GRAND CHALLENGES AND GREAT TEAMS

1
GRAND CHALLENGES AND GREAT TEAMS

• David Patterson
• University of California at Berkeley
• Eckert-Mauchly Award Lunch
• International Symposium on Computer Architecture
• Beijing, China
• June 24, 2008

2
Outline

• Systems Grand Challenges for 21st Century
• 20th Century Architecture Achievements
• Jim Grays Definition of a Grand Challenge
• What Next? A Few Remaining Problems in
  Information Technology, 1998 Turing Lecture
• 21st Century Architecture Resources
• 21st Century Teams
• Thanking My Teams

3
20th Century COMPUTER ARCHITECTURE RESEARCH
ACHIEVEMENTS

• Invent Stored Program Computer
• Built 1st Digital Computers ENIAC, EDSAC,
  UNIVAC,
• Start 1st Computer Companies
• Eckert-Mauchly Corporation
• Invent Memory Hierarchies
• Virtual Memory, Cache
• Computer Architecture vs. Implementation
• Instruction Set Architecture
• Vacuum Tubes, Magnetic Core, Magnetic Disks, VLSI
• Cost performance improved 100 Billion in 55 years
• Instruction Level Parallelism
• Pipelining, Vector, Branch Prediction,
  Superscalar, Out-Of-Order Execution,
  Speculation,

4
Properties of a Research Goal

• Simple to state.
• Not obvious how to do it.
• Clear benefit.
• Progress and solution are testable.
• Can be broken into smaller steps
• So that you can see intermediate progress.

http//research.Microsoft.com/Gray/talks/Turing2.
ppt
5
21st Century COMPUTER SYSTEMS GRAND CHALLENGES

• Computer systems that make it easy to write
  correct, portable, parallel programs that
• Run efficiently with 100s of processors per chip
• Scale as processors increase according to Moores
  Law
• Computer systems that never fail you
• Gray Nine 9s Only 1 second of outage per
  century
• Secure computer systems that protect privacy
• Computer systems that reduce absolute energy
  while improving cost-performance
• 1975 refrigerator 1800 kwh, 1200 ? 600 kwh,
  500, larger
• IT advances help big problems facing humanity
  global change, energy, hunger, poverty,
  education, disease,
• Will people still be here in 22nd century? Did IT
  help or hinder?

6
ARCHITECTURE RESEARCH and COMPUTER SYSTEMS GRAND
CHALLENGES

• Require computer architecture innovation, but
  cant be solved just by computer architects
• Need teams with experts in applications,
  programming, systems software, architecture and
  hardware
• 20th Century CS traditionally neglects
  applications
• Embrace applications experts in 21st Century

7
21st Century Computer Architecture RESEARCH
Opportunities

• 2010 to 2030 Few researchers can afford to build
  credible VLSI prototypes
• But ECAD allows design, layout, simulate VLSI
  components to estimate power, size, speed
• But FPGA HW design reuse, fast enough to run
  full SW stack (gt100 MHz), limited resources
• Cloud/Utility Computing ? Research at scale
• Open Source Software
• Industrial strength SW to measure and run
• Channel for innovative HW/SW systems

8
Grand Challenge Summary

• Great 20th Century Track Record
• Stored program, Memory hierarchy, ISA, ILP
• 5 New Challenges for New Century
• Parallel computers easy to program, scalable
• Never fails you (99.9999999 1 sec/century)
• Secure and protects privacy
• Reduce absolute energy while increasing MIPS/
• Leverage IT advances to help all of humanity
• Build prototypes with FPGAs, ECAD, OSS
• Multidisciplinary teams apps to HW

9
Outline

• Systems Grand Challenges for 21st Century
• 20th Century Architecture Achievements
• Three of Jim Grays Grand Challenges
• What Next? A Few Remaining Problems in
  Information Technology, 1998 Turing Lecture
• 21st Century Teams
• 21st Century Architecture Resources
• Thanking My Teams

10
Wrestling TeamsHigh School College
11
Reduced Instruction Set Computer (RISC)
1980-1984

• Carlo Sequin I
• Simpler better for Moores Law?
• 1984 consult for Sun gt SPARC
• 1988 sales gt1B
• Since 2004 ship gt1B embedded RISCs/yr
• Sponsor Defense Advanced Research Project Agency
  (DARPA)

12
RISC Colleagues Beyond Berkeley

• Joined RISC forces with IBM Stanford
• John Cocke (801) John Hennessy (MIPS)

Eckert Mauchly Award (1985)
Eckert Mauchly Award (2001)
13
RESOLVING RISC-CISC DEBATE

• In the USA, how do we resolve debates?
• We ask celebrities!
• Who is the biggest celebrity in the world?

14
RESOLVING RISC-CISC Debate

• Angelina Jolie as Kate Libby (aka as hacker Acid
  Burn) in movie Hackers (1995)

15
RESOLVING RISC-CISC Debate

• Angelina Jolie RISC architecture is gonna
  change everything.

Blue Man Group (silence)
16
HENNESSY AND PATTERSON TEAM

• TWO BOOKS, 8 EDITIONS TOTAL

17
Redundant Array of Inexpensive Disks (RAID)
1987-1993

• Randy Katz and I Use many PC disks to build
  better storage?
• RAID I built on 1st SPARC
• Sun 4/280 with 28 disks
• RAID II custom HW, 144 disks
• 1st Network Attached Storage
• Today, RAID 25B industry
• RAID students join industry and academia,
  started own companies (VMware, Panassas )
• RAID sponsors IBM, Sun, NSF, DARPA

18
Network of Workstations (NOW) 1993-1998

• Tom Anderson, Eric Brewer, David Culler, and I
  Build a supercomputer from standard desktops /
  switches?
• NOW-2 105 Sun Ultra-1 Workstations Myrinet
• Eric Brewer uses NOW to build Inktomi search
  engine startup
• Brewer pioneers 3-tier Internet service
  architecture
• Leads to clusters of 1000s of desktop computers
  at many Internet companies (Google, Ebay, Amazon,
  )
• Sun, HP, DARPA sponsor NOW

19
EECS _at_ UC Berkeley TEAMS
RISC
RAID
RAD Lab Par Lab
NOW
20
EECS _at_ UC Berkeley TEAMS

• RAD LAB 2006-2010
• If the datacenter is now the computer, what
  HW/SW lets one person design and efficiently
  operate software as a service?

RAD Lab

• Par Lab 2008-2013
• What HW/SW makes it easy to build correct,
  efficient, and scalable parallel programs?

Par Lab
21
COMPUTER INDUSTRY TEAMS
22
The Patterson TEAM

• Parents David and Lucie
• Our Children David and Michael
• Their wives Zackary and Heather
• Grandchildren Andrew, Grace, Owyn
• Siblings Linda, Don, Sue
• In Laws Lloyd, Pat, John
• Nieces Nephews Laura, John, Lyle, Alissa,
  Melissa, Bryan, and Paul

23
My Wife

• Linda Ann Crandall
• Met at 12 years old
• Stated dating at 16
• Married at 19
• Kids at 21

24
A TEAM IS SMARTER THAN ANY SINGLE GENUIS.