Beta
×

Welcome to the Slashdot Beta site -- learn more here. Use the link in the footer or click here to return to the Classic version of Slashdot.

Thank you!

Before you choose to head back to the Classic look of the site, we'd appreciate it if you share your thoughts on the Beta; your feedback is what drives our ongoing development.

Beta is different and we value you taking the time to try it out. Please take a look at the changes we've made in Beta and  learn more about it. Thanks for reading, and for making the site better!

First Quantum Computing Gate on a Chip

Zonk posted more than 7 years ago | from the next-advance-is-really-tiny-stargates dept.

Supercomputing 166

An anonymous reader writes "After recent success in using quantum computing for superconducting qubits, researchers from Delft have formed the first Controlled-NOT quantum gate. 'A team has demonstrated a key ingredient of such a computer by using one superconducting loop to control the information stored on a second. Combined with other recent advances, the result may pave the way for devices of double the size in the next year or two--closer to what other quantum computing candidates have achieved, says physicist Hans Mooij of the Delft University of Technology in the Netherlands. Unlike today's computers, which process information in the form of 0s and 1s, a quantum computer would achieve new levels of power by turning bits into fuzzy quantum things called qubits (pronounced cue-bits) that are 0 and 1 simultaneously. In theory, quantum computers would allow hackers to crack today's toughest coded messages and researchers to better simulate molecules for designing new drugs and materials.'"

cancel ×

166 comments

Sorry! There are no comments related to the filter you selected.

Double the size of a single not gate? (5, Funny)

Spazntwich (208070) | more than 7 years ago | (#19630183)

I know grammar has been taking a hit in society as of late, but now even our computers are blatantly spewing out double negatives?

We're not in for an unrough ride, gentlemen.

Re:Double the size of a single not gate? (1)

MyLongNickName (822545) | more than 7 years ago | (#19630239)

Where is the second negative?

Re:Double the size of a single not gate? (1)

xouumalperxe (815707) | more than 7 years ago | (#19630739)

It's a double-size not gate.

Re:Double the size of a single not gate? (2)

MadnessASAP (1052274) | more than 7 years ago | (#19630247)

More to the point I thought the idea was to make these things smaller not bigger? Unless I'm missing something.

Re:Double the size of a single not gate? (0)

Anonymous Coward | more than 7 years ago | (#19630319)

That reminds me:
What happens when you divide by zero in a quantum computer?

Re:Double the size of a single not gate? (1)

Aladrin (926209) | more than 7 years ago | (#19630409)

According to the summary, it's the same thing as dividing by 1 on a quantum computer.

Let's try it... -universe asplodes-

Re:Double the size of a single not gate? (2, Funny)

Jazz13 (914138) | more than 7 years ago | (#19630703)

Rodney McKay steps through the resulting tear in space time and smacks you up the back of the head.

It depends... (1)

msauve (701917) | more than 7 years ago | (#19630833)

on whether someone is observing (or not, which is not and in negative logic).

Imagine slashdot on a quantum server! (3, Funny)

EmbeddedJanitor (597831) | more than 7 years ago | (#19631253)

You'd never know if an article was a dupe or not.

Re:Imagine slashdot on a quantum server! (1)

Spazntwich (208070) | more than 7 years ago | (#19631433)

Maybe they could be both. Or an article could be a dupe of itself.

Personally, I'd just enjoy seeing them post a picture of a dead cat as a story.

A solid milestone... (5, Interesting)

teebob21 (947095) | more than 7 years ago | (#19630241)

I find it interesting that the first electronic computing gates devised were the AND/OR gates, using basic diode logic. Quantum computing research develops the NOT gate first. I think this has something to do with the esoteric nature of quantum computing. AND/OR gates require two inputs to change to a single value, where NOT is merely an inverter. The idea of entanglement makes the inversion process a likely first step in quantum research.

For those wondering why this is important, the first true electronic gates were invented in the early 1920's. This predates point-contact transistors by about 20 years, invented in 1947. 60 years later, here we are with transistor computing in every aspect of our lives.

At the rate quantum computing is advancing, I think we can expect to see quantum transistors (in the lab, at least) by 2020. A true useful quantum computer may be available less than 50 years from now. Hopefully by then someone will pick up the slack and have the Linux kernel ported to the Q-CPU architecture!

Re:A solid milestone... (2, Interesting)

Ant P. (974313) | more than 7 years ago | (#19630277)

What is a quantum computer good for, anyway? So far all I've seen is cracking encryption and other stuff involving gigantic calculations. Is there anything in the mainstream market it'd be useful for, like sound/video processing?
Come to think of it, arithmetic encoding is a bit like encryption...

Re:A solid milestone... (1)

Yez70 (924200) | more than 7 years ago | (#19630345)

Consider the possibilities of complex artificial intelligence and this could be applied to virtually any aspect of our lives.

Re:A solid milestone... (0)

Anonymous Coward | more than 7 years ago | (#19630375)

Any kind of blind search (code cracking, chess, genetic algorithms, you name it) takes the square root [wikipedia.org] of the time it does on a classical computer with the same space available. That's pretty neat.

Re:A solid milestone... (1)

asuffield (111848) | more than 7 years ago | (#19630421)

Any kind of blind search (code cracking, chess, genetic algorithms, you name it) takes the square root of the time it does on a classical computer with the same space available. That's pretty neat.


However, the only kind of problems for which we would use blind search are very hard and take a very long time already, and the square root of a very slow algorithm is usually still a very slow algorithm. While they are "interesting" problems, they are largely esoteric in nature. Any practical applications of quantum computing are probably going to have to be more creative than the method you're referring to. We don't really know much about non-trivial quantum computing algorithms yet.

