6th HLF – Interviews with mathematics and computer science laureates: David A. Patterson

Video in TIB AV-Portal: 6th HLF – Interviews with mathematics and computer science laureates: David A. Patterson

Formal Metadata

6th HLF – Interviews with mathematics and computer science laureates: David A. Patterson
Title of Series
No Open Access License:
German copyright law applies. This film may be used for your own use but it may not be distributed via the internet or passed on to external parties.
Release Date

Content Metadata

Subject Area
Laureates at the 6th HLF sit down with Tom Geller, Tom Geller Productions, to discuss their career, mentoring and their experience at the Heidelberg Laureate Forum (HLF). These renowned scientists have been honored with most prestigious awards in mathematics and computer science: Abel Prize, ACM A.M. Turing Award, ACM Prize in Computing, Fields Medal and Nevanlinna Prize. The opinions expressed in this video do not necessarily reflect the views of the Heidelberg Laureate Forum Foundation or any other person or associated institution involved in the making and distribution of the video. Background: The Heidelberg Laureate Forum Foundation (HLFF) annually organizes the Heidelberg Laureate Forum (HLF), which is a networking event for mathematicians and computer scientists from all over the world. The HLFF was established and is funded by the German foundation the Klaus Tschira Stiftung (KTS), which promotes natural sciences, mathematics and computer science. The HLF is strongly supported by the award-granting institutions, the Association for Computing Machinery (ACM: ACM A.M. Turing Award, ACM Prize in Computing), the International Mathematical Union (IMU: Fields Medal, Nevanlinna Prize), and the Norwegian Academy of Science and Letters (DNVA: Abel Prize). The Scientific Partners of the HLFF are the Heidelberg Institute for Theoretical Studies (HITS) and Heidelberg University.

