Interview with David Patterson—Recipient of ACM A.M. Turing Award 2017, UCLA Bruin
The ACM A.M Turing Award is the highest honor in Computer Science and is often referred to as the “Nobel Prize in Computing”. The 2017 Turing Award was awarded to David Patterson and John Hennessy for their pioneering work in computer architecture design, in particular, for the impact of their work in the ARM processor which is used in nearly every modern smartphone.
David Patterson and John Hennessy created a systematic and quantitative approach to designing faster, lower power, and reduced complexity microprocessors. Their approach led to lasting and repeatable principles that generations of architects have used for many projects in academia and industry. The impact has been stunning: many tens of billions of processors use reduced complexity architectures. (Source: amturing.acm.org)
David Patterson did his AB, MS and PhD at UCLA after which he spent four decades at UC Berkeley as a Professor of Computer Science. During his illustrious career, David Patterson has received accolades such as the John von Neumann Medal, Diane S. McEntyre Award for Excellence in Teaching, and the inaugural Outstanding Alumnus Award of the UCLA Computer Science Department among several others.
We reached out to Professor Patterson and he very generously agreed for an interview. Professor Patterson, apart from being a prolific academic and an outstanding educator, is also very witty and fun to talk to and has many interesting life stories. Here’s the interview.
DP: Prof. David Patterson, MM: Mihir Mathur, HL: Helen Lee
HL: Thank you so much for taking the time to talk to us!
DP: Happy to do that for my alma mater!
HL: So what was your initial reaction after winning the Turing award, were you expecting it?
DP: I was certainly shocked. What happens is that you hear about it when you’re nominated, and we were nominated many years ago. When you’re nominated for something, you can’t help but think you’re gonna get it, but as the years went by and we didn’t hear I was like ‘OK, this isn’t gonna happen’. So I was shocked that many years later we got the award. Apparently, there are so many good candidates that there’s a long queue!
HL: Who were some of the other candidates you were competing with?
DP: Well the person who received it last year was the person who invented the World Wide Web! So that was a pretty good choice.
(Sir Tim Berners Lee won the Turing Award in 2016)
MM: Professor Patterson, you’ve written a lot of books on Computer Architecture. In fact, we at UCLA study from your book for one of our courses (CSM151B). So which is your favorite book that you’ve written and how was your experience writing so many books?
DP: I would say, just because of the impact, it would have to be the first one I wrote with John Hennessey — Computer Architecture: A Quantitative Approach. I had done a couple of other books that weren’t really in my field, one was on Computer Literacy and one was on the Programming Language called SmallTalk, which is a predecessor to Object Oriented Systems. We decided to write one in our field because we were very frustrated with the quality of the textbooks out there—they were basically just catalogue descriptions of features. So you could kinda say what the pros and cons were but you had a really hard time quantitatively deciding what was the best thing to do. So people in the industry would have used their intuition or hunch about what is the right thing to build.
So, John and I complained for a few years. Then it became clear that I was going to be Chair of Computer Science at Berkeley. And we thought, at the time, that when you become chair, your life is over as you can only do bureaucratic things. So we forced a deadline to write. In 1989, we worked really hard to write that book. We treated it like a chip—we had an alpha release, a beta release and we would class-test those releases and incorporate ideas.
The reaction was overwhelming. We thought we were writing primarily for college students, graduate students particularly, but it became very popular with people in industry. In fact, it was popular even with software people to try and understand how hardware works. At the time, Microsoft stocked it in their supply stores!
HL: You did your BS, MS and PhD at UCLA. What are some of your favorite memories from UCLA?
