Visual summary of operating lessons from David Patterson.

Lessons from David Patterson

David Patterson is a computer scientist who co-developed RISC and is a professor emeritus at UC Berkeley. This collection gathers his own words from a Computer History Museum oral history and several interviews. — Oral History of David Patterson (Computer History Museum).

Part 1: Academic Research & Mentorship

  1. On courage and fortune favoring the bold: David Patterson: "Having courage to do things kind of goes along with the advice is fortune favors the bold." — Turing Award Winner: TPUs vs GPUs vs CPUs, Computer Architecture, RISC vs CISC (Developing.dev podcast with Ryan Peterman; David's turns, transcript).
  2. On having no role models: David Patterson: "I don't think I would've gone to graduate school if he hadn't done that because I had no role models and no plans." — Oral History of David Patterson (Computer History Museum, 2007; his answers).
  3. On not getting advice on the academic system: David Patterson: "Alas at that time, or perhaps with my advisor, I didn't get any real advice how to interview, how the academic system worked." — Oral History of David Patterson (Computer History Museum, 2007; his answers).
  4. On seeking feedback: David Patterson: "For some reason my nature is to get lots of feedback on everything I do, maybe because I wasn't a very good writer at the time." — Oral History of David Patterson (Computer History Museum, 2007; his answers).
  5. On working on important problems: David Patterson: "I agree with Richard Hamming’s opinion: If you don’t work on important problems, you are not going to do important things." — A Virtual Interview with John Hennessy and David Patterson (SIGARCH; Patterson's answers).
  6. On the mentor who got him into research: David Patterson: "If he hadn't taken the initiative to try and get this undergraduate student involved in research while an undergraduate, there's no way I would be where I am today." — Oral History of David Patterson (Computer History Museum, 2007; his answers).
  7. On bonding a team: David Patterson: "Somehow I was on good teams, and I learned how to bond people together and form a team and to be successful as a team." — Oral History of David Patterson (Computer History Museum, 2007; his answers).
  8. On interacting with industry: David Patterson: "I always have used the crutch of interacting with industry to help figure out whether this problem's important or not." — Oral History of David Patterson (Computer History Museum, 2007; his answers).
  9. On giant public universities: David Patterson: "I’m the first of my family to graduate from college, got all my degrees at UCLA and spent all of my life at UC Berkeley, so all’s I know is giant, large, public universities." — Dave Patterson and John Hennessy on Recode Decode (Vox transcript; Patterson's labelled turns).

Part 2: Hardware Design Philosophy

  1. On keeping the instruction set simple: David Patterson: "If we kept the instruction set simple, we could iterate them faster, it could potentially be more power efficient, and things like that." — Turing Award Winner: TPUs vs GPUs vs CPUs (Software Engineering Daily episode 1810; his turns).
  2. On polysyllabic instructions: David Patterson: "What it came down to is for the more sophisticated instruction sets, you can think about polysyllabic words in the vocabulary, you needed fewer of them to execute the program, but they might run more slowly." — Turing Award Winner: TPUs vs GPUs vs CPUs (Software Engineering Daily episode 1810; his turns).
  3. On a common goal with IBM and Stanford: David Patterson: "The RISC stuff, I tried to join forces with IBM and Stanford because you could think of us having a common goal." — Oral History of David Patterson (Computer History Museum, 2007; his answers).
  4. On what an instruction set is: David Patterson: "All right, so when I explain it to my relatives, what I say is when software talks to hardware, there's a vocabulary, and the technical name for that vocabulary is instruction set." — Turing Award Winner: TPUs vs GPUs vs CPUs (Software Engineering Daily episode 1810; his turns).
  5. On the core of the instruction set: David Patterson: "If you look at the very core of that instruction set, it's similar to RISC I, but since we have so many transistors today and we had that idea of adding features when you need them." — Turing Award Winner: TPUs vs GPUs vs CPUs (Software Engineering Daily episode 1810; his turns).
  6. On compiler writers and architects: David Patterson: "Often the compiler writers, the architects, would come up with a sophisticated instruction, and the compiler writers would say, well, we don’t need that." — Turing Award Winner: TPUs vs GPUs vs CPUs, Computer Architecture, RISC vs CISC (Developing.dev podcast with Ryan Peterman; David's turns, transcript).
  7. On evaluating performance early: David Patterson: "And so performance was pretty early, how do you evaluate performance was pretty early in what we were doing, so that we could set the framework for the whole rest of the book." — Oral History of David Patterson (Computer History Museum, 2007; his answers).
  8. On the cost of microcode interpreters: David Patterson: "So you pay the extra overhead of the microcode interpreter to have these sophisticated instructions, and then the compiler doesn’t even use them." — Turing Award Winner: TPUs vs GPUs vs CPUs, Computer Architecture, RISC vs CISC (Developing.dev podcast with Ryan Peterman; David's turns, transcript).
  9. On replacing one big thing with many small ones: David Patterson: "It seems radical but replace the big one by a lot of small ones, and if you add some extras you get reliability." — Oral History of David Patterson (Computer History Museum, 2007; his answers).
  10. On a more elegant implementation: David Patterson: "Two other guys, Manolis Katevenis and Bob Sherburne built, without the constraints of the course and without the deadlines of the course, built a much more elegant implementation of the ideas, and a much better implementation of the ideas." — Oral History of David Patterson (Computer History Museum, 2007; his answers).

