Slashdot Mirror
1 Molecule Computes Thousands of Times Faster Than a PC
alexhiggins732 writes with this tantalizing PopSci snippet: "A demo of a quantum calculation carried out by Japanese researchers has yielded some pretty mind-blowing results: a single molecule can perform a complex calculation thousands of times faster than a conventional computer. A proof-of-principle test run of a discrete Fourier transform — a common calculation using spectral analysis and data compression, among other things — performed with a single iodine molecule transpired very well, putting all the molecules in your PC to shame."
I think we are going to see a lot more of this sort of thing as humans get better and better at organizing matter into computing machines. The future is looking very very bright!
"linux is just DOS with a UNIX like syntax" -- Galactic Dominator (944134)
Hoping this will be the breakthrough that keeps Moores Law in business :P
~don't feel threatened by my pineal~
Something that could play Crysis with no lag!
Add more table salt.
In a way. thats just the same as claiming a laser can caluclate a 2D FFT if you look at the frauenhofer diffraction of an aperture.
Or that single candle can render better than any GPU by the way a room looks like when its illuminated by it.
You just have to redefine a basic property of your system as "calculation"
HI O WISE PRINCE. WHT TOOK U SO DAM LONG?
I really hate it when people come up with the simple "Quantum computer 1000 times faster than conventional computer". It's not just overly simplistic, it's wrong.
Quantum computers can turn some problems that require exponential time to solve into a polynomial time. So instead of taking 2^n time, it might take n^3 time. That's cannot in any realistic way be described as being "X times faster".
AccountKiller
...one molecule ought to be enough for anybody!
Popsci can't write, submitter can't read, and timothy doesn't know what a DFT is.
Just imagine a beowulf cluster of these. (ducks)
They dope it with steroid and massive amount of MSG.
Fuck systemd. Fuck Redhat. Fuck Soylent, too. Wait, scratch the last one.
i accidentally the whole cup
I'm no quantum theory expert, but does this represent the limit? Or are there some hypotheses about doing calculations with smaller particles?
I've never seen a quantum computing device smaller than the size of a small room, so I'm not really sure how fair it is to compare it to a PC.
Really the PC doesn't even use full atoms for calculations, it uses electrons and electron holes in the atoms, and its at least 2000 times smaller than any quantum device I've seen.
You don't really get to say its one molecule when its a device made up of a fuckton of molecules and you are comparing too it a PC which uses subatomic elements to actually do the work.
You have a fast calculator ... the size of a room ... which I can put 2000 slower and easier to make calculators in and end up faster.
Sure, eventually, they'll make it smaller and smaller, but your comparison is like saying using an f16 to deliver mail is faster than using a postal truck to deliver milk. Just because you make two statements that share a verb doesn't mean you've made a comparison thats in any way meaningful.
Persistent Volume manager for Kubernetes - https://github.com/dwimsey/openshift-pvmanager
I can't wait until something like this is available in a tattoo with a thought interface and neural wifi - upgrade as necessary! Wouldn't it be awesome to share with someone else that creative spark you ignited on "Item 9", or replay your acid trip when you're bored? Then start hacking all the molecules around you to do your bidding! Well, hopefully this isn't line 1 of why I got sent to the funny farm.
A one molecule computer faster than a PC. I find that hard to believe. My Asus Netbook is powered by one "atom", and it's still dog slow.
Take Nobody's Word For It.
I want one so i can overclock it by adding neutrons.
They whose government reduces their essential liberties for temporary security, receive neither liberty nor security.
"a common calculation using spectral analysis and data compression" should read "a common calculation used in spectral analysis and data compression" instead.
.... say bye bye to encryption...
"drill that hole baby .. *fart*"
seen at digg
meteorite impact
Encryption that relies on hard problems on computers we're accustomed to, sure, but there are also quantum crypto methods that will become cheap and available in the future. There are special things you can do with quantum crypto (that you can use today if you're rich) that you can't do with ordinary crypto too -- like detecting when people listen in. I don't think this represents the end of home crypto, perhaps a long vacation though.
Imagine if you weren't allowed to use roads because a bus company complained about your driving 3 times. --skunkpussy
does the molecule run Linux?
