I'm surprised that a mint copy of the original was so hard to find. I would have thought all kinds of schools as well as tech companies would keep collections of historical industry mags like "Electronics".

Where's JoeP? New South Wales has made a single atom transistor. I heard an interview with one of the researchers on NPR, and this may be scalable. She said that only a few hundred of these transistors would be needed for computing, but I'm not sure why that would be.

Science Daily - In a remarkable feat of micro-engineering, an international team of researchers, including physicists at the University of New South Wales in Australia, have created a working transistor consisting of a single atom placed precisely in a silicon crystal.

She mentioned that the transistor would switch between energy states rather than voltage levels to represent digital values. Here is the link to the NPR interview, it's p. interesting. Reaching For The Limits of Tiny Transistors : NPR

Quote:

I guess in the world of quantum computing, you are actually using a completely different way of doing the computing. So you're using what's called parallelism, quantum parallelism. So instead of doing calculations in the classical world one after the other, albeit very fast, you actually do these calculations in parallel.

And there are certain kinds of calculations where if you can do things in parallel, you can get a predicted exponential speed-up in the computation. And I guess that's one of the reasons why quantum computation is quite an exciting field at the moment. If you can actually realize them then there are certain tasks that you can do which would be much faster than you can do with classical computers.

ES?

__________________
Chained out, like a sitting duck just waiting for the fall _Cage the Elephant

Porn has been a driving force not in technological advancement per se, but in consumer adoption of technological advances.

Back in the day, mostly only perverts owned their own projectors, then their own VCRs; and in early Usenet, the binary newsgroups were largely porn (there was a very misleading and very widely quoted study at the time that conflated bandwidth vs. content to create the impression that the internet was like 95% porn or something, in fact), and really, porn has always been a major factor in increasing consumer demand for faster download speeds and greater storage capacity.

Check you out, Doctor X, accidentally touching on a salient point!

I'm surprised that a mint copy of the original was so hard to find. I would have thought all kinds of schools as well as tech companies would keep collections of historical industry mags like "Electronics".

The 'Electronics' journals were the first to be translated to digital format. Once that was accomplished, print journals were discarded wholesale. We're right in the midst of the same process with medical science and service journals.

I guess in the world of quantum computing, you are actually using a completely different way of doing the computing. So you're using what's called parallelism, quantum parallelism. So instead of doing calculations in the classical world one after the other, albeit very fast, you actually do these calculations in parallel.

I think she may be simplifying it a bit, as we already have true parallelism outside of quantum computing. It has existed for decades and is happening in your computer right now if you have a dual-(or more)-core processor. It's happening on an even greater scale in your graphics card as it calculates which color to paint each pixel on your screen at the exact same time. So, while parallelism is definitely cool, that's not what makes quantum computing so powerful. Regular parallelism can't give us the kind of exponential speed-up necessary to solve the types of problems she's talking about here:

Quote:

Quote:

And there are certain kinds of calculations where if you can do things in parallel, you can get a predicted exponential speed-up in the computation. And I guess that's one of the reasons why quantum computation is quite an exciting field at the moment. If you can actually realize them then there are certain tasks that you can do which would be much faster than you can do with classical computers.

So, for example, factoring numbers belongs to a class problems that are called NP-complete, which means that they can't be solved in polynomial time. The larger your problem gets, the amount of time it takes to solve the problem grows exponentially. But! With NP-complete problems, if you already have a proposed answer, you can verify it in polynomial time. So if I ask you to factor 12,118,699, it would take a very long time. But if I told you that the factorization was 3557*3407, you could check my work lickety split.

Last I heard (late last year) they had this one quantum computer factor the number 15. Big whoop, right? I can do that in my head. But seriously, factoring numbers is a very hard problem for traditional computers to do, but it has cool applications in spying and war-making, so the nerds make their prototypes do it as a parlor trick to get governments to throw money at them and it is very effective.

So with quantum computing, you have this black box (I'm a computer scientist, not a physicist, so I get to believe in magic and let them do the work, but I assume this is where the quantum parallelism and exponential speed-up happens) where I give it my problem, and it gives me an answer in polynomial time, but there's a possibility it won't be right. But suppose I know it has a 10% chance of being right. I just run my program 10 times and check each answer. That still takes loads less time than solving the problem straight out.

That's as much as I can remember from a 3-hour workshop I took last year. I'm taking the full course this fall, so hopefully I'll be more useful after that.

This might come closer to answering your question than anything I said:

Quote:

Information is measured in ‘bits’, and a bit may have two positions (described typically as 0 or 1). Quantum computers however don’t use these bits, and instead they use quantum bits, or ‘qubits’. But while a bit must be a 0 or a 1, a qubit can be both 0, 1, or a superposition of both. This difference might seem small and subtle, but in fact, it is absolutely humongous: a mere hundred qubits can store more classical ‘bit’ information than there are atoms in the Universe.