Re:A solid milestone... (2, Insightful)

Anonymous Coward | more than 7 years ago | (#19630539)

Most of the hard logistics problems are forms of nonlinear optimization and usually reduce to integer programming, which is solved by a pretty blind search (excluding tricks like branch and bound, pruning; it's still a tree search). Improving these solutions by a few percent translates into a lot of money very quickly.

Esoteric, maybe (it'll be a computer in a lab, not a PC), but usable nonetheless.

Used for what? (3, Funny)

symbolset (646467) | more than 7 years ago | (#19630485)

At home you will use these for ever more sophisticated rendering of artificially intelligent virtual reality porn.

At work it will be more useful in the advanced simulation of a mechanical process for imprinting letter glyphs on sheets of wood fiber.

Re:A solid milestone... (2, Funny)

Schemat1c (464768) | more than 7 years ago | (#19630849)

What is a quantum computer good for, anyway?
You could use it to store your recipes.

Re:A solid milestone... (1)

Climate Shill (1039098) | more than 7 years ago | (#19631251)

What is a quantum computer good for, anyway?

IIRC, quantum computers have one killer app, which is that they can simulate other quantum systems (i.e. Stuff). If you try simulating a lump of high temperature superconductor on a classical computer you won't get very far, but on a quantum computer you just might.

#disclaimer: This is slashdot, so I reserve the right to have been talking out of my a(r?)s[se].

Re:A solid milestone... (2, Funny)

pdbaby (609052) | more than 7 years ago | (#19631791)

Regex pedantry here, but you're being overly permissive - you allow semantically invalid words: as, ars & arss.

As a side-note, if you like hearing from your local regex pedant, please remember to donate g(ener|ratuit)ously: we survive only through your funding

...it's like I know I should be ticking 'Post Anonymously' but I just can't stop myself

Re:A solid milestone... (2, Funny)

afaik_ianal (918433) | more than 7 years ago | (#19632073)

you allow semantically invalid words: as, ars & arss.

No they don't. The only matches are ass, ase, arss and arse, but your point still stands. When being a pedant, it's also polite to provide a correction. The regex they were after is: a(rse|ss).

...it's like I know I should be ticking 'Post Anonymously' but I just can't stop myself

At least tick "No Karma Bonus", please :).

Re:A solid milestone... (2, Insightful)

Simon80 (874052) | more than 7 years ago | (#19630343)

I wouldn't know for sure, but I don't think it's valid to compare any form of computing based on binary logic with quantum computing. I searched for "quantum transistor" and found this [sandia.gov] , which makes use of the term to refer to transistors that rely on principles of quantum mechanics to function properly. This would be relevant to conventional computing, but not quantum computing. If I understand correctly, quantum computing is not a replacement for binary logic computing, but an alternative or supplement.

Re:A solid milestone... (0)

Anonymous Coward | more than 7 years ago | (#19630515)

Transistors *already* use quantum principles to operate. What we're looking for is some kind of fundamental "qbit holder" that stores all superpositions in, for lack of a better term, a created mini-universe. Computing through those involves somehow "selecting the desired universe", whatever the hell that means. It wouldn't operate like a transistor by any means, it's just a convenient analogy for what is the "bit holder" building block in current computing.

Re:A solid milestone... (1)

WarJolt (990309) | more than 7 years ago | (#19631129)

Transistors replace tubes. Qbits replace transistors. Transistors do the same logical job as tubes. Qbits do the same logical job as Transistors. The transistor drastically changed the way electronics were designed. I suspect Qbits will do the same thing. We still use tubes today however, so obviously there are some things that tubes do differently then transistors. When I bought my tube amp I pretty much stopped using one of my transistor amps. Qbits, transistors and tubes behave differently, but they can do some of the same jobs. Qbits I suspect will slowly replace certain logical blocks on a semiconductor. Qbits and transistors behave differently. They will still be used for the same computational tasks.

Re:A solid milestone... (3, Informative)

afaik_ianal (918433) | more than 7 years ago | (#19632299)

You're getting confused between the distinction between bits/transistors, and qubits/quantum gates. The difference between tubes and transistors is nothing like the difference between transistors and quantum gates. Bits and qubits are just as dissimilar.

Like transistors, tubes are just amplifiers. Contrary to popular belief, both are analogue components, but they have quite different responses to changes in current (the graphs are a different shape). Many people consider tube amplifiers to produce better sound quality than transistor amplifiers, and tubes will almost always handle clipping better than transistors. There's actually not much a transistor can do that a vacuum tube can't.

Either tubes or transistors can be used to implement logic gates. These gates can be built up to store and process bits of information. In the digital world, they are functionally similar. The advantages of tubes from the analogue world go out the window, and the advantages of transistors (size, speed and cost) come into their own.

Quantum computers are another thing entirely. The computers you and I are using at the moment are deterministic Turing machines. Quantum computers are non-deterministic. Quantum gates deal not in bits, but in qubits. They don't think in terms of binary data, they work with quantum states.

There are many things that these quantum components can do that conventional transistors cannot, but I don't think we'll be seeing quantum gates implementing binary logic in computers any time soon (if ever). The article unfortunately confuses the issue by making it sound like they've implemented a binary operation. CNOT is not a binary operation - it is a quantum operation.

I hate analogies as much as the next guy, but if tubes are the horse and cart, and transistors are our cars, then quantum computing is going to be our interstellar space craft - they suck for doing the weekly shopping.

