A superconducting chip developed at IBM demonstrates an important step needed for the creation of computer processors that crunch numbers by exploiting the weirdness of quantum physics. If successfully developed, quantum computers could effectively take shortcuts through many calculations that are difficult for today's computers.
IBM's new chip is the first to integrate the basic devices needed to build a quantum computer, known as qubits, into a 2-D grid. Researchers think one of the best routes to making a practical quantum computer would involve creating grids of hundreds or thousands of qubits working together. The circuits of IBM's chip are made from metals that become superconducting when cooled to extremely low temperatures. The chip operates at only a fraction of a degree above absolute zero.
IBM's chip contains only the simplest grid possible, four qubits in a two-by-two array. But previously researchers had only shown they could operate qubits together when arranged in a line. Unlike conventional binary bits, a qubit can enter a "superposition state" where it is effectively both 0 and 1 at the same time. When qubits in this state work together, they can cut through complex calculations in ways impossible for conventional hardware. Google, NASA, Microsoft, IBM, and the U.S. government are all working on the technology.
Nature Communications: Demonstration of a quantum error detection code using a square lattice of four superconducting qubits
(Score: 0) by Anonymous Coward on Friday May 01 2015, @03:04PM
Is that what's new here?
(Score: 1, Funny) by Anonymous Coward on Friday May 01 2015, @04:42PM
First off, it's spelled cubits. And secondly, they're not new, they've been around since ancient Egypt.
(Score: 2) by Megahard on Friday May 01 2015, @06:58PM
Yes, it can now simultaneously solve all 2x2 tic-tac-toe games.
(Score: 2) by bob_super on Friday May 01 2015, @03:25PM
Is there a human-readable (my quantum physics classes are twenty year behind me) description of the concept "shortcuts through many calculations"?
I'd like to understand to understand how you tell a quantum computer to do a specific thing, and accurately retrieve the answer, but the journalists reporting breakthroughs always wisely steer clear of the "how/why does it work" section...
(Score: 4, Informative) by VortexCortex on Friday May 01 2015, @06:49PM
My code exists in a superposition of compiling and testing, so I'll try to give an example.
Let's say you have a quantum adding circuit with some word-size of qubits. In a traditional adding circuit you can only add two separate bit patterns at once through the circuit; However, with a quantum adding circuit you could feed in two qubit patterns, each pattern could represent a superposition of multiple values. When the quantum adding circuit performs its single add instruction it has performed the addition of all those different superpositions at once, and resulted in a new pattern of qubits that represents all the possible outputs from all the different inputs. Now, not all qubits need to be in superposition states, so you don't have to test every value at once, but a true quantum computer would be able to. This is similar to a single circuit being a parallel circuit.
E.g., say you want to add the values 0 or 1 to the values 4 or 6, but we don't know which ones yet. We need three qubits, and I'll use "?" to represent superpositions. In binary the instruction could look like: result = 00? + 1?0; Result would then contain a single set o qubits in superpositions that represents all the values: 4,5,6 or 7 (decimal); 100, 101, 110, or 111 (in binary), or 1?? (the notation falls apart here because we need to entangle even values with the 0 input, odd values with the 1, values less than 5 with the 4 and greater than 5 with the 6 input). Other calculations could then be performed on the resulting value. For instance, we might have a constraint that says the answer we're looking for is prime. A quantum prime "filter" would collapse the result to being only 5 or 7, and collapse the first input superposition from being 00? to 001.
The classic example application is cryptography. Imagine you have a hash digest (a specific pattern of bits resulting from feeding a specific series of bits into a hashing algorithm). Let's say you also want to forge a document, modifying it such that when fed into the hash algorithm the digest (fingerprint) still matches. Most hash algorithms can generate many "collisions" whereby two or more inputs produce the same digest output since the digest length is typically shorter than the input data; However, hash algorithms are designed to be difficult to discover collisions without applying brute force (trying all possible inputs until you find a match). To forge the document and inject your exploit into it you'll need to know what other bits besides the exploit-payload you need to flip to produce a hash digest that matches the original. With today's computers you'll have to brute force the flipping of a number of bits corresponding to the size of the hash algorithm's internal state, which could possibly take billions of years. In your quantum computer's registers the message you want to feed in could be kept strictly 1s and 0s, but the other parts that could be bit-flipped could be represented by superpositions of both 1 and 0 -- To make changes less noticeable perhaps you specify bits that toggle capitalization of some select ASCII characters in the document.
A traditional computer would have to perform a number of hash/test iterations equivalent to about half of every possible input combination (on average). However, when you perform your hashing algorithm only once with a quantum dataset you'll get a single result that essentially represents every possible bit-flip at once, i.e., the result will contain many superpositions upon superpositions via entanglement. The next step is to feed the resulting value into a quantum circuit that collapses the superpositions into the desired bit pattern matching the digest of the document you wanted to spoof. This then also collapses the superpositions of the input dataset and reveals what input you need to feed in to spoof the hash. Thanks to its qubit superpositions the quantum computer performs this calculation in a single iteration (constant time) by computing and testing every combination of the possible bit-flips.
