Big Win for the Losers at D-Wave
Does D-Wave's first big sale disprove the quantum computing naysayers?
Hi, this is Steven Cherry for
’s “Techwise Conversations.”
Our January issue is different from the rest of the year. We usually do a special report called “Winners & Losers”—five technology projects we think show remarkable promise and five we think just the opposite.
“D-Wave Systems, a Canadian start-up, recently booted up a custom-built, multimillion-dollar, liquid-helium-cooled beast of a computer that it says runs on quantum mechanics.
That's right. D-Wave, a 55-person company operating out of an office park in Burnaby, British Columbia, claims to have built that almost mythical machine, that holy grail of computing, the stuff of sci-fi novels and technothrillers—the quantum computer. Such a system would exploit the bizarre physics that apply on ridiculously small scales to compute ridiculously fast, solving problems that could stymie today’s supercomputers for the lifetime of the universe.
Now, building a practical quantum computer has proved hard. Really hard. Despite efforts by some of the world’s top physicists and engineers and the likes of IBM, HP, and NEC, progress has been slow. Ask the experts and they'll tell you these systems are a decade—or five—away.”
Harsh words, and you can almost hear the gods of karma preparing their punishment. Sure enough, a couple of weeks ago, aerospace giant Lockheed Martin, maker of the F-22 Raptor fighter jet and the Mars Odyssey orbiter, bought a quantum computer from D-Wave for $10 million. So were we wrong to call D-Wave a loser? Are quantum computers really ready to make the leap from the laboratory to the commercial market? What is Lockheed Martin getting for its money that it can’t get from a regular computer?
My guest today to answer these questions, as well as, presumably, to gloat a little, is Geordie Rose, founder and CTO of D-Wave. The company is still headquartered in Burnaby, British Columbia, but he joins us by phone today from Banff in neighboring Alberta.
Geordie, welcome to the podcast.
Geordie Rose: Thank you, Steven
Steven Cherry: Let’s start by talking about quantum computers. Basically, instead of a binary 1 versus 0 choice, quantum computers have additional states?
Geordie Rose: That’s right. So the fundamental unit that stores information in a conventional computer is a bit, a 0 or a 1, and physically that bit is represented by some physical thing in your chip. It can be a voltage, which is either below some value or above some value. It can be a current, again, which is magnitude, say, one direction or another. It can be the direction of a magnetic thing, like a little magnetic grain in your hard drive. Ultimately those abstract symbols that we manipulate when we build algorithms and code are embodied physically in the chip. Now in quantum computing the bits are a generalization of that concept to something which are generally referred to as quantum bits or qubits where the device upon which the information is stored doesn’t behave classically. So things like this allow you to do problem-solving procedures that are much more efficient. You can do many, many more things in parallel, and while the gains aren’t automatic, they certainly can—when you wrap up all of the details, they can be quite significant.
Steven Cherry: So what problems are Lockheed Martin trying to solve here? What are they getting for their money that they can’t get from a regular computer?
Geordie Rose: What our system does is solve optimization problems and a generalization of that which are called sampling problems, but I won’t go into that. So an optimization problem is very simply what you would think it is. Imagine you have to find the biggest, or the smallest, or the least, or the largest of something, which could be very complicated. For example, you might want to know what is the biggest financial return that I could get on any set of 100 financial instruments, and I have the freedom to choose any 100 I like, and maybe the freedom to chose the timescale and all this. So you could imagine that if I had to choose 100 financial instruments, there are lots of different combinatorial possibilities, many, many more than I could ever enumerate. So what these optimization problems end up being is both foundational to many of the problems that current computers are bad at. So what Lockheed wants to do with this particular system is see if we can recast a particular type of problem which is called “verification and validation” in optimization language and then use our system to then make inroads into this problem. So verification and validation is, given a big complex system, of which there are many nowadays and the aerospace industry is rife with them, large machines that have many different parts, lots of software of a variety of different types from different vendors, lots of different very complex hardware pieces. They all have to work together and they can’t fail for unintended reasons. They can’t have behavior that wasn’t in the spec.
Steven Cherry: How much customization is involved with a customer like Lockheed Martin? I mean I’m just wondering if anybody with an optimization problem and an extra $10 million lying around should just buy a D-Wave computer or is there more involved than that?
Geordie Rose: There’s quite a bit more involved. We partner with that organization, supply them with all our technical know-how on the hardware side including machines, and then work together with them on the software side to marry their problem into algorithms and procedures that will run efficiently on this new type of machine. So this is typically viewed as a multiyear procedure where it’s a learning exercise for a lot of it, where we teach the organization how to do this. We help them along the way, we scratch our heads and think about their hard problems, and then the intent is that at the end of the relationship we’ll have demonstrated significant progress into this very hard problem, which will then, you know, cascade onwards and will eventually hopefully solve the problem or at least part of the problem for the customer organization.
Steven Cherry: That actually leads directly to my next question, which is, When will we know if Lockheed Martin made a wise purchase here?
Geordie Rose: Well, as the CTO of D-Wave, I would say that they’ve made a wise purchase regardless of the outcome because this new type of technology is like an entirely new branch of technology. So there are things about what we do that are really quite different than the approaches that people take nowadays. And it’s showing a lot of promise, so even the results that we’ve gotten now on these early stage chips, they’re dramatically different than conventional processors. But just as an example, the chips are superconducting, which means they’re made out of metal and we operate them at very low temperatures, and at those temperatures these metals don’t generate heat when they act—when the processors are operating. So the power profile of these things is ridiculously low. In fact the energy consumption of one of these chips compared to a CMOS chip is something like 10 to the 10 times less. So we’re talking about, like, ten orders of magnitude less energy generated during a computation. So I think that they made a wise decision to look at it. Whether or not it actually pays off in the problem that they’re trying to solve, I’d say we’re looking at something between, say, three to seven years. Because the technology now is not mature enough to actually directly target the core hard problem that they’ve got. Their problem is so big that you need to have much larger scale systems than what we’ve built today, but you know we have a road map and we’ll get there. So I think as the technology matures together with the understanding of how to solve the problem from the software algorithmic side, you’re going to see a win or not somewhere in the, you know, three- to seven-year time frame I would think.
Steven Cherry: Well, fair enough. Well, congratulations on the sale and thanks for not rubbing our noses in our loser label.
Geordie Rose: You’re welcome.
Steven Cherry: We’ve been speaking with D-Wave founder and CTO Geordie Rose about whether the future of quantum computing has finally arrived.
For IEEE Spectrum’s “Techwise Conversations,” I’m Steven Cherry.
This interview was recorded 8 June 2011.
Segment producer: Ariel Bleicher; audio engineer: Francesco Ferorelli
Follow us on Twitter @spectrumpodcast
NOTE: Transcripts are created for the convenience of our readers and listeners and may not perfectly match their associated interviews and narratives. The authoritative record of IEEE Spectrum's audio programming is the audio version.