Using quantum interference - the vibrations of the atoms themselves - the team was able to run the complete discrete Fourier transform extremely quickly by encoding the inputs into an optically tailored vibrational wave packet which is then run through an excited iodine molecule whose atomic elements are oscillating at known intervals and picked up by a receiver on the other side. The entire process takes just a few tens of femtoseconds (that's a quadrillionth of a second).
- vibrations, atoms, excited, femtoseconds.... that sounds like it describes my sex life, can I too, calculate a Fourier transform a few thousand times faster than my desktop? I want 'it' to be useful for once!
You can't handle the truth.
How many FPS out of Halo will I get with this discovery?
also, plain old XOR (with OTP white noise key the size of the cryptogram) remains unbroken.
45 5F E1 04 22 CA 29 C4 93 3F 95 05 2B 79 2A B2
WOW!
the molecule might have the right answer, but i imagine that it can only give you a proabilistic answer
Quantum computers can turn some problems that require exponential time to solve into a polynomial time.
you mean transforming nondeterministic polynomial (NP, or deterministic exponential) into polynomial (P) problems, then this is wrong:
"There is a common misconception that quantum computers can solve NP-complete problems in polynomial time. That is not known to be true, and is generally suspected to be false."
The word "some" doesn't save you either, if you do it for one NP-complete problem, you'd just gotten yourself a Fields Medal :)
I know for a fact that some molecules can in a few milliseconds solve complex protein folding problems that take supercomputers years. Folding@Home could be revolutionized.
http://physics.aps.org/pdf/10.1103/PhysRevLett.104.180501.pdf
But the only so-called quantum cryptography I have heard of is hardly cryptography. It is a way to generate a shared secret between 2 computers that happen to be directly linked by optical fiber, while detecting any attempt at eavesdropping. It is worthless unless the two computers that want to communicate have a dedicated fiber line that connects them.[1]
Further while it prevents eavesdropping, it does not prevent full blown man-in-the-middle, where the fiber is severed, and converted into a pair of fibers with the adversary in the middle.
[1] Actually there are some tricks that allow for non-dedicated lines by using a chain of fully trusted relays with quantum secured channels between them, but that is still insane. A network of fully trusted relays simply cannot scale.
Stylish sheet to fix many problems in Slashdot's D3: https://gist.github.com/801524
...You could run Crysis on about half a ton of iodine?
Unless your Microsoft(tm) white noise generator generates the key 00000000.......
And say hello to theoretically unbreakable (not 10^15 years unbreakable, literally unbreakable) quantum entanglement-based one time pads.
but does it have MIDI?
No. NP-complete is different from NP. There are several NP (but not NP-complete) problems that quantum computers can solve in polynomial time: integer factoring, for example.
Also, NP doesn't mean deterministic exponential. There are sub-exponential problems in NP too.
The silicon processor in my PC is just one molecule as well and it can do much more than a FFT.
Excuse me, but please get off my Pennisetum Clandestinum, eh!
One time pads already are unbreakable.
Actually, symmetric cryptographic algorithms like AES don't take much of a hit from quantum computers at all. Just double the key size, and the speed advantage of Shor's algorithm is completely negated. Contemporary asymmetric algorithms like RSA can be broken in polynomial time, but there are so-called post-quantum algorithms that cannot be broken efficiently by any known quantum algorithm, and do not require a quantum computer to use. Such algorithms are not theoretically guaranteed to be unbreakable like quantum key exchange cryptosystems are, but they are no more vulnerable to quantum computers than RSA is to deterministic computers today.
I just wish they would finally come up with something that is production-ready.
There are so many uses for FFT, it’s not even funny. And all normal algorithms always will be imperfect and slow.
Instant FFT (and inverse FT) would (also) instantly change the world.
Any sufficiently advanced intelligence is indistinguishable from stupidity.
Think of all the silicon that is going to be put out of work by one atom of iodine.
I'm taking this with a grain of salt-
New slashdot layout sucks.
Also, take a look at BQP (bounded error quantum polynomial time). It is suspected (the same way as P != NP is suspected) that BQP includes all of P, some of NP (but not NP-complete), and some of PSPACE.
