D-Wave Systems Inc. has finally announced the date of its first public demo! :D  On February 13th, 2007, in Mountain View, California, they will demonstrate the world's first 16-qubit quantum computer, which is also the world's first quantum computer that can do something useful.  Then, on the 15th, they will demonstrate it again in Vancouver, BC.

For those of you who would like to see the demo in person, (and can make it to one of those two locations, that is), you can register on the official D-Wave website.  They'll probably be posting video of the demo online after it happens too.

Also, as the CTO says in his post:
One very cool thing that we?re planning to do in Q2/2007 is to provide free access to one of these systems to people who want to either develop or port applications to it?so if you have an idea for an app that needs a fast NP-complete problem solver, start thinking about what you could do with some serious horsepower.


On a related note, if any of you are up for an optimization showdown, start looking into the problems known as Maximum Independent Set (which is identical to Maximum Clique and Minimum Vertex Cover), Integer Programming, and the many variations of Boolean Satisfiability.  I'll take on anyone up for Max Independent Set. ;)  Most such algorithms can be easily adapted to make use of an oracle like a quantum computer, and it'll be a while before huge problems can be entirely solved on the quantum computers, so it's still quite useful to them.
Posted on 2007-01-21 02:03:46 by hackulous
wow... too bad i live half globe from there.

i hope there will be some demo and texts online.

Looking forward to some "assembly coding" for quantum computers. I wonder what will it's architecture look like.
Posted on 2007-01-21 04:40:59 by vid
Isn't it more like a "quantum computing device" than a "quantum computer"?
Posted on 2007-01-21 04:59:07 by f0dder

Looking forward to some "assembly coding" for quantum computers. I wonder what will it's architecture look like.

vid, I also had that wrong impression that you now have - that QCs could be used in a way similar to how PCs are. The qubits, as I understand them, are like a matrix, where you feed values, and get the result as a matrix: of probabilities of all results. It solves problems, where the least movement/energy is required for the solution. In a nutshell, a traveling salesman's problem to decide on the best route, is one type of problems, for instance .
Though: D-Wave's QC takes ~$2000/day to work - it's for the liquid nitrogen, used to cool the chip down to less than 1 Kelvin, to make some of its parts super-conductors (and thus make it easy to use the quantum effects). So, we won't be seeing QCs from D-Wave at home, but it's great that they're going to lend it for public-use :) .

hackulous, is your app to compare two molecules one of the two demo-apps to be run during the show  :) ? I'm quite curious on how the Java/C code to define and sequence a problem, and then get the result looks like :)
Posted on 2007-01-21 06:42:18 by Ultrano

The qubits, as I understand them, are like a matrix, where you feed values, and get the result as a matrix: of probabilities of all results.

Yes and no.  You feed it a matrix of sorts (possibly a hypermatrix in the future), and get back a single answer vector.  You can run it multiple times to get probabilities.  They're not counting the matrix in the number of qubits, just the answer size.  It would be misleading to count it as 58 qubits with 42 being couplers and 16 being state (even though they're effectively the same thing in a different shape).  Then again, the input on each also has more precision than just 1/0, and previous examples of quantum computers haven't been so humble.  The "7-qubit" paperweight to factor 15 into 3 and 5 has a 4 important bits of input and only 2 important bits of output. :lol:
It solves problems, where the least movement/energy is required for the solution. In a nutshell, a traveling salesman's problem to decide on the best route, is one type of problems, for instance .

True.  The problem that happens to be easiest to map onto the hardware at the moment is Max Independent Set, which is the problem of finding the largest set of nodes in a graph such that none of those nodes have an edge between them.

Though: D-Wave's QC takes ~$2000/day to work - it's for the liquid nitrogen, used to cool the chip down to less than 1 Kelvin, to make some of its parts super-conductors (and thus make it easy to use the quantum effects). So, we won't be seeing QCs from D-Wave at home, but it's great that they're going to lend it for public-use :) .

Also yes and no.  Thankfully, refrigeration technology is advanced enough these days that liquid nitrogen (and getting much colder than that) isn't too expensive.  Geordie once said in some lecture he gave that you could probably run it on an average 110V electrical socket if you wanted.  I can't recall if they use nitrogen, but they probably use liquid helium at some point, which although a bit more expensive, isn't a ton more expensive, and I'm pretty sure that it gets reused (re-refridgerated).  That said, the system itself will likely be way too expensive for a home consumer for a while.  Big companies, however, might want their own and could probably easily afford them.

hackulous, is your app to compare two molecules one of the two demo-apps to be run during the show  :) ? I'm quite curious on how the Java/C code to define and sequence a problem, and then get the result looks like :)

It is. :D  I still can't believe that I got to write it.  Of course, I had a lot of input on what people would like in it after the initial draft, some of which made it tons better, and they've undoubtedly been doing tweaks in the month and a half since I was there.  In many ways, it's a lot simpler than you'd think.  Unfortunately, I can't say how it works in much detail.

Also, I don't want to take much credit, because it's the hardware teams who've really made history.  They worked tons of overtime to get it done too.  Plus, much of the crazy math, physics, and engineering that they had to go through at every step is so advanced that even if I dropped everything else right now and tried to learn it, I probably couldn't.  I once looked at a whiteboard and honestly thought it was Greek before realising that it was math.  (They do use Greek for the mask names, as you can see on the image Geordie posted.)

In response to f0dder's question, there are a lot of fatal problems that D-Wave has found over the years with what people previously considered a "quantum computer", and even though it's a relatively new field, some of the misconceptions on what might work are hard to shake.  If you want a computer like your average desktop that uses nothing but quantum processing, not only would it be a huge waste of money (because much of what your computer does wouldn't benefit), but you'd also have to wait a century or so before you could get the 1GB worth of qubits (effectively, the RAM) that you'd want to run your programs. ;)  A lot of otherwise respectable theoreticians had implied that that such a thing would make sense, but considering that we're at 16 qubits and not 8,589,934,592 qubits, they may have to reconsider their ideas.  It even turned out that the popular "gate model" approach isn't scalable in practice, so it's hard for them to get above 7 qubits.  (The previous record is apparently 12 qubits, but who knows whether it actually works or not.)
Posted on 2007-01-21 13:06:55 by hackulous
some chances of quantum "extension procesors"? Something like FPU, MMX, etc. but it will perform computations it is useful for.

I am still pretty confused about how exactly it works, even though i studied some material on it
Posted on 2007-01-21 14:18:58 by vid
vid: I doubt that will happen the next few years - from my (very vague :)) understanding it's more like a big external "quantum processing device" that you interface with...
Posted on 2007-01-21 14:28:06 by f0dder

some chances of quantum "extension procesors"? Something like FPU, MMX, etc. but it will perform computations it is useful for.

Yup, that's one of the possibilities.  Another is to have it as an external (or eventually internal) device plugged-in as f0dder suggested.  Another is to have it as a standalone supercomputer to be used for hard problems.  The external device and the standalone supercomputer are probably nearest in the future because of it being out of the price range of consumers.  The cost of production should go down a lot with automation of course, but I have absolutely no idea how much the stuff costs in the first place.  There are other concerns, such as size, but they can probably be overcome eventually too.  They've already got a lot of plans for huge improvements in just about every facet of the system and production thereof, much of which should be in place for the next demo, whenever it is.  At least half of them I don't know about or couldn't possibly understand (like the XZ coupling Geordie mentions to allow it to simulate other quantum systems).

It would be awesome to have a Quantum Processing Unit in an average computer, though  8)  You could probably overclock the CPU too with the super-cooling needed :lol:
Posted on 2007-01-21 15:02:47 by hackulous