Re:A solid milestone... (5, Informative)

pablob (567257) | more than 7 years ago | (#19630505)

I find it interesting that the first electronic computing gates devised were the AND/OR gates, using basic diode logic. Quantum computing research develops the NOT gate first. I think this has something to do with the esoteric nature of quantum computing. AND/OR gates require two inputs to change to a single value, where NOT is merely an inverter. The idea of entanglement makes the inversion process a likely first step in quantum research.

Keep in mind that this is a Controlled-NOT gate (a two-input gate) and not a simple NOT gate (a one-input gate). It has been proven that if you can implement two-qubit C-NOT and arbitrary one-qubit operations, you can implement a universal quantum computer (that is, one that can run an arbitrary "quantum program").

There is a deeper reason for quantum computers not to use AND/OR gates, which is their irresibility (AND and OR are two-input to one input gate, which makes them irreversible). Quantum Mechanics is intrinsically reversible, so a quantum computer should be reversible too and that's why it is not common to hear about Quantum AND or Quantum OR.

Pablo B.

Re:A solid milestone... (1)

Climate Shill (1039098) | more than 7 years ago | (#19631305)

this is a Controlled-NOT gate

Is there a reason why it's called a NOT gate rather than XOR ? Is there some quantum wierdness that makes the thing asymmetric and causes A-inverts-B to mean something different to B-inverts-A ?

Re:A solid milestone... (1)

Max Littlemore (1001285) | more than 7 years ago | (#19631391)

Yes and no...

Re:A solid milestone... (3, Informative)

pablob (567257) | more than 7 years ago | (#19631455)

Is there a reason why it's called a NOT gate rather than XOR? Is there some quantum wierdness that makes the thing asymmetric and causes A-inverts-B to mean something different to B-inverts-A?

I guess it's mostly historical, and that XOR is usually associated with a two-input/one-output gate. Controlled-NOT looks pretty intuitive (the target is negated only if the control is 1), but probably XOR is better because A-inverts-B is exactly the same as B-inverts-A (which seems counterintuitive when you think of it as a CNOT, but it's reasonably simple to see that is the case by just looking at the corresponding truth table).

But I would venture that most people working on trying to build quantum computers would be more familiar with CNOT than XOR in the context of quantum computation (QXOR sounds weird, doesn't it?)

Pablo B.

Re:A solid milestone... (0)

nahdude812 (88157) | more than 7 years ago | (#19631497)

NOT is only true if A is true and B is false. A not B.

true NOT true = false
true NOT false = true
false NOT true = false
false NOT false = false

XOR is true when A and B are different.

true XOR true = false
true XOR false = true
false XOR true = true
false XOR false = false

The difference between NOT and XOR is in the false-true comparison. Actually officially (and maybe easier to conceptualize) "A not B" is a short hand of "A and not B," where not is an operator that applies only to B and inverts it. There is more info at the Wikipedia article on logic gates [wikipedia.org] .

Re:A solid milestone... (1)

IWannaBeAnAC (653701) | more than 7 years ago | (#19631647)

You made that up, didn't you? Even the Wikipedia article you mention doesn't say anything about a 2-input gate called 'NOT'. I've never heard of A NOT B as shorthand for A AND (NOT B) before, but I'll forgive you if you can post a link to where that terminology is used. But that gate has nothing to do with XOR, and definitely has nothing to do with the quantum C-NOT gate. The whole point of quantum gates is that they are reversible, of which AND and OR (and there negative-true counterparts) are not.

Re:A solid milestone... (5, Interesting)

asuffield (111848) | more than 7 years ago | (#19630521)

For those wondering why this is important, the first true electronic gates were invented in the early 1920's. This predates point-contact transistors by about 20 years, invented in 1947. 60 years later, here we are with transistor computing in every aspect of our lives.


However, it is important to realise that the theory of computation had been in development since the early 1800s (and the logic underlying that had been around for centuries); by the time the first electronic devices were created, we already had a good understanding of what they could be used for, because we had been doing exactly the same things by hand for over 50 years at that point (the word "computer" originally meant a person who performed such computations, and an "electronic computer" was just a device to replicate the task that person was doing).

We can't do quantum computations by hand, so we have no real experience with the theory, and the underlying statistical methods are relatively recent developments: quantum computers do not use the classical logic that we're all familiar with. This is a massive setback compared to the development of the electronic computer - and advances in theory usually can't be accelerated all that much. It is likely to be between 50 and 100 years before we know enough to build non-trivial applications out of quantum computers. Not because we can't build the hardware, but because we don't know how to write any software to run on them. The entire field of software development will have to be reinvented, and we don't actually know that it will be useful for anything. Unlike the first electronic computers, which had very real and obvious applications performing the tasks that were currently being done by hand, we have only vague theories and ideas about what quantum computers might be useful for. (Even the much-quoted method for breaking certain encryption algorithms is based on various assumptions that aren't proven; we don't know for sure whether quantum computers will actually be able to run it, yet)

We'll get there eventually, but it will probably take a long time and we can't really predict at this stage whether it'll be particularly useful. From what we know so far, these things are going to be incredibly arcane and obtuse to work with, and that is going to make it difficult. We might see it in our lifetimes, but I wouldn't place any bets on it, it might take much longer. The things we're playing with today may turn out to be the Babbage engines of quantum computing.