Hardware FT outputs are ... old hat. Any lens outputs FTs for free. But ... only humans do calculations ... hardware just **is**. That's right btye.boyz ... maths/algorithms/logic is performed only in the human mind.
PopSci sources http://www.sciencedaily.com/releases/2010/05/100504220042.htm which sources:
Story Source:
Adapted from materials provided by American Physical Society, via EurekAlert!, a service of AAAS.
Journal References:
1. Kouichi Hosaka, Hiroyuki Shimada, Hisashi Chiba, Hiroyuki Katsuki, Yoshiaki Teranishi, Yukiyoshi Ohtsuki, Kenji Ohmori. Ultrafast Fourier Transform with a Femtosecond-Laser-Driven Molecule. Physical Review Letters, 2010; 104 (18): 180501 DOI: 10.1103/PhysRevLett.104.180501 - http://prl.aps.org/abstract/PRL/v104/i18/e180501
2. Ian Walmsley. Ultrafast computing with molecules. Physics, 2010; 3: 38 DOI: 10.1103/Physics.3.38 - http://physics.aps.org/articles/v3/38
is no longer valid. The story isn't there.
Nope. If you're saying a quantum computer can solve NP-complete problems, that's generally believed not to be true (in the same sense that it's generally believed than P != NP). Even if that's not what you're saying, you would have to show that a quantum computer could break, say, AES, in polytime. Grover's algorithm can cut the number of bits required to brute-force in half, but to guard against it, just double the number of bits - 256-bit AES should still be hard to break. So private-key encryption (AES and the likes) are safe unless shown otherwise susceptible.
That leaves public-key crypto, which is probably what you're thinking about. True, with a quantum computer, one can factor numbers in polynomial time (Shor's algorithm), and thus break RSA. But RSA isn't the only public-key cryptography algorithm around. First, even in the worst of worlds short of BQP containing NP, one can make digital signatures using any sort of one-way function: the construct is called a Lamport signature. Second, there are public-key cryptosystems that seem hard to break using quantum computers - the McEliece system, which is based on error-correcting codes, is one of them.
If it turns out BQP contains NP, then there's always quantum cryptography, using the laws of physics to hide the message. However, the practical implementation of quantum cryptography is quite difficult, since if the laser emits more than a single photon for each "burst" (entangled bit), the scheme can be broken. To get over the loss rate implied by single photons, the crypto would either have to be by the use of lasers in free air, or have lots of quantum repeaters along the fiber - more than would be the case for a traditional message.
In any case, if BQP contains NP, we'd have a magic machine that can solve any puzzle that can be quickly verified. Finding optimal solutions to general engineering problems would just be a matter of churning the specs through the machine. The world would change -- very quickly and very radically -- and concerns about crypto would seem slight in comparison.
... because quantum computers are NOT known to solve any problem that "requires exponential time" in polynomial time.
This is one of the most frequent misconceptions about quantum computing. Take a look at Scott Aaronson's blog:
http://www.scottaaronson.com/blog/
It states "Quantum computers are not known to be able to solve NP-complete problems in polynomial time" for a reason ...
Quantum computers can factorize numbers in polynomial time. However determining prime factors is NOT known to be NP-complete, even though it is commonly assumed to be "hard".
But don't trade in your conventional computing power just yet. Like other quantum information platforms, molecular computing is in its infancy; we understand some of its mechanisms, but it's difficult to execute and there are still a lot of unknowns. Further, researchers aren't quite sure how they could integrate such technology into something that works the way we're used to our computers working.
It's called a brain.
Big apple, new Yorik, undig it, something's unrotting in Edenmark.
"New technology is faster than old technology!"
nah... you underestimate them.
1111-1111-1111-1112
45 5F E1 04 22 CA 29 C4 93 3F 95 05 2B 79 2A B2
It was indeed a mere observation of conjuncture. That said, it has been an extraordinarily useful one in the form of a challenge to humankind. Without it we would not have progressed the way we have. Intel is using Moore's law as a road map, forcing other companies *coughAMDcough* to innovate just to keep up. And that is why we have the enormous speeds available today. So we have a prediction that shaped the future. Why bother? Because our dreams shape our world.
If video games influenced behavior the Pac Man generation would be eating pills and running away from their problems.
trusted relays doesn't sound very secure.
