Robert “Bo” Ewald
President, Chief Revenue Office,
Few names are as recognizable in high performance computing industry circles as Bo Ewald, a man who has become famous for his work in supercomputing. Whatever ones perspective is, without a doubt Ewald has been worth watching and is now the president of one of the more fascinating companies emerging on the scene: D-Wave, “The Quantum Computing Company.” HPCwire managed to pin Ewald down for a few questions on quantum computing, trends he thinks are important, and the year ahead.
HPCwire: Bo, you’ve been involved with high performance computing for a long time, and are now heading up D-Wave in the U.S. – a computing company dedicated to a whole new kind of computing, quantum computing. Can you tell us about the company and the challenges you face as you work to move D-Wave into prominence?
Bo Ewald: D-Wave is the first company with a commercial quantum computer and we are off to a great start with leading customers like Lockheed, USC/ISI, Google and NASA/Ames having 512 qubit systems. The architecture of the system is quite different than anything that most people have ever encountered – it is simultaneously simpler and more complicated which is probably appropriate since it is a quantum computer. Simpler because the machine really executes one instruction that loads the qubits and the interaction between them, and more complicated because you transform your algorithm into a form of an optimization problem that the system solves natively.
In that arena, it is probably like conventional computing was in the 1950’s or 60’s (even before I was doing it!) – people creating new algorithms and solving problems that previously couldn’t even be attempted. So, to move the field ahead, we need more people developing algorithms, applications and software tools to keep pace with the rapid scaling of the hardware to more qubits and perhaps interconnect topologies.
HPCwire: Can Quantum Computing do anything to help solve the Moore’s Law problem that gets talked about so much? How does quantum change computing – how fast and where?
Bo Ewald: D-Wave’s quantum computing is fundamentally on different speedup, solution space and power consumption curves than conventional systems. In moving from the D-Wave One (128 qubits) to the D-Wave Two (512 qubits) in two years, the company measured speedups of several orders of magnitude due to both hardware and algorithm improvements. And, as part of a benchmark test by one of our customers, the D-Wave Two ran some optimization problems about ten thousand (yes ~10,000) times faster than commercial solvers at the time.
However, taking a balanced view, there are many applications that won’t map onto this architecture yet, or that run more slowly on it. But, the early results show the phenomenal promise of quantum computing as the capacity of the systems and their architectures improve. And, since these are supercooled, superconducting supercomputers – they use minimal power – ours uses only about 15.5 KW and we don’t expect to use any more than that for the next few generations.
HPCwire: As you look down the road at 2014 and beyond, what are the trends developing in high performance computing that you see as important, and how does quantum computing interact with those trends?
Bo Ewald: One of the key things to watch is the hybrid computing trend of more specialization, followed by more generalization, followed by . . . For example, looking at the development of the GPU products and markets, GPU’s were developed first to off load graphics processing from the more general purpose CPU’s. Then, as they became more useful and successful, GPU hardware and software have evolved to make them more general purpose.
I think that we’ll see the same with quantum computing – for example, the D-Wave systems will operate in concert with a big HPC system or data analytics engine. The D-Wave machine will work on optimization, machine learning, or perhaps Monte Carlo parts of an application as directed by the host. My guess is that over time, we’ll then see the quantum computer becoming more general purpose as well.
HPCwire: On a personal note, can you talk about your personal life? Your family, background, any hobbies?
Bo Ewald: I wasn’t so sure at the time, but I was fortunate to grow up in a hard-working, motivated family. Our parents both worked, they were interested and involved in what we were doing, set great examples for us to follow, and gave us a secure foundation from which it was easier to make leaps to follow your dreams or try something new.
I’ve also been surprised and pleased to learn that as long as you stay active (and have a little luck and good genes) you can stay in reasonable shape mentally and physically, even when you are as old as I am! I still think I’m about 28, and have a lot of experience at it, but I’m shocked that I can still lift weights about as well as I did when I was 28, can still run, ski, . . . Of course, that probably means that I wasn’t very good at those things to begin with.
HPCwire: One last question – is there anything about yourself that you can share that you think your colleagues would be surprised to learn?
Bo Ewald: They might be surprised to know that I usually cook Thanksgiving and Christmas dinners. I favor family recipes from my mother and grandmother, including twice-raised “Parker House” yeast rolls. It’s funny – at family get-togethers we never talk about work – in large part because my daughter Cydney (Sun & Dell), her husband Mike (Sun & IBM) and I all worked at different companies in the same markets!