On the heels of Intel’s reaffirmation that it will deliver the first U.S. exascale computer in 2021, which will feature the company’s new Intel Xe architecture, we bring you our interview with our 2019 Person to Watch Jim Keller, head of the Silicon Engineering Group at Intel. [Ed. note: the interview was conducted prior to Monday’s Aurora announcement, but allusions to the future design are made.]
Jim joined Intel from Tesla, where he served as vice president of Autopilot and Low-Voltage Hardware. Prior to Tesla, he held roles at AMD (where he led the development of the Zen architecture) and P.A. Semi (which was acquired by Apple Inc.).
Overall, Jim has more than 20 years of experience in x86 and ARM-based microarchitecture design across a broad range of platforms, including PCs, servers, mobile devices, and cars. He holds a bachelor’s degree in electrical engineering from Pennsylvania State University.
HPCwire: Hi Jim, congrats on all of your success and recognition. As a lauded chip architect, you could work anywhere, why Intel?
Jim Keller: There are a couple reasons. The scale of the opportunity and the magnitude of the impact that Intel has in the world is really interesting — there’s a bunch of new innovation to do, and Intel can do it at a really interesting scale. We are facing a really big change in how people use computers, the types of computing, and what computing is doing to our world. It’s exciting to be a part of that journey. We are developing new technologies for a bunch of applications to meet these needs, and I like that kind of challenge.
Working at Intel is also a lot of fun. I have been very happy to meet so many people who care, are engaged, and want to do great stuff. I love being involved with people who care.
HPCwire: Is there a red thread that runs throughout your career? Your ethos to design?
Number one is I always work on interesting problems. Number two is I really get into the details of what I’m doing and what the results of my design – both good and bad – are. Details and problems are inspiring to me. You learn more from your mistakes than your successes.
HPCwire: Is Jim Keller concerned about Moore’s law?
Not even a little. Moore’s Law is not dead, and we will elaborate on our plan to continue it over time. My observation is that the end has always been near, but there has always been and always will be new things we can do to keep pushing that end out. The real challenge is, what happens as we continue to scale technology? What new architecture do we need and how do the problems change? That’s the interesting place.
HPCwire: Are the requirements of AI and HPC aligned or in contention when it comes to the future of silicon? How so?
They’re completely aligned and have more in common than you think – transistor density yields computational intensity. The evolutionary journey used to be called Bell’s Law. Every ten years there was a new kind of computer, and to an interesting extent, there were shrinks in performance enhancements for each new class of computer. The recent adventure is we have gone from scalar computing to vector to matrix to spatial. Each step has been a quantum leap in the use of transistors for increasingly complicated computational methods. As best as I can tell, more is better, and there will be additional new architectural adventures enabled by Moore’s Law.
HPCwire: Outside of the professional sphere, what can you tell us about yourself – personal life, family, background, hobbies, etc.? Is there anything about you your colleagues might be surprised to learn?
I’m a fairly energetic person. I have two children that constantly surprise and amaze me, and they both think I’m a goof, which is fine with me. I kitesurf, snowboard and run, and I’m also a pilot.