New things are always on the horizon
I bet the interface speed is great too using I/Odine (somebody had to....)
This thread discussing whether an analog computation by a physical device is really a computation (it is, a world of non-digital computers once existed) provokes another question in my mind. What haven't we seen electronic analog technologies developed to implement neural network computation, which is intrinsically analog in the first place? Why must the combining of inputs in a neural device be simulated numerically?
This bears somewhat on the prospects of quantum computing. It seems to me that electrical processes in standard materials that precisely mimic biological neural processes should be relatively easy to develop (compared to quantum computing), and that they would be far faster than numerical simulation. Yet special neural devices (in the same sense that a transistor is a device) and circuits do not seem to be prominent in neural computing. If this relatively accessible technology space has not been turned into commercial hardware, it suggests the immense difficulty that quantum computing will have.
Starships were meant to fly, Hands up and touch the sky - Nicky Minaj
42
So you are telling me that a dedicated, built from scratch minimal "circuit", specifically designed to solve a particular mathematical problem, where even in the inputs have been specially converted to make it easier for the circuit to process, and the "output" is likewise interpreted in the native capacity of the circuit, is many times a faster than a general purpose entire computer, that has been designed to complete not only thousands of specific tasks, but also untold number of unknown ones, under a huge variety hardware platforms?
Duh, not shit Sherlock.
I understand they're trying to say that quantum computing could be many times faster than today's CPUs, but a simple comparison like this is a non-comparison. Talk about comparing apples and oranges... When you have something resembling a general purpose circuit that can take arbitrary inputs and outputs, then maybe we can start comparing to a modern CPU...
In a real emergency, we would have all fled in terror, and you would not have been notified.
I can't wait until they port GCC over!
but *MY* computer doesn't leave indelible purple stains on the tablecloth!
Can it run Doom!?
"you'd just gotten yourself a Fields Medal :)"
And refuse it like http://en.wikipedia.org/wiki/Grigori_Perelman
One time pads already are unbreakable.
One-time pads require true randomness in order to be unbreakable, which (last time I checked) wasn't a feasibility for traditional desktop computers. Now I'm sure there's a subset of Slashdotters who sport their own custom half-life sampling hardware... but regardless, there is no realistic implementation for this algorithm. Any randomly generated pad would have to be securely transferred to both parties. If you are already going through all of the trouble of distributing massive pads, it would make more sense to simply hand the person you are trying to communicate with the plain-text of the message.
However, I don't fear the proposed "end of internet privacy," since current quantum computers take up an entire room, and possess only a handful of qubits. I imagine that by the time the quantum computer has been scaled to the point where we see modern-day cryptography such as RSA break, quantum encryption will already be implemented on a broad scale.
If it is true, then we should practice this molecule instead of PC. What do you say? http://www.articlesbase.com/health-articles/acai-max-cleanse-does-acai-many-cleanse-remove-extra-pounds-1954235.html
Is the article really talking about quantum computing? it seems it is talking about molecular computing. I didn't see anything relating to quantum, i.e. taking into advantage the superposition of states etc.
The title or subtitle of this should really be; Quantum Particle Performs Thousands Of Times Faster Than Simulations of Quantum Particle.
The authors found a behavior of a particle which is useful computationally - and compared performance of that against simulations. Most simulations of anything are like that - slower than the original behavior they're meant to simulate. This is not at all surprising. But this is far different from having a general purpose computing engine, and I don't believe you could make such from purely from Fourier functions in any economical fashion. But ooh, particles are fast!
This technique should improve the ability to generate a truly random number at least,... at least for a little while.
...it still can't play Crysis on max settings.
That's pretty stupid. Idea is you solve the distribution once and can relay messages many times after that. Give a few harddrives worth of one-time pad to your submarine captain or embassy or agent and broadcast messages to him as long as you like.
That's my point. If there are very few, they can be run in a very transparent manner and the small number of interested parties (the sysadmins of each computer for example) can verify the security. But as there are more, they inherently become less secure.
That is what I meant by "a network of fully trusted relays simply cannot scale."
Stylish sheet to fix many problems in Slashdot's D3: https://gist.github.com/801524