DP: I think just how fortunate I was to end up where I am today. In my junior year, I was a Math major; there was no Computer Science major yet. There was a CS PhD program, but no undergraduate degree. In one quarter of my junior year, a Math class I wanted to take was cancelled so I had to find something to fill up my schedule and there was a 2 unit Computer Course. When I got into it, I could see the magic of software—the ideas that could come alive inside a computer. I was hooked. So in my senior year, I took every Computer course I could find. I did really well on one of the classes, and I said innocently to the person teaching it that I’d much rather do computer stuff than work at my dad’s factory in South Central LA. So on his own, he went and found me a job as an undergraduate student at a research project. So in my last quarter or two I didn’t have to work in the factory and I got to program, which I loved. I liked it so much that I asked my wife if she thought it was okay to get a masters and she said that’d be great, if you can fly, fly high. Then I was put in the office with three other grad students and they were all getting their PhDs so I thought, well, maybe I should get a PhD!
I could have very easily been a Math major, never taken a Computer course, never got to grad school. The opportunities at UCLA, and the kind of people who went out of their way to help me— that’s how I got my PhD.
MM: So did you become interested in architecture at UCLA or was it later on?
DP: Well that’s another one of those fortunate roads. My wife and I got married as teenagers and we were young enough that we waited a whole year before we had kids. I was working as a Research Assistant 20 hours a week to support my wife and two young kids, and two years into my PhD program the research grant came to an end. So my research advisor got me an opportunity to interview at Hughes Aircraft. It was about 5 miles from campus—close enough that I could ride my bike. They offered me a job which instead of software (which I had been doing), was to build computers that would go on airplanes. That was my real hardware education.
Had I not lost my job as a Research Assistant, which sounded terrible at the time, I wouldn’t have got the job at Hughes. And had I not gone to Hughes, I wouldn’t have got hardware trained and wouldn’t have got a job at Berkeley for hardware. I feel like a crinkle ball bouncing through the pins so I was able to get this job.
HL: So you’ve won a lot of teaching awards and research awards. How do you manage to be dedicated to both teaching and research?
DP: I think its a common misperception that those are in conflict. What I have seen is that there’s a correlation between great research and great teaching. The principle on which higher education in the United States is based on is that we can attract great people who want to push the frontiers of research and they’ll also be great in the classroom. Though, I’m sure you’ve had examples of people who’re terrible teachers. They’d also be terrible researchers! There’s some people who only do one and not the other, but boy, I’ve seen a lot of examples of them together.
There’s a picture taken when I was chair of the department. On the Berkeley campus, the highest teaching award is the Distinguished teaching award. There were five of us [in the picture] who won that award, and I’m looking at that photo now and all five of us have been elected to the National Academy of Engineering, and three of the five, including me, have won the Turing award! That’s the highest research award in Computer Science and they won the highest teaching award at Berkeley. That’s concrete evidence of the correlation!
MM: So now that you are retired from Berkeley, we saw that you are working at Google. Can you share with us some of the projects that you are working on and how your experience has been?
DP: After turning 70, I couldn’t put in the 60 hour work weeks that you need to be a successful faculty member, so I knew I needed to cut back. Now, I am working 1 day a week at Berkeley, 1 day a week at my beach house, and then 1, 2, or 3 days a week at Google. What I’m working on is what we call Domain Specific Architecture. With the end of Moore’s Law, general purpose microprocessors are hardly improving. Last year, with a standard benchmark, they improved 3%. That’s like doubling every 20 years. The only way to make significant improvement in performance and energy efficiency is to build things that do narrow tasks but extremely well. An attractive target is the excitement of machine learning and deep neural networks, so what I’m working on at Google is domain specific hardware for deep neural networks. Two of them have been revealed: Tensor Processing Unit (TPU) Version 1 and Version 2.
HL: What is your favorite non-CS thing to do?
DP: Playing soccer. I should mention that I wrestled when I was at UCLA. I played on the Rugby team as well. There were 3 levels on the Rugby team and I played on the second and third level. I like sports. After I came to Berkeley, my kids started to play soccer, but I had never played it before. I still play soccer two times a week even at this advanced age and it’s a lot of fun.
MM: Have you ever played soccer with your students?