As to the how? Still no idea.

Actually, wait. Now that I think about it, I have a vague idea. Stand by.

So you have these qubits and they are "entangled" with each other. The presentation I saw was on the CS theory of it, not the physics, so I don't know how exactly it relates to quantum entanglement, but I imagine they used the same word because the one has something to do with the other. Anyway, these qubits are represented as nodes on a graph (hypercube, IIRC), and they are entangled if they share an edge. And the state of one qubit depends on the states of all the others it's adjacent to. And then there's a lot of math, like what happens to the whole graph if this one node changes state? So I guess you can represent a whole lot of information that way, a lot more than just 2^n (or 3^n if you're counting superposition as a state).

__________________

Last edited by Ensign Steve; 02-28-2012 at 11:14 PM.

It's a bit like the double-slit experiment where you fire one particle through the slits. The probability of where it lands implies that it has to 'sniff out' both possible routes through the slits - though it can only actually pass through one of them. This can be extended to 3, 4, ... n slits - the single particle has to somehow 'traverse' all the possible routes so that it knows the probabilities that it will land at all the different places on the screen (or detector). It can only traverse one route, classically, but it behaves as though it's aware of all the other possibilities.

This doesn't make any sense from a classical point of view. One way of understanding what happens is to imagine that there are lots of different parallel universes and that in each universe the particle takes a different route. The final position where the particle lands in our universe then implies that it somehow has knowledge (up until we measure where it is) of what is going on with its brother particles in all the other universes.

The quantum computer uses the same idea - we encode the problem in one or more quantum bits - these bits could be, for example, the polarization or spin of some particles. Then the particles do their weird quantum stuff inside the computer, and when we measure the result we see what happened to them. The fact that the quantum computer is many times faster than a classical one can then be explained by the many copies of the quantum computer existing in all the parallel universes. Say there are 1000 parallel universes - that's like having 1000 classical computers all working together on the same problem at the same time.

Say there are 1000 parallel universes - that's like having 1000 classical computers all working together on the same problem at the same time.

Buuuuuuuuuut . . . does not every time a particle have a choice, the universe splits into two parallel universes, so the more computing, the more choices, the more and MOAR UNIVERSES UNTIL WE CAN NO LONGER CONTAIN THEM ALL AS "THE KIDS THESE DAYS" POAST MOAR AND MOAR ON FORA WE DO NOT LINKE TO PLAY THEIR ONLINE YATZE GAMES ANDSURF P0RN UNTIL IT ALL COLLAPSES INTO A TINY QUANTUM SINGULARITY OF COMPRESSED LOLCATS, MEMES VIAGRA ADDs mickthinks's uvula and, of course, porn.

I guess in the world of quantum computing, you are actually using a completely different way of doing the computing. So you're using what's called parallelism, quantum parallelism. So instead of doing calculations in the classical world one after the other, albeit very fast, you actually do these calculations in parallel.

I think she may be simplifying it a bit, as we already have true parallelism outside of quantum computing.

I think she was just correcting herself.

Quote:

It has existed for decades and is happening in your computer right now if you have a dual-(or more)-core processor.

and even if you don't.

Quote:

It's happening on an even greater scale in your graphics card as it calculates which color to paint each pixel on your screen at the exact same time. So, while parallelism is definitely cool, that's not what makes quantum computing so powerful. Regular parallelism can't give us the kind of exponential speed-up necessary to solve the types of problems she's talking about here:

Quote:

Quote:

And there are certain kinds of calculations where if you can do things in parallel, you can get a predicted exponential speed-up in the computation. And I guess that's one of the reasons why quantum computation is quite an exciting field at the moment. If you can actually realize them then there are certain tasks that you can do which would be much faster than you can do with classical computers.

So, for example, factoring numbers belongs to a class problems that are called NP-complete, which means that they can't be solved in polynomial time.

There is no proof or disproof whether they are solvable in polynomial time on this planet that I know about. That's all. And if there is a polynomial time algorithm to solve one of these exactly, it will work for all of the others. That's why a lot of people think that they are "probably not" in P.

Had a discussion with a friend [Imaginary.--Ed.] which led to an inspection of some discussions on religion on other sites, the levels of which reminded me of the genesis of My Personal Maternal Meme Symbol: the tendency of the Willfully Ignorant but Perpetually Offended Zealots to descend to "YOUR MOM!" slurs. The substance of such I will not repeat. Right up there with "ur a fag" with various suggestions of all sorts of homophobic fantasy. I write "fantasy," since I have to wonder about the minds that dream up such activities. They spent time on it, time they did not spend learning things like SCIENCE.

It stands as one of the supreme Last Resorts of the Weak Mind--next to, of course, "Ur a Fag HA!" and appeals to Hitler.

I think about these things a lot on the bus. . . .