seems like we'd just use them now), which would create two security classes of users, one of whom could penetrate the other's security at will. Yikes.
My deviantArt site
Coventional quantum computing is described by a network diagram. This can be translated into a sequence of computational steps, one or two qubit gates acting on selected qubits. The simplest QC architecture would be to run one gate at a time.
Parallel exucution of gates can be arranged (as long as gates act on different qubits) but this is highly dependent on the actual physical system used (ion trap, neutral atom trap, optical lattice, solid state nuclear spin, electron dots, SQUIDs etc).
The key figure of merit is the ratio of gate execution time to the decoherence time. Current estimates of error correction efficiency place the upper bound of this ration at 10^-4 or so (this actually also depends on the ratio of the number logical qubits to physical qubits, sacrificing one for the other). Since quantum dots have very short relaxation times, this places severe constraints on the high speed control electronics. I'll wait for the pre-print or paper before coming to any conclusion on the report. There's still the problem of constructing the damn thing, the purity of the silicon, cooling, EM noise and readout (which isn't mentioned in the article). I'm wary of the heterostructure approach, getting pure silicon to work is hard enough (ask the UNSW guys).
Cheers,
D.
(Not a solid state expert)
There's no problem factoring big primes. I can do it in constant time. Big composites is a problem.
Free Java games for your phone: Tontie, Sokoban
I truly take pride in this discovery... mostly because I attend UW. But I suppose a love of physics helps in that area, too.
Anyways, here's a somewhat technical article regarding the research (PDF).
Oh, and "On Wisconsin!"
IWARS.
People, in general, disappoint me. Politicians even more so.
Quantum computers could render assymetric crypto next-to-useless, and as-such may permenantly set electronic privacy back decades for all but the super-powerful.
Those that claim quantum cryptography will redress this problem don't understand that quantum crypto will likely be even more expensive than secure symmetric cryptography.
In essence, the advent of quantum computers may be the turning point, the point where advances in computer communication are no-longer tools of freedom, but become, once more, tools of the powerful.
...to break RSA. Specifically, I believe that Shor's Algorithm requires 3n qubits, where n is the number of bits of the number you're trying to factor. Multiply by a factor of five to allow some error correction, and you need about 15k qubits to crack 1024-bit RSA.
I work in the field (still an undergrad, but I'm doing some research), and I had the opportunity to meet Michael Nielsen a little while ago when he visited the Perimeter Institute and the University of Waterloo. Nielsen is one of the two authors of the book you mentioned. Out of curiousity, what university do you go to, Misanthropic?
http://www.princeton.edu/~pear/index.html/ www.fourmilab.ch/rpkp/
http:/
Have a long hard look at that first link before you ignorantly dismiss this person's opinion.
There is a lot of research into this - the ability for thought to influence the outcome of random calculations and events. It's been years since I looked into any of this, the most common experiment is a depiction of a random number generator that you can make devitate from a true random distribution over time by willing it to do so.
Maybe there's something there, maybe there isn't, but you don't just dismiss or accept it out of hand without looking at experimental evidence yay or nay.
..don't panic
Doesn't Windows make your computer a quantum computer?
You never know its stability state until you attempt an operation. Upon doing so you can't tell what it will do next.
(With apologies to Mr. Schrodinger and Mr. Heisenberg)
"They do not preach that their god will rouse them, a little before the Nuts work loose." Kipling, 'The Sons of Martha'
As for the use of quantum computers in AI - at present, nobody has provided an example of a vaguely AI-related problem that quantum computers of the type currently being studied would be useful for. Somebody may do so in the future, of course. In any case, anything that can be done on a quantum computer can be simulated on a normal one (in a theoretical sense, it may take till the end of the universe to do so). They don't give you the ability to compute anything "non-algorithmic".
Any sufficiently advanced technology is indistinguishable from a rigged demo
--Andy Finkel (J. Klass?)
There was a recent discussion about quantum computers (QCs) on sci.crypt. The consensus is, given a powerful enough QC, all public-key methods (RSA, Diffe-Helman, Elliptic Curve systems, etc) are badly broken by Shor's algorithm.
But symmetric ciphers (AES, DES, Blowfish, Serpent, etc) only have their effective key length cut in half, as a consequence of Grover's algorithm for searching an unordered list in O(sqrt(N)) time. So 64-bit keys become crackable with 2^32 work, and 128-bit keys in 2^64 work. Using 256-bit symmetric keys is considered sufficient to negate the threat of QCs.
I'm not sure about other cryptographic constructs such as PRNGs (Yarrow, ANSI X9.17) or hash functions (SHA-1, MD5), but I'm guessing at worst you would just have to double the size of the internal state to achieve security levels comparable to today.
Disclaimer: IANAC (I am not a cryptographer) but I do know quite a few.
Democracy is two wolves and a sheep voting on lunch.
Either this story has been severely garbled by journalists or its an outright lie designed to get funding.
-- SIGFPE