Re:A solid milestone... (2)

Dantu (840928) | more than 7 years ago | (#19631841)


We can't do quantum computations by hand, so we have no real experience with the theory, and the underlying statistical methods are relatively recent developments: quantum computers do not use the classical logic that we're all familiar with


I'm not an expert on Quantum computers, but I think the math/computer science is WAY ahead of the physics on this one. Please correct me if I'm wrong, but I think a Quantum computer is to a normal computer as a NFA is to an DFA (http://en.wikipedia.org/wiki/Nondeterministic_fin ite_state_machine). In this case it's really not a new idea at a fundamental logical level. I realize these only apply to regular languages, not general Turing problems, but that's because you can convert one to the other, so in THEORY they are equivalent anyway. Similarly, conventional computers can solve the same set of problems that quantum computers can solve, they just require an expansion of the problem which in the real world requires much more time and/or space to complete.

Of course, we don't have (that I know of) nice high-level compilers for them yet, but I think that's a whole different issue.

Re:A solid milestone... (4, Interesting)

jp102235 (923963) | more than 7 years ago | (#19630673)

Well, inverting logic is well, logical (no pun intended) to most modern digital logic designers of the CMOS type. CMOS logic (and its variants) are inherently inverting. That is, the basic gate in CMOS is an inverter. The next higher complexity of gates in CMOS is a NAND and NOR (AND NOT / OR NOT). To make an AND gate requires a NAND and an Inverter... same thing for OR : a NOR and an Inverter. Although the quantum abstraction of computation may not be the same as CMOS (inverting layers of logic) it is not surprising at all that the designers tried to make an inverter first. Had they started during the days of relays, we might have had a different gate altogether. JP

Re:A solid milestone... (2, Informative)

fatphil (181876) | more than 7 years ago | (#19630733)

Controlled-Not is not Not. Controlled-Not is "if the control line is 1, then not the input, else preserve the input", i.e. XOR. (But being quantum, it must also output the control line too, so that the operation can be reversed.)

Re:A solid milestone... (1)

IWannaBeAnAC (653701) | more than 7 years ago | (#19630917)

This isn't an ordinary NOT gate, it's a controlled NOT, or C-NOT. This gate does have two inputs. The state of the first input is either inverted or not inverted, depending on the state of the second input. So it is actually very similar to a conventional XOR gate.

By all means, pull numbers out your ass on what your predictions on when you think a quantum computer will be built, but you should at least put in a disclaimer that you have no idea what you are talking about. WTF is a `quantum transitor' ? There are a lot of designs of transistors that depend on quantum mechanical effects to operate, and are therefore called `quantum transistors'. These exist today, but these are ordinary transistors and have nothing to do with quantum computing. If you are trying to refer to some device that has a similar relationship to a quantum gate that a transistor has to a conventional gate, then I'm afraid you are out of luck - AFAIK there is no such device. For the designs referred to in the article, a Josephson junction could maybe be considered a fundamental building block - but it acts nothing like a transistor.

Re:A solid milestone... (1)

TheRealMindChild (743925) | more than 7 years ago | (#19631047)

Does that mean I can finally play Doom 3 in high quality mode?

Only a NOT gate? (0)

Anonymous Coward | more than 7 years ago | (#19630249)

Is this newsworthy? Wait until they can make a full adder, on a chip. Then I'll be impressed.

Like in Borat? (2, Funny)

SilentOneNCW (943611) | more than 7 years ago | (#19630251)

Not Jokes:

It's a Quantum Gate.... NOT!

Re:Like in Borat? (1)

AngryBacon (1094489) | more than 7 years ago | (#19630271)

It a NOT Quantum Gate, would be more accurate. :D

Re:Like in Borat? (1)

mattcoz (856085) | more than 7 years ago | (#19630613)

Maybe I'm getting old, but I'll always associate not jokes with Wayne's World, not Borat.

Fuzzy qubits of unknown distinction? (1)

DanTheManMS (1039636) | more than 7 years ago | (#19630263)

Sound a lot like Tribbles to me.

Re:Fuzzy qubits of unknown distinction? (2, Funny)

Joebert (946227) | more than 7 years ago | (#19630469)

Tribbles & Qubits: The new Pacman

DONT USE SOFTWARE WRITTEN BY INDIANS (0, Troll)

s16le (963839) | more than 7 years ago | (#19630265)

I would never run software on my computer that was coded by an Indian. The biggest reason is that they never wash their hands after using the restroom. They go straight from the urinal back to the cubicle and their back to hacking away on their keyboards.

Think about that- when you run indian-coded software, you wind up with dried urine and dead skin shedded from an indian person's dick in your CPU. Many of them don't clean themselves properly after defecating either, so you end up with bits of indian feces traveling through your address bus, execution stack... just don't use software coded by indians!!!

They say that it works (4, Funny)

zmollusc (763634) | more than 7 years ago | (#19630305)

but how can they test it when the output is always either 0, meh, pfft or 1?

Re:They say that it works (2, Funny)

Joebert (946227) | more than 7 years ago | (#19630487)

That's why they built a NOT gate, even if they failed, they'd still succeed.

Re:They say that it works (1)

fatphil (181876) | more than 7 years ago | (#19630773)

Because you can chose the basis in which you are going to perform the measurement to determine the answer, you can either measure along the 0-pfft axis, or the meh-1 axis. That simplifies things greatly.

Re:They say that it works (1)

Torvaun (1040898) | more than 7 years ago | (#19631273)

If a qubit is both 0 and 1 at the same time, what the hell does an inverter do? Make it 1 and 0 at the same time instead?

Re:They say that it works (1)

Your.Master (1088569) | more than 7 years ago | (#19631575)

Brings it from 0 and 1 at the same time to 1 and 0 at the same time.

Quantum states (4, Interesting)

arashi no garou (699761) | more than 7 years ago | (#19630323)

I'm no quantum theory expert by a very long shot, but it was my understanding that there are 32 quantum states of electrons, not just on/off (1/0) like in the binary computer world. So, if we now have a quantum NOT gate, doesn't that mean there are 32 possible states of the NOT gate? Also, according to the article the CNOT gates they created can be both 0 and 1 simultaneously. In my mind this would cause errors and actually stop the flow of information instead of speeding it up.