Part 3: The Semiconductor Industry & Moore's Law

  1. On Moore's law slowing: David Patterson: "We could see that Moore's law was slowing down and thought what the future looked like is you'd have a core instruction set, and then you'd add special-purpose instructions for specific domains." — Turing Award Winner: TPUs vs GPUs vs CPUs (Software Engineering Daily episode 1810; his turns).
  2. On the end of Dennard scaling: David Patterson: "With this ending of Moore’s Law and Dennard scaling, there’s no other choice." — Dave Patterson and John Hennessy on Recode Decode (Vox transcript; Patterson's labelled turns).
  3. On the golden age: David Patterson: "I mean, if I was younger and had more energy, I would, this golden age sounds pretty good to me." — Dave Patterson and John Hennessy on Recode Decode (Vox transcript; Patterson's labelled turns).
  4. On the end of rapid performance gains: David Patterson: "In the 1980s, 1990s, 2000s, you have a laptop and your friend’s laptop would be like four times as fast as yours." — Turing Award Winner: TPUs vs GPUs vs CPUs, Computer Architecture, RISC vs CISC (Developing.dev podcast with Ryan Peterman; David's turns, transcript).
  5. On what the end of Moore's Law means: David Patterson: "Now I think what people assume that means is if we are no longer on Moore’s Law, that technology is not improving." — Turing Award Winner: TPUs vs GPUs vs CPUs, Computer Architecture, RISC vs CISC (Developing.dev podcast with Ryan Peterman; David's turns, transcript).
  6. On a chip that failed on power: David Patterson: "And Intel actually had a microprocessor that failed one of their generations because they just couldn’t get the power down enough." — Turing Award Winner: TPUs vs GPUs vs CPUs, Computer Architecture, RISC vs CISC (Developing.dev podcast with Ryan Peterman; David's turns, transcript).
  7. On why multicore became necessary: David Patterson: "So then what that forced us to go to multicore is that before that, it was easiest for the programmer, for everybody, if there was one very sophisticated processor that did everything, but we couldn’t do that anymore." — Turing Award Winner: TPUs vs GPUs vs CPUs, Computer Architecture, RISC vs CISC (Developing.dev podcast with Ryan Peterman; David's turns, transcript).
  8. On the doubling of transistors: David Patterson: "Those of us who believed in Moore’s Law believed eventually the microprocessor would be the way we did all our computing, with the doubling of transistors every year or two." — Turing Award Winner: TPUs vs GPUs vs CPUs, Computer Architecture, RISC vs CISC (Developing.dev podcast with Ryan Peterman; David's turns, transcript).
  9. On data centers and local utilities: David Patterson: "But in particular regions, if somebody builds a lot of data centers in the same region, then that utility company could be taxed, and that's new." — Turing Award Winner: TPUs vs GPUs vs CPUs (Software Engineering Daily episode 1810; his turns).