DP: Yes, I’m one of the older players, but the age range goes from 20s on up. In fact, at a computer architecture conference that I attended, they had a soccer game. I showed up and played, and they were surprised that I was good. They thought, “Oh it’s an old guy, I’ll just run by him”, but I didn’t let them do that!. And Berkeley CS last year had a student vs. faculty game. The faculty won!
MM: We also saw online that you hold a record for weightlifting.
DP: Yeah I can tell you the story there. Because of wrestling, I lifted weights my whole life. In teaching a computer architecture class, especially at the undergraduate level, I try to talk about the importance of metrics. I found that students like to hear about our personal lives, so that they aren’t just being lectured by a robot. So I told them about bench pressing. Now, the thing about bench pressing is that when you ask about how strong someone is, you usually ask about their bench press. But you have to account for age. I did a metric to account for age and I called it pound-years, where you multiply your age by how much you can bench press (as a joke). And I said, a 20 year old bench pressing 350 pounds is 7000 pound-years, which is a lot, but here I am, 50 years old and I can bench 325 pounds, so that’s 16000 pound-years, so I’m much better than them.
I then decided to look up weightlifting records, and I saw that for my age group and my weight, I was pretty competitive! So I entered a bunch of competitions after and won them, and I set a record for bench pressing. I also entered one competition that I thought was a bench press competition, but it actually had 3 weight events: bench press, squats, and deadlifts which I hadn’t practiced. I ended up setting a record for deadlifts too.
MM: That is so impressive!
DP: If you are interested, at my official retirement party, I gave a lecture called: “How to be a Bad Professor”, and I talk about a lot of my life stories there. I also gave a “How did I get here” talk which talks about how I got to Berkeley. If you want, take a look.
MM: Yeah I actually watched your Ten Commandments about being a bad professor last night. Those were pretty hilarious.
DP: I’m always trying to get in some humor!
MM: What’s your advice for students who are interested in research?
DP: Get involved in a research project as an undergraduate student. Although I didn’t consciously do it, a classic thing to do is do well in a class and talk to your professor afterwards about whether there are things that you can be involved in. There are a lot of research going on and not all the projects are writing papers. I think just being around a research project and graduate students or professors can be eye opening. It has certainly changed my life, but it is hard to tell whether you like research until you try it. There is a surprising number of opportunities. You do need to have a couple classes behind you to be useful, but you can’t be shy.
What struck me is that I hold office hours and I have days when I’m in the office by myself and I’m just reading emails. I didn’t think I was an intimidating person or anything, but people didn’t come by. Take advantage of office hours because faculty has to be there for that and get to know the faculty and ask about their research projects and see if there is anything that you are interested and want to get involved in. I would start off with just trying to help out and then it could lead to getting paid, at UCLA and Berkeley and there are a lot of research opportunities. I know that today, people do internships in the summer, which is a nice thing to do in terms of getting to know what companies are like. But if you give a chance to be involved in a research project in the summer, there’s a lot of time because faculty aren’t teaching and grad students aren’t taking classes over the summer. There is a lot of work that gets done in the summer, so if you give a chance to work on a research project during the summer, that’s a tremendous opportunity if you are thinking about going to grad school.
MM: That’s all the questions that we had. Thank you for taking out the time to talk to us and we are really looking forward to sharing the interview with everyone at UCLA.
DP: Yeah no problem!
I was at UCLA when John Wooden was winning all the games and would go to those games. The ironic thing is that the people who were around at that time now have buildings and streets named after them. UCLA is a great place to go and the UC system is just an amazing public university. In a field like ours, you really don’t want to be using 30 year old lecture notes to talk about computing. By having researchers lead the way in education, you are having the important, recent information. I hope everyone appreciates everything that the UC schools offers them.
HL: Thank you so much for talking to us. We really appreciate it!
Thanks for reading this post. For more such stories, follow Tech@UCLA.