Someone with some understanding of this stuff please elaborate, before my head asplodes.

Re:Quantum states (4, Funny)

LighterShadeOfBlack (1011407) | more than 7 years ago | (#19630481)

but it was my understanding that there are 32 quantum states of electrons, not just on/off (1/0) like in the binary computer world. So, if we now have a quantum NOT gate, doesn't that mean there are 32 possible states of the NOT gate?
Well, yes and no...

*ba-dum-tsch* Thank you very much, I'll be here all week.

Re:Quantum states (0)

Anonymous Coward | more than 7 years ago | (#19630501)

That doesn't mean we have 32 NOTs. Think about it, we have a new 32-digit format just like bin, hex and dec on your calculator.

btw Bill Gates has way more states than that

Re:Quantum states (0, Troll)

QuantumG (50515) | more than 7 years ago | (#19630525)

I'm no quantum theory expert by a very long shot
No shit.

Do some research [wikipedia.org] .

Re:Quantum states (1)

arashi no garou (699761) | more than 7 years ago | (#19631105)

Wow, how very snarky of you! You could have done like the other three people here and kindly explained to me that there is a vast difference between what I had in my head (quantum physics) and quantum computing. But no, you had to show your obvious superior wit and intelligence by saying the infinitely wise "no shit".

Oh, wait, you posted a link to wikipedia. Perchance, that is the limit of your knowledge on the subject as well then?

Re:Quantum states (1)

QuantumG (50515) | more than 7 years ago | (#19631119)

What if it is? It's more than you bothered to lookup before posting.

Re:Quantum states (4, Informative)

pablob (567257) | more than 7 years ago | (#19630559)

32? They should be 42!

More seriously, a qubit (short for "quantum bit") has two well defined states, 0 and 1 (|0> and |1> for those QM buffs) just as a regular (classical) bit. The difference is that the classical bit has to be in either 0 or 1, while the qubit can be in what is called a "superposition" of those. So you could have a qubit in the 0 state, or in the 1 state, or in the "x% 1 and y% 0" state (where x+y=100). Part of the magic of quantum computers comes from this fact: using the proper operations, you can feed your quantum computer a register which is set up to "all possible inputs" so that it applies the algorithm to all possible values. Some people call this "Massive parallelism". You have to be careful, because Quantum Mechanics does not allow to extract the result of all those calculations (that would be great), so you have to go through some tricks to get useful information out of that "parallel processing".

I hope this helped some!

Pablo B.

Re:Quantum states (1)

asuffield (111848) | more than 7 years ago | (#19630591)

it was my understanding that there are 32 quantum states of electrons


That is almost, but not quite, entirely unrelated to quantum computing (it's got something to do with quantum physics, that's about where the relationship ends - I think it's about the chemistry of atoms). I don't believe it's true except under specific circumstances, anyway.

So, if we now have a quantum NOT gate, doesn't that mean there are 32 possible states of the NOT gate?


Quantum computers operate on qubits, which take on any state along a continuous probability line between 0 and 1 - you can think of it as a floating point number with an arbitrary degree of precision. As such, there are as many possible states as there are real numbers; this quantity is both infinite and uncountable. The actual math operates on the complex plane, but for the purposes of considering the number of possible states, it's equivalent to this description using the real numbers (if I've remembered my transfinite theory correctly - it's been a few years - but it's definitely infinite and uncountable).

Actual quantum computers are likely to have fixed finite limits on their precision, but we're not really far enough along to be sure how that will work out yet. There are a number of competing theories on the subject.

Re:Quantum states (5, Informative)

mindriot (96208) | more than 7 years ago | (#19630737)

Note: I am not a physicist, this is just what I remember from a quantum computing lecture I attended years ago. Of course, rather than believing 100% in what I wrote, you're probably better off double-checking on Wikipedia and Google...

The quantum states you're referring to do have something to do with this. However, their number isn't what's important.
The interesting thing about the quantum computing world is that such states can be in superposition, that is, it is unclear whether or not the state is one or the other. You can only know that if you measure the state, the outcome will be state A in, say, 30% of the time, and state B in 70%. Now, you could probably extend this to 32 different states, but since we're used to bits, we'll build something where we just use two (for instance linearly polarized photons -- 0 degrees = 0, 90 degrees = 1).

Now, there exist methods (or they're being researched) that allow you to put your bit into a superposition of its states. This could, for instance, be so that measuring the state will produce a 0 exactly half the time. Maybe you could put your Schroedinger cat in the box -- dead=0, alive=1...

This by itself is not particularly exciting. But you could do that for multiple bits (say a 32-qubit word) so that measuring it, you get uniform probability to measure any number between 0 and 4294967295. Where it really gets interesting is when you apply quantum operators to your state: They can transform the state without destroying the superposition, i.e. without measuring it. For instance, if your superposition currently gives you 30% chance for measuring a 0 and 70% for a 1, then a CNOT gate would reverse that probability.

Note, however, that a CNOT is a "controlled not": it has two inputs, the control and the target. The control passes through unchanged, but the target is flipped if and only if the control is 1 (i.e. the target output value is identical to the XOR of the input values). In a quantum world, this lets the two bits be entangled: For instance, if the target bit is 1, then the output of the target is 1 iff the control is 0 (target = NOT control). Now suppose that we create a superposition on the control input -- then the control output will be that same superposition, but the target output will be (NOT control) for all control values. In other words, we've just computed a function for all possible input values at once. And you can build these things larger, to do more useful things, such as with a 32-qubit input.