Part 4: Building Systems

  1. On a RAID patent: David Patterson: "I think I have never filed a patent at Berkeley, but they are probably, as Randy looks chronologically, they probably came up with the RAID 5 idea before we did." — Oral History of David Patterson (Computer History Museum, 2007; his answers).
  2. On reliability making RAID popular: David Patterson: "I think Randy and I, for many years, thought of this as a storage performance thing, and what clearly made it popular was the reliability piece." — Oral History of David Patterson (Computer History Museum, 2007; his answers).
  3. On clusters of workstations: David Patterson: "So, that led to the Network of Workstations [NOW] project and John and I have had these friendly rivalries here." — Oral History of David Patterson (Computer History Museum, 2007; his answers).
  4. On RAID as a way of thinking: David Patterson: "But I always thought the RAID paper was about, it was like the case for RISC, right, this whole technological thing, right." — Oral History of David Patterson (Computer History Museum, 2007; his answers).
  5. On naming the options: David Patterson: "To explain what we were doing and why it was different, we needed to write this taxonomy and, this was a paper that I ended up being the first author of and I just numbered the options, right." — Oral History of David Patterson (Computer History Museum, 2007; his answers).
  6. On scale and building supercomputers: David Patterson: "Well, I spent the sabbatical at Thinking Machines in 1990 and Oral History of David Patterson saw all the effort they did to build these things." — Oral History of David Patterson (Computer History Museum, 2007; his answers).
  7. On searching with Google: David Patterson: "You think about how many people, including me, use Google all the time, just search every single day." — Oral History of David Patterson (Computer History Museum, 2007; his answers).
  8. On a rival that shipped RAID first: David Patterson: "And the story there, we were broadcasting this to the winds, what happened at just this time, EMC [Corporation] in Boston, which was building cheap DRAM [Dynamic Random Access Memory] to plug into IBM mainframes." — Oral History of David Patterson (Computer History Museum, 2007; his answers).
  9. On the scale point: David Patterson: "And, I could see how long it would take a company, a small company the size of Thinking Machines to put this all together with the software and stuff like that." — Oral History of David Patterson (Computer History Museum, 2007; his answers).

Part 5: Artificial Intelligence & Domain-Specific Architectures

  1. On memory-focused design: David Patterson: "We should be thinking of being more memory-focused, thinking of the problem of how do we get access to the data rather than thinking of it as computing is the focus of what's being done." — Turing Award Winner: TPUs vs GPUs vs CPUs (Software Engineering Daily episode 1810; his turns).
  2. On the serving side: David Patterson: "The serving side is there's more opportunity, I think, for hardware people to innovate." — Turing Award Winner: TPUs vs GPUs vs CPUs (Software Engineering Daily episode 1810; his turns).
  3. On neural networks and GPUs: David Patterson: "And within a few years, everybody switched over, and not only were they doing neural networks, but they were using GPUs." — Turing Award Winner: TPUs vs GPUs vs CPUs, Computer Architecture, RISC vs CISC (Developing.dev podcast with Ryan Peterman; David's turns, transcript).
  4. On narrower workloads: David Patterson: "So if you tell me I only have to run a narrow class of programs and I don’t have to necessarily run all the operating systems, everything else, well, yes, I could shuffle those resources and do something much more efficient for something." — Turing Award Winner: TPUs vs GPUs vs CPUs, Computer Architecture, RISC vs CISC (Developing.dev podcast with Ryan Peterman; David's turns, transcript).
  5. On the timing of machine learning: David Patterson: "So just coincidentally, where we were technologically, right around 2012, 2015 is when machine learning AI burst on the scene." — Turing Award Winner: TPUs vs GPUs vs CPUs, Computer Architecture, RISC vs CISC (Developing.dev podcast with Ryan Peterman; David's turns, transcript).
  6. On what a graphics processor left out: David Patterson: "It didn’t need to do everything that a general-purpose processor needs to, didn’t have to support virtual memory, it didn’t have to support a compiler, which is a pretty radical idea for architectures." — Turing Award Winner: TPUs vs GPUs vs CPUs, Computer Architecture, RISC vs CISC (Developing.dev podcast with Ryan Peterman; David's turns, transcript).
  7. On getting performance from memory: David Patterson: "What they decided to do to do the graphics job is they need to get a lot of performance out of the memory system." — Turing Award Winner: TPUs vs GPUs vs CPUs, Computer Architecture, RISC vs CISC (Developing.dev podcast with Ryan Peterman; David's turns, transcript).
  8. On when GPUs arrived: David Patterson: "So what happened is around 2000 is when GPUs came along, and these were what we call domain-specific architectures." — Turing Award Winner: TPUs vs GPUs vs CPUs, Computer Architecture, RISC vs CISC (Developing.dev podcast with Ryan Peterman; David's turns, transcript).
  9. On a niche product that took off: David Patterson: "So there was this niche product that happened to do pretty fast single-precision floating point and even half-precision floating point, and they were pretty cheap." — Turing Award Winner: TPUs vs GPUs vs CPUs, Computer Architecture, RISC vs CISC (Developing.dev podcast with Ryan Peterman; David's turns, transcript).
  10. On specializing: David Patterson: "But by specializing, you can reduce the costs and you can produce the energy and the carbon footprint by doing that." — Turing Award Winner: TPUs vs GPUs vs CPUs (Software Engineering Daily episode 1810; his turns).