Related Material

Mathematics Turing test Internet forum Computer Musical ensemble
Multiplication sign Direction (geometry) 1 (number) Set (mathematics) Open set Mereology Food energy Perspective (visual) Expected value Programmer (hardware) Mathematics Different (Kate Ryan album) Endliche Modelltheorie Mainframe computer Area Programming language Pattern recognition Sound effect Bit Computer Reduced instruction set computing Microprocessor Degree (graph theory) Shooting method Process (computing) Computer science Right angle Quicksort Turing test Laptop Divisor Observational study Virtual machine Student's t-test Coprocessor Event horizon Computer Power (physics) Moore's law Computer hardware Operating system Metropolitan area network Computer architecture Domain name Turing test Artificial neural network Projective plane Content (media) Counting Planning Cartesian coordinate system Limit (category theory) Numerical analysis Faculty (division) Compiler Word Pointer (computer programming) Software Integrated development environment Factory (trading post) Internet forum Universe (mathematics) Video game
Area Turing test Open source Projective plane Virtual machine Set (mathematics) Open set Reduced instruction set computing Stack (abstract data type) Theory Field (computer science) Moore's law Mathematics Befehlsprozessor Machine learning Bit rate Internet forum Computer hardware Operating system Information security Computer architecture
Area Turing test Multiplication sign Internet forum Quicksort Student's t-test Numerical analysis Computer architecture
Turing test View (database) Distribution (mathematics) Multiplication sign Physical law Computer Computer Mathematics Process (computing) Internet forum Musical ensemble Videoconferencing Computer architecture
first a very basic question what is your work for which you received the most recognition I've got the Turing award for my work on reduced instruction set computers at risk this it's not a very hard idea it was just very controversial at the time so basically when software talks to hardware there's a vocabulary so the hardware can understand it and NDC's in my notion was contrary to the prevailing wisdom which at the time he thought you'd a very rich vocabulary too for software talk to hardware and our idea was instead to have a very simple vocabulary not a lot of words in very simple words then the technical question was when you read these words if they're more complicated you wouldn't need as many that how fast could you read them and it turned out he had to read about 25% more instructions of the simple instructions and the complicated ones but we could read them about five times faster so the net effect was about a factor of four better so that's why you know today 99% of processors are RISC processors they say it was controversial at the time were their predecessors were their papers for example or people whose work inspired you to follow this path papers not so much you know John and I both started research projects at our two universities where we asked well what makes sense for microprocessors we'd see what the larger computers the mainframes and mini-computers have done but shouldn't we let's start from first principles to do that there was an earlier project at IBM led by John Kok who was also a Turing Award winner who had the idea of simpler instruction sets for bigger computers so that was a precedent of our work but unfortunately they weren't allowed to write papers about what they did so it was kind of a a rumor and that Johncock would visit us and inspire us but yeah this it was more what makes sense for microprocessors it wouldn't got us going so is this your
first HLF yes my first hls we just got the award in March and so it's my first chance oh that's right you're you're the the newest thing yeah I'm really I'm the newbie yes and when did you get in oh I came in Wednesday so today's Monday so yeah fourth day's going yeah so you've had a chance to interact a little bit with the students and yeah well mostly last night there was a German 50th anniversary party and for a co man that's was the earlier event not too many students were there so do you have any expectations or hopes for the event no I was came to head see what it was like but it was kind of fun though there was a woman from Nigeria who was sampled on the opening ceremony and asked her what she wanted to do and she said she wanted to see Dave Patterson so so I went out of I make make sure we connected last night so that fulfilled her wish and we talked about you know what she's teaching in Nigeria and whether there are some ideas I could help with pointers some course material and to help her out so that was fun so going back to to how you got started were there any specific mentors I mean we talked a little bit about and who were they and how do they love what you know as I think when I give my talk I'm gonna give my life story there but I had no plans to go to graduate school or study computer science I was a math major and there was not yet an undergraduate degree in computer science so as a math major in the end of my junior year a course was cancelled so I had to take something so I took a computer course and loved it and then in the middle of my senior year I took a course and from and did well and just casually mentioned to the instructor that but should rather be doing this I was working part-time in kind of Ana factory to support myself as it went through school I should rather do computer science than work in a factory and set it completely innocently and he went and found me a job as undergraduate and as a result of that I started working with graduate students and that got me interested in going to grad school and I talked to my wife into well master's degree seems pretty cost-efficient and then I was in a room with four students the other three were all getting PhD so that seemed like a good idea but if had that you know faculty member and had on his own found me a job I'm sure I wouldn't go to grad school because I had no plans to do it so it changed my life and where was this at use UCLA University of California Los Angeles and who's the well he was a just a PhD students time because his name's John Lou bear and he became a professor at the University of Washington it's just after he helped me out and he we had to hang around UCLA for a while but while he was looking for a job so he's teaching that course now I know that you you were at Berkeley for a long time yes yes 40 years how did you approach teaching and mentorship and promising in students well I guess you know I felt it was it dependent you know for undergraduates you know they needed more direction and I of course looked for opportunities to return the favor that happened to me to get them involved with research for graduate students I consider them kind of like young colleagues Ryan just there they're clearly very smart if they get into graduate school at Berkeley and just don't have a lot of experience and so my idea was what I tried to do is create research projects that if I was a student I just died at work on them I make it and what I would think would be an incredibly exciting project from the from the students perspective and then get them involved and let them find their passion and what they're gonna work on I'd love to ask you how the people here the HLF have been but of course you've only been here for a short time yeah so they it's seem interesting people certainly excited about what they're working on how do you encourage excitement well I think I'm kind of naturally a cheerleader and an optimistic person and kind of a lot of energy so I think I can