The problem is, you thus get all possible results at the same time, but it's a superposition, and after measuring, you'll only have one result. Why is this useful? Because for one, you can construct some algorithms that transform the problem in such a way as to give a guaranteed result; in other cases, you'll do multiple samples and after a while you'll get your result -- and for some problems, you'll get it orders of magnitude quicker, on average, than on classical computers.

For instance, the Deutsch-Josza algorithm is such an example. Assume I have a function that does one of two things -- it is either constant over the whole input domain, or it is balanced, that is, it returns 0 for exactly half the possible inputs and 1 for the other half. The function, to you, is a black box. How do you determine quickly whether the function is constant or balanced? On a classical computer, you have to test one more than half the inputs, in the worst case, to find out whether the function is balanced or constant. Using the Deutsch-Josza algorithm, you can solve the problem in *constant* time on a quantum computer.

In other words, quantum computing may be interesting for some number-crunching applications. Of course the true capabilities of such a system are not yet completely understood. But I would think that for desktop computing it's probably not too relevant...

Re:Quantum states (1)

CaspianXI (935014) | more than 7 years ago | (#19632157)

I'm an undergraduate physics student assisting a professor in research in quantum mechanics. Although I've only touched quantum computing slightly in my research, I'd like to confirm that much of what mindriot said makes sense according to quantum mechanics as I understand it (I say "as I understand it" because the great Dr. Feynman once said that "no one understands quantum mechanics").

One point I'd like to mention -- the qubit is in both states A and B simultaneously, as far as we can tell. Measuring the qubit to be in state A 30% of the time does not indicate that is actually in that state 30% of the time -- rather, the act of measuring it tends to force it into state A 30% of the time.

Quantum mechanics involves incredibly tiny particles that are easily interrupted because even our most delicate instruments will greatly interrupt the particles we're trying to measure. If this doesn't make sense, imagine that you're sitting in front of a table in a dark room -- there's a marble spinning on the table, and you need to determine which direction it's spinning using a ping-pong ball. You roll the ping-pong ball toward the marble, which diverts the direction of the ping-pong ball, which rolls off and hits a sensor. Now knowing the position of the ping-pong ball, you calculate how its path was diverted by the marble, giving you the direction in which the marble was spinning.

BUT -- the ping-pong ball also hit the marble, which may have pushed the marble off in some direction. This may have made the marble stop spinning altogether, or the marble may now be rolling in some direction instead of spinning.

With qubits, which often assume states "spin up", "spin down", etc., we have the same problem in measuring them. If you don't see the problem in this, imagine if the act of opening a file on your computer altered its contents. Qubits will probably be used more in memory than disk space... but still, if you place a value in memory, you don't want it to change when it comes time to retrieve it.

Quantum computing involves a very delicate balance between how much information we want to "see" in the system and how to leave it alone. Once again, my research is only tangentially related to quantum computing, so I can't answer the question of how they've solved this dilemma (or if it's been solved), but I find it interesting nevertheless.

Re:informative hom0homo (-1, Troll)

Anonymous Coward | more than 7 years ago | (#19631053)

fags should all die.

Cracking (4, Informative)

z-man (103297) | more than 7 years ago | (#19630351)

In theory, quantum computers would allow hackers to crack today's toughest coded messages.

That's an overstatement. A quantum computer will not suddendly magically crack the strongest codes. Yes, certain algorithms designed for quantum computers, like Grover's algorithm, will reduce the time needed to find the key of a symmetrical cipher with about half the number of bits in the key. However, given for example a 256-bit key you would still have ~2^128 keys to check and afaik 2^128 still takes quite sometime to crack....

Re:Cracking (1)

z-man (103297) | more than 7 years ago | (#19630423)

Note, my above post is about symmetric cryptography, not asymmetric cryptography. When it comes to asymmetric (public key) cryptography, which relies on computationally difficult problems such as prime factoring, quantum computing has a lot more potential. Factoring a prime number in O(log n^3) (Shor's algorithm) is an enormous improvement and would be devastating for traditional public key cryptography. Of course, we'd still have quantum cryptography.

Re:Cracking (0)

Anonymous Coward | more than 7 years ago | (#19630823)

Or Lamport signatures [wikipedia.org] , at least for digital signatures. There's a cryptosystem based on the same idea, but I can't find its name.

Re:Cracking (1)

frieko (855745) | more than 7 years ago | (#19630649)

IANACSOQP but aren't all NP-complete problems essentially equivalent at some fundamental level? So if QC can do one NP-complete problem in P time, it can do all NP problems in P time?

Re:Cracking (0)

Anonymous Coward | more than 7 years ago | (#19630745)

All NP-complete problems are equivalent in that a puzzle for one can be transformed into one for the other in polynomial time; that's what makes them NP-complete. So yes, if the QC can do one problem, it can do them all, but it's not likely that the QC can do so. For the QC to solve NP-complete problems, either P==NP or the probabilistic variant (don't remember its name) == BQP, both of which are considered to be unlikely.

There's Grover's algorithm, but that doesn't do you any good if the conditions above are false. If they are, then an NP-complete problem must have a worst case superpolynomial time complexity, and since Grover only reduces it by the square root, it isn't enough to bring NP over to P.

(Also, if someone manages to make a Kieu-type hypercomputer or a "QC" that uses quantum gravity [hpi-web.de] to solve PSPACE-complete problems, then the above is moot; but such a QC wouldn't be the same kind of quantum computer as is described here.)