Part 6: Open Source Hardware (RISC-V)

  1. On the origin of RISC-V: David Patterson: "What happened was, the origin story for RISC-V is in 2010, we were going to do research around parallel computing that was sponsored by Intel and Microsoft." — Turing Award Winner: TPUs vs GPUs vs CPUs (Software Engineering Daily episode 1810; his turns).
  2. On the thirst for an open architecture: David Patterson: "Then in talking to them, we found out there was this thirst for an open architecture that people could use rather than have to get a proprietary architecture." — Turing Award Winner: TPUs vs GPUs vs CPUs (Software Engineering Daily episode 1810; his turns).
  3. On a common language across computing: David Patterson: "Most of the people involved really like the idea that potentially we could have a lingua franca across all computing." — Turing Award Winner: TPUs vs GPUs vs CPUs (Software Engineering Daily episode 1810; his turns).
  4. On proprietary instruction sets: David Patterson: "But yes, that x86 is tied to that company who's having difficulties and there have been many other instruction sets that have gone away because of the fortune there." — Turing Award Winner: TPUs vs GPUs vs CPUs (Software Engineering Daily episode 1810; his turns).
  5. On optional features: David Patterson: "So, we have, RISC-V allows optional features for all kinds of applications like encryption, or machine learning, or things like that." — Turing Award Winner: TPUs vs GPUs vs CPUs (Software Engineering Daily episode 1810; his turns).
  6. On teaching and research: David Patterson: "So I enjoyed teaching and I wanted to see if I could both do teaching and research at a place like Berkeley." — Dave Patterson and John Hennessy on Recode Decode (Vox transcript; Patterson's labelled turns).
  7. On tying research to courses: David Patterson: "Courses are absolutely going to start and stop, so I thought, “Why don’t I tie the research to the courses and then we’d have deadlines and be able to make steady progress?” That’s where the name came from." — Dave Patterson and John Hennessy on Recode Decode (Vox transcript; Patterson's labelled turns).
  8. On the RISC architectures at Berkeley: David Patterson: "Well, we did four risk architectures at Berkeley in the 1980s and Hennessy did a couple that are called MIPS at Stanford." — Turing Award Winner: TPUs vs GPUs vs CPUs (Software Engineering Daily episode 1810; his turns).
  9. On executing more simple instructions faster: David Patterson: "And it turned out that RISC style tended to execute about 30% more instructions, but we could execute them about four or five times faster." — Turing Award Winner: TPUs vs GPUs vs CPUs (Software Engineering Daily episode 1810; his turns).