This is all a bit over simplified, but one should get the gist.
(Score: 0) by Anonymous Coward on Friday May 01 2015, @07:31PM
So you are saying that with just a few qbits, all encryption would be readily breakable in just a few iterations instead of years measured by orders of magnitude?
That seems very far fetched. Is there any proof of this or just theory?
(Score: 2) by rts008 on Friday May 01 2015, @11:58PM
I'm pretty sure it's just theory at this stage.
To my knowledge, no actual working quantum computer 'large/powerful enough' to test this theory has been built yet.
All of the quantum computers I've heard about are rudimentary, like the one in the article.
IMO, quantum computers are on the same level today, as when the first electronic cumputers were being built in labs. Not even on the computing level of a cheap digital watch, at least yet.
If anyone knows better feel free to chime in, I'm far from an expert on quantum computers. :-)
(Score: 2, Informative) by Fauxlosopher on Saturday May 02 2015, @06:54AM
Basically yes, with the exception of at least one [worldtruth.org] type of encryption method [cqc2t.org] based around the use of a "one-time pad [wikipedia.org]".
Properly-used OTP-encrypted messages are supposed to completely uncrackable and completely impervious to any breakthroughs to include quantum computing. However, a major problem with OTPs is secure delivery of the pad itself (which is also the key) from the message sender to the recipient, which is why OTP-based methods are not used as often as others like public key cryptography [wikipedia.org].
(Score: 2) by bob_super on Friday May 01 2015, @09:09PM
Thanks for the long explanation. It does make sense.
My question was more basic, though: How do you feed the operands, trigger a specific operation, and retrieve/collapse the result?
(Score: 0) by Anonymous Coward on Friday May 01 2015, @03:30PM
It is more like massively parallel. You can try all the combinations to the lock instantly...
If anyone suggests they've found a shortcut in math they are usually full of shit.
(Score: 3, Informative) by takyon on Friday May 01 2015, @03:56PM
https://en.wikipedia.org/wiki/Quantum_computing [wikipedia.org]
[SIG] 10/28/2017: Soylent Upgrade v14 [soylentnews.org]
(Score: 2) by mtrycz on Friday May 01 2015, @04:54PM
Thanks.
One thing I can't decipher is "polynomial speedup". Does this mean that O(2^n) becomes something like O(n^x) for some x, or does that mean that O(n^x) becomes O(n^(x-c)), hence possibly O(n^2) -> O(n).
Or something else entirely?
In capitalist America, ads view YOU!
(Score: 2) by takyon on Friday May 01 2015, @05:07PM
I think it's like this: https://jsfiddle.net/g4q08m4a/ [jsfiddle.net]
Exponentials will outpace any polynomials eventually.
[SIG] 10/28/2017: Soylent Upgrade v14 [soylentnews.org]
(Score: 2) by takyon on Friday May 01 2015, @05:14PM
More readable: https://jsfiddle.net/g4q08m4a/1/ [jsfiddle.net]
[SIG] 10/28/2017: Soylent Upgrade v14 [soylentnews.org]
(Score: 2) by maxwell demon on Friday May 01 2015, @05:25PM
The first one is more than polynomial speedup (e.g. seen for Shor's algorithm), the second one is polynomial speedup (e.g. seen for Grover search).
In the case of Grover search, it's in particular O(n) -> O(n1/2).
The Tao of math: The numbers you can count are not the real numbers.
(Score: 1) by dak664 on Friday May 01 2015, @06:32PM
Well that's the idea. But it seems to me the discussion of computation time is lacking; adjacent qubits have to superimpose over some time scale before settling into the answer, and that answer is statistical to boot.
An excited nucleus is basically a qubit(?) and it can take years or centuries to beta decay when coupling to the continuum.
(Score: 1) by gallondr00nk on Friday May 01 2015, @05:44PM
Windows programs have been doing that for decades.
(Score: 2) by kaszz on Friday May 01 2015, @05:47PM
Current memory technology seems to have the same capability.
(Score: 2) by rts008 on Saturday May 02 2015, @12:08AM
Well, to be fair, Windows can also neither 0 and/or 1 at the same time.
It's worse than Schroedinger's cat, or not.
I don't think any of my Windows PC's have had a cat in the box though, every time I have thrown one out the window in frustration, not once did any of them land on their feet.
(Score: 0) by Anonymous Coward on Friday May 01 2015, @07:46PM
Until someone can actually run an experiment where they find the solution to say, a 1024 bit problem in one computational iteration, I don't buy this. After all, if a qbit can be both 1 and 0 simultaneously, they can solve every computational problem in just one cycle if there are a sufficient number of qbits and two cycles if there are not, since the two cycles can create superpositions upon superpositions of themselves.
If this worked, I would expect that some government agency somewhere already has a device that can break all encryption of any strength in a tiny fraction of a second. There is a functionally unlimited sum of money available for the creation of such a device, and according to this, quantum computation on a chip is available for purchase, therefore it would be of extreme foolishness to think that the theory is correct and no one has taken advantage of it simultaneously.