I get I inspire people to want to work on things I think that's one of my part of my skill set I was a I did sports I still do a lot of sports and so I kind of use the metal model of the coach who's there to help you help young people achieve to the best of their abilities but that's what I think of myself now the environment that they're in is very different from the one that you were in certainly as far as the content and perhaps you mean the HLF students sir I mean in general anybody of that age who's entering computer architecture for example right what do you think are the most important differences between then and now between when I started and now as I'll say my lecture the biggest thing is the end of Moore's law is for 50 years there's been Moore's law with the number of transistors would double every year or two so that's you know you were it was like skeet shooting right you're tried to lead a couple years that when the technology is gonna be available which it was a if you design for today it'd be antiquated in a couple of years so you have to guess where the technology is going to be but shockingly that's over in fact when I say that on Thursday people aren't gonna believe me because they've heard people say Moore's law is over but it's it's over so we can't count on that anymore and so we're gonna have to rely more on computer architects to come up with new ideas how to do better computers even though the transistors are happen aren't gonna get much better so this is going to be even a greater challenge going forward I know I'm probably asking you to repeat things you're going to give in your talk but this won't come out until after your talk show I hope you don't feel so what do you think are the the most promising areas to follow yeah well since Hennessy and I just got the award I guess six months ago we wanted to make that part of our Turing lecture is to lay out challenges as an opportunity so the challenges are the the ending of Moore's Law there's another thing called Dennard scaling where you know it used to be that you can make put more transistors in because they didn't use more energy the voltage would drop and they could hit that in but that's over so power is now a limit so two big things we counted upon or over the opportunities we think are we've got these new modern programming languages like Python which are really productive for programmers but they're really inefficient on hardware so doing some kind of compiler hardware innovations to make these modern programming language what's faster seems like a big opportunity in another thing that we've had kind of run out of ideas given the end of Moore's law and giving it a Dennard scaling how to make general-purpose computers go faster that they're hardly improving at all you know 10 15 years ago they were doubling in performance every 18 months thanks to new ideas and Moore's Law and people would throw away their laptops not that they were broken it's just they were so slow relative to their friend's laptop they didn't want to use it anymore well that that yours long long past us so as far as architects are concerned the only way forward of making hardware faster is to narrower the domain that they work on so not general-purpose processors but domain-specific architectures is the phrase is being used what sort of applications do you think are most appropriate for for such architectures the one that's getting the most excitement right now is machine learning neural networking so that particular domain everybody's talking about I was talking to a PhD candidate who who was this is a bunch of conferences and no matter what
conference he went to computer architecture theory the most excitement at the session was the machine learning session so for all fields people are excited about learning about for its potential since you machine learning has a tremendous appetite for computing what the issue is going forward and CPUs aren't getting much faster we're gonna need new architectures for machine learning so it's a very exciting area right now optimized arced a hardware hardware optimized for machine learning to accelerate the rate of late machine wow that actually answers the next thing that I was thinking of asking which is what do you think of the what well what are the things in the next five years but actually in the last five years what do you think is has had the greatest impact aside from the ending of Moore's Law no I would say in the last five years I don't know that there's been it's been more kind of the ending of things than rather some great new thing that's happened well I know let me change that answer I think that we were doing at Berkeley another thing I'm going to talk about is a brand new thing for computer architecture which is to borrow ideas of the open source movement of machine from like a Linux operating system so so far all of the what's called the instruction sets this vocabulary that I talked about early have been proprietary so the most popular are owned by Intel and arm the new idea is to develop a RISC instruction set that nobody owns that anybody can use and so that all companies can embrace so this idea has been around about three or four years now and so we started up like that Linux Foundation it's called risk five for the fifth project those are risk five foundation and this has hundreds of you know 200 members now so it's very rapidly increasing in popularity so there's a lot of excitement about the idea of an open architecture like there are open operating systems do you see interest in following this among among young researchers yes particularly it's there's a tremendous excitement in China around this five there's interest in the security community one of the advantages that
risks five since it's open is anybody can use it and you can make changes to it and put it out there and yourself so in the past when there with only proprietary instruction sets you have to kind of beg are more Intel to put your ideas in but with risk five you could do it yourself put it up on the network and see how well they work because it has a real software stack and you know you're gonna get users and everything so it's really dramatically increasing the
number of people who can do computer architecture research you don't have to work with it with one or two companies now now I'm going to move again away from your working back to back to the students here in mentorship and and that sort of thing and I know this is your first time so it's sort of hard to gauge but what do you hope that they get from you get from me well the computer architecture is an exciting area that it's you know there's been some times recently where things seem to be moving
pretty slowly that you know that industry was where you had to work in industry to be able to affect what was going on and the fact that it's really blossomed and unohana seen and I the title of our during lecture is a new golden age we think it's going to be another exciting era from in computer architecture and if you you know even if
you're outside of computer architecture if you need things to go faster to be more energy-efficient you're gonna need to work with computer architects to deliver that because in Moore's laws no more is there anything else that you'd like to add a question you think should have been asked but wasn't first question that should have been asked [Music] I'm not I don't know I guess interests oh and you did a good job oh thank you very much and thank you again for making time I hope you enjoy the the conference yeah thank you [Music]


  361 ms - page object


AV-Portal 3.20.1 (bea96f1033d39fbe77f82542458e108105398441)