Part 7: Writing, Teaching, and Communication

  1. On settling arguments with numbers: David Patterson: "And because there weren’t ways to settle the arguments quantitatively, people would just argue about it." — Turing Award Winner: TPUs vs GPUs vs CPUs, Computer Architecture, RISC vs CISC (Developing.dev podcast with Ryan Peterman; David's turns, transcript).
  2. On finishing the book: David Patterson: "John took time off too and so we decided it was now or never, and what we thought is, I had actually had some experience of writing textbooks." — Oral History of David Patterson (Computer History Museum, 2007; his answers).
  3. On writing the case for RISC: David Patterson: "So, I wrote this draft of the paper, the case for reducing instruction set computers, and I sent him a draft, and he gave lots of comments back." — Oral History of David Patterson (Computer History Museum, 2007; his answers).
  4. On aggressive schedules: David Patterson: "Publishers hear this all the time: authors saying that they have an aggressive schedule, but we went on an aggressive schedule." — Oral History of David Patterson (Computer History Museum, 2007; his answers).
  5. On frustration with textbooks: David Patterson: "We just got increasingly frustrated with the textbooks, which were still from like the supermarket catalog era of describing architecture." — Dave Patterson and John Hennessy on Recode Decode (Vox transcript; Patterson's labelled turns).
  6. On choosing a title: David Patterson: "I picked the Monty Python quote, 'And now for something completely different,' and that was in the first chapter." — Oral History of David Patterson (Computer History Museum, 2007; his answers).
  7. On introductory courses: David Patterson: "I didn’t know that the system would work, but we have introductory courses in computer science with more than a thousand students." — Dave Patterson and John Hennessy on Recode Decode (Vox transcript; Patterson's labelled turns).
  8. On angering powerful companies: David Patterson: "These were two assistant professors, not tenured faculty, who were making powerful companies very mad at them, maybe me more than John because of my personality." — Oral History of David Patterson (Computer History Museum, 2007; his answers).

Part 8: Career Strategy & "Bad Careers"

  1. On real-world microcode: David Patterson: "So, I came back and I could see the real world experience in building production microcode and just how chock full… errors were there all the time when they were doing change orders." — Oral History of David Patterson (Computer History Museum, 2007; his answers).
  2. On microcode bugs: David Patterson: "It doesn't make sense to design microprocessors this way and, with this extra RAM, and cost, and patching and the field." — Oral History of David Patterson (Computer History Museum, 2007; his answers).
  3. On starting the next thing: David Patterson: "I think another thing, as I look back on my career, is I often start talking about the next thing I'm going to work on before that." — Oral History of David Patterson (Computer History Museum, 2007; his answers).
  4. On what industry offers: David Patterson: "Plus, obviously the industry will help you figure out what the problems are." — A Virtual Interview with John Hennessy and David Patterson (SIGARCH; Patterson's answers).
  5. On making time for fun: David Patterson: "I think when you’re a kid, you don’t have to tell kids to play, but as you get adult, you get so busy you don’t think you have time to have fun." — Turing Award Winner: TPUs vs GPUs vs CPUs, Computer Architecture, RISC vs CISC (Developing.dev podcast with Ryan Peterman; David's turns, transcript).
  6. On a successful co-author: David Patterson: "So, it was nice to have a really successful, brilliant co-author both to share the work and make us work on it, and in retrospect I don't think I could have picked anybody else to work with or had it work out as well." — Oral History of David Patterson (Computer History Museum, 2007; his answers).
  7. On killing your own design: David Patterson: "We should embrace MIPS and kill our own design.' And it was really hard for me to say that, and Bill Joy said at the time, he says, 'Look, we've talked to the MIPS guys." — Oral History of David Patterson (Computer History Museum, 2007; his answers).
  8. On what makes a career feel good: David Patterson: "The people who worked with people like ministers or teachers or doctors felt really good about what they do with their careers." — Turing Award Winner: TPUs vs GPUs vs CPUs, Computer Architecture, RISC vs CISC (Developing.dev podcast with Ryan Peterman; David's turns, transcript).
  9. On calling Berkeley: David Patterson: "Late in the interview season my wife insists that I call Berkeley, and I call up the chair of the department, Elwyn Berlekamp." — Oral History of David Patterson (Computer History Museum, 2007; his answers).