PROTOTYPE COMMERCIAL QUANTUM COMPUTER DEMO'ED

A Vancouver startup, D-Wave Systems, claims to have demonstrated the world's first commercially viable quantum computer. They were quite confident of this, considering that their press release trumpeted a success at least ten minutes before the demonstration was scheduled to begin at the Computer History Museum in Mountain View, Calif.

They've certainly reason to be confident. Even though a quantum computer is commonly considered a research affair that's as much as 20 years away from usefulness, D-Wave not only formed a company to develop a unit but also got venture capitalists to fund it. They picked up US $14 million last May.

Today, the company was supposed to show off a quantum computer sporting 16-qubits, the most of any quantum computer, commercial or otherwise, but still way too few to do anything important. What's a qubit? Qubits, or quantum bits, are what make quantum computers different from their digital ancestors. A digital bit can be either a one or a zero but not both at the same time. A qubit can. And that lets it do many calculations at once. So quantum computers should be capable of solving certain horrendous problems faster than conventional computers. Certain types of searches, the "traveling salesman" problem, and finding the factors of large integers fall into this category.

At it's heart the D-Wave computer, called Orion, is a chip of niobium that's been cooled to near absolute zero. It relies on a dark-horse technology known as adiabatic quantum computing. It and D-Wave have many critics.

The computer solves only one type of problem, which mathematicians call a two-dimensional Isling model in a magnetic field, but through some software trickery, other problems can be recast as this problem. At the demonstration, they planned to show off its flexibility with two programs. First, they were to show how Orion runs a pattern-matching application that searches a database of molecules. The second could be called the wedding planner's dilemma, in that it is designed to figure out the best seating arrangement for a group of people according to certain constraints, such as Uncle Sam can't sit next to Aunt Jean but has to be at the same table as Grandpa Harry.

In the second half of this year, according to D-Wave founder Geordie Rose, the company will give free access to an Orion to let people develop applications for it.

Spectrum solicited some feedback on the Orion's development from Lieven Vandersypen, an associate professor at Delft University. Vandersypen's group is working on an alternate quantum computing scheme, and he was part of the team at Isaac Chuang's lab at MIT that did the first demonstration of a quantum computing plan that, if scaled up, could be used to defeat common encryption algorithms.

Here's what Vandersypen thought:

"First, it's quite remarkable that they have persuaded investors to put serious money in their enterprise at such an early stage. It sounds like they have a clear vision of where quantum computing is going, and how to approach it. Whether it is realistic, time will tell.

"Until now, D-Wave hasn't published any major advances or breakthroughs in the scientific literature. With respect to their announcement, there is little detailed information available to support, and thus judge, the validity of the claims (as would be the case in a scientific publication).

"From what I was able to find on the Web about the hardware, it looks impressive to put 16 of these superconducting devices together and wire them up in a special fridge. Still, the level of control appears to be very minimal at the moment—the problems it can solve seem to be hardwired. Of course, one has to start somewhere.

"The 'software' approach is somewhat unconventional (called adiabatic quantum computing). The current understanding is that in principle it can be mapped one-to-one onto the conventional quantum circuit model. For NP-complete problems [Ed.: a complexity theory reference to "non-deterministic polynomial time" complete equations], this means that a quadratic speed-up is possible, but not more. There is, of course, no magic needed to solve a small-scale NP-complete problem (although quantum magic may have been used in their demo—it is not fully clear to me from what was made public). As they say, the proof of the pudding will be the solution of a large-scale problem in a time faster than conventional computers.

"In any case, I'm intrigued by their announcement and curious to see whether something serious will come out. So I guess the bottom line is 'we'll see'."

For today, "we'll see" will have to suffice as the last word on this topic.

Related Stories

Tech Talk

IEEE Spectrum’s general technology blog, featuring news, analysis, and opinions about engineering, consumer electronics, and technology and society, from the editorial staff and freelance contributors.

Newsletter Sign Up

Sign up for the Tech Alert newsletter and receive ground-breaking technology and science news from IEEE Spectrum every Thursday.

Advertisement
Advertisement