Re:Cracking (1)

fatphil (181876) | more than 7 years ago | (#19630805)

For values of "essentially equivalent" equal to "can be reformulated into, and the result therefrom converted back, in polynomial time", yes.

Note that factoring is not known to be, and suspected (by gut feel only) to not be, NP complete.

Re:Cracking (0)

Anonymous Coward | more than 7 years ago | (#19630753)

Ok, so we just build a quantum-buster-buster?

Hey, it worked for the trace-buster in The Big Hit ;-)

Quantum gates? (5, Funny)

Mikachu (972457) | more than 7 years ago | (#19630403)

They're opening the quantum gates now? They're insane! Who knows what might pour out of them... I hope they're at least doing it on the moon.

The future of the human race is up to one lone marine now. Thanks a lot, scientists.

Re:Quantum gates? (1, Insightful)

Anonymous Coward | more than 7 years ago | (#19631223)

Wrong, get Gordan Freeman to throw the switch, no amount of Marines are gonna be enough.

In 20 (50?) years... (1, Funny)

Anonymous Coward | more than 7 years ago | (#19630407)

Dude you're getting a Delft!

I fear for the programmer's sanity (1)

Keichann (888574) | more than 7 years ago | (#19630413)

Can anyone remember the name of that assembler that only had the 'not' operator? Maybe it's time for a port :)

Re:I fear for the programmer's sanity (1)

fatphil (181876) | more than 7 years ago | (#19630795)

There was the OISC, is that what you're thinking of?
I think its only operation was something like subtract A from B, and if negative jump to C.
However, google will probably do better than my addled memory.

Six of one, half a dozen of the other? (0, Redundant)

nick_davison (217681) | more than 7 years ago | (#19630445)

...the first Controlled-NOT quantum gate... a quantum computer would achieve new levels of power by turning bits into fuzzy quantum things called qubits (pronounced cue-bits) that are 0 and 1 simultaneously.
0 and 1 simultaneously, through a NOT gate... becomes 1 and 0 simultaneously? Sounds useful. ;)

A long way off yet! (1)

Chemisor (97276) | more than 7 years ago | (#19630455)

> the result may pave the way for devices of double the size in the next year or two

Well, at the current rate of progress, we might see a Quantum Pentium III in about 26-52 years, depending on whether its "next year" or "two". I might be dead of old age by then.

Re:A long way off yet! (0)

Anonymous Coward | more than 7 years ago | (#19630547)

Who's to say the number of qubits on a chip doesn't *square* every 18 months?

Re:A long way off yet! (1)

poopdeville (841677) | more than 7 years ago | (#19630719)

Let's hope not. We're still stuck at 1.

Hans Mooij of the Delft University of Technology? (1, Redundant)

jddj (1085169) | more than 7 years ago | (#19630509)

"Mooij you're gettin' a Delft!"

oblig.. (0, Redundant)

hldn (1085833) | more than 7 years ago | (#19630513)

but does it run linux?

GateChip One (0)

Anonymous Coward | more than 7 years ago | (#19630551)

It's a gate, on a chip.

Inversed qubit? (2, Interesting)

Lobais (743851) | more than 7 years ago | (#19630611)

If a qubit is both 0 and 1 at the same time, what is the point of inversing it? Would it then be 1 and 0 at the same time?

Mod parent up. (0)

Anonymous Coward | more than 7 years ago | (#19630685)

I don't know if it is insightful or funny, but I hope they mod you up.

---

My image-word for this post is: redneck.

Re:Inversed qubit? (0)

Anonymous Coward | more than 7 years ago | (#19630931)

Maybe we'll be able to settle the 2+2=5 thing after all!

Re:Inversed qubit? (0)

Anonymous Coward | more than 7 years ago | (#19631087)

Some qubits are more 0 than 1. The inversion then turns them into qubits that are more 1 than 0 (by the same probability).

Very roughly speaking.

Re:Inversed qubit? (2, Informative)

bh_doc (930270) | more than 7 years ago | (#19632021)

Absolutely correct. The way it's written in the text is a bit misleading. A qubit can be 0, 1, or some proportion of 0 and 1. It's like a continuum between 0 and 1, and a qubit can take any place along that continuum.

The tricky thing is, quantum mechanics won't let you measure anything other one of two values (not strictly true, but go with me on this). But those values don't necessarily have to be 0 and 1, just so long as they are orthogonal to each other.

Okay, to understand that last bit you need to bring in the fact that the proportions of 0 and 1, x and y say, can actually be any complex number, positive, negative, imaginary, whatever, just so long as their magnitudes-squared add to 1. So, then, instead of measuring for the 0 and 1 states, you could measure for "0.5" and "-0.5" states.

This isn't even getting to the confusion of putting these things into gates.

Qubits? (0)

Anonymous Coward | more than 7 years ago | (#19630623)

I've never really understand why they started calling them qubits.

A 'bit' is simply shorthand for "binary digit". Quantum digits, however, aren't binary, since they can represent much more than a simple 0 or 1. By adding the 'Qu-' to the term, we are essentially calling them "Quantum Binary Digits," which is in itself an oxymoron.

I understand that it's just supposed to be a nickname, but I think if we have the power to make up words to represent these "fuzzy" quantum digits, that "quigit" (like 'widget' with a K tacked on front) would be both more accurate and more fun to say.

Re:Qubits? (1)

gardyloo (512791) | more than 7 years ago | (#19630689)

I think if we have the power to make up words to represent these "fuzzy" quantum digits, that "quigit" (like 'widget' with a K tacked on front) would be both more accurate and more fun to say.
Perhaps. But then all the Gnome people would insist on calling them "gwidgets" and the only real winners would be the xfce users....

Does this make D-Wave's quantum computer obsolete? (1, Interesting)

Anonymous Coward | more than 7 years ago | (#19630647)

Recently, D-Wave's founder Gordie Rose was asked in his blog [wordpress.com]


(comment 2):"How come Delft U has been able to perform a CNOT with two qubits using superconducting technology? I thought Rose/D-wave claimed it was extremely difficult to do discrete quantum gates with superconducting technology. What are the present & future limitations of the Delft "quantum computer?"


Rose IGNORED the question. The quantum computer built by D-Wave [wikipedia.org] is an adiabatic computer (which is an analog computer), whereas the Delft people have built a discrete gate quantum computer. Does the Delft computer make D-Wave's computer obsolete?

while this may all be true (-1, Flamebait)

Anonymous Coward | more than 7 years ago | (#19630815)

muslims have burned and vandalized over a fucking drawing. they have called for and have indeed murdered people for speaking out against their so called god and so called profit. how can people let this type of violence continue with no ramifications?
 
FUCK MUSLIMS!

Argh (1)

Mazin07 (999269) | more than 7 years ago | (#19630841)

Dangit, and I'm having enough trouble in computer science as it is, without all this fuzzy simultaneous 0/1 nonsense.

NOT Moore's Law (3, Funny)

VisceralLogic (911294) | more than 7 years ago | (#19630883)

the result may pave the way for devices of double the size in the next year or two

Hmm, seems like they've successfully performed a NOT on Moore's law.

For what purpose? (1)

arrenlex (994824) | more than 7 years ago | (#19630949)

Transistors are capable of more than on and off -- they can handle many intermediate stages of charge (fundamentally low, medium, high), which computing applications do not currently exploit. Why not add a third state by using technology that already exists? What are the benefits of quantum computing over the existing versatility of transistors?

Re:For what purpose? (1)

edschurr (999028) | more than 7 years ago | (#19631515)

My very incomplete understanding is that the quantum computer doesn't actually use three states, but rather both states simultaneously. I.e. the qubit is in a superposition of both 'on' and 'off'; it's not one of 'on', 'off', or 'other'. The difference matters. The quantum computer can be thought of as computing all possible inputs at the same time, but when the output is measured you get only one of the results, at random I guess. Apparently this can be taken advantage of in a few ways. Basically, some problems can be computed in a better time than the classical computer could do it.

I reiterate that my grasp of this is weak, and especially where terminology is concerned.

Sounds familiar... (1)

guruevi (827432) | more than 7 years ago | (#19631039)

by turning bits into fuzzy quantum things called qubits (pronounced cue-bits) that are 0 and 1 simultaneously

Sounds like any ol' woman to me, nothing to worry about, we have been handling it for centuries.

Maybe not so great... (1)

FridayBob (619244) | more than 7 years ago | (#19631181)

... In theory, quantum computers would allow hackers to crack today's toughest coded messages ...
That may sound pleasing, but how about if it were changed to this:

... In theory, quantum computers would allow governments to crack today's toughest coded messages...
It's not like we're going to be the first ones to get our hands on these devices, you know (if they ever allow us to have them).

Wow. ? (3, Funny)

DoofusOfDeath (636671) | more than 7 years ago | (#19631285)

This is awesome no it's not!

A little insight (1)

Frion (942886) | more than 7 years ago | (#19631431)

Having taken a class on quantum computing last semester I would really like to add in some facts here. First to say qbits are both 1 and 0 at the same time is not entirely clear. Qbits are represented by arrays of length 2. These can be either [1,0] or [0,1]. Where the confusion occurs is when these are a superposition of the two, which means basically means that there is a probability that the result would be one of the two. What results from this is knowing the result most of the times, but sometimes being uncertain. For the uncertain cases there are ways to use the probabilities where in almost all cases only the more probable case will result. Also it is not completely correct to say we have no idea of how these will work. We have a pretty damn good idea, we just have not tested it yet. In fact, most of quantum computing is just simple linear algebra, as the qbits can be represented as arrays and the gates that control them can simply be represented by 2 by 2 matrices. Obviously this is only the basics of it, not touching on entanglement or any algorithms(which can all be represented by multiplying matrices). Anyway I did a pretty bad job of explaining all that, but the point is that this is a big deal and we are way ahead of understanding how these things should work in the future over understanding how to make a machine that will make them work.

perl -qm (1)

Doc Ruby (173196) | more than 7 years ago | (#19631701)

I want to log into that machine and run some quantum Perl scripts on it [cpan.org] . Nothing like an existing library of code to kickstart a new architecture.

Quantum Not? (1)

zeketp (888795) | more than 7 years ago | (#19632265)

How does a quantum NOT gate work exactly? Normal NOT gates make zeroes into ones and visa versa. So this makes something from a simultaneous one and zero into a simultaneous zero and one?! How does this even help perform calculations? How do you use that info? *Not zero or one, but both? WTH?* Sorry if this is an obvious question, Discrete mathematics isn't in the curriculum for aerospace engineering. Does this do tons of simple calculations very fast? If not, I don't see as much of an application in it. Maybe if it could do some higher level calculus *in hardware* then it has real value.
Load More Comments
Slashdot Login

Need an Account?

Forgot your password?

Submission Text Formatting Tips

We support a small subset of HTML, namely these tags:

  • b
  • i
  • p
  • br
  • a
  • ol
  • ul
  • li
  • dl
  • dt
  • dd
  • em
  • strong
  • tt
  • blockquote
  • div
  • quote
  • ecode

"ecode" can be used for code snippets, for example:

<ecode>    while(1) { do_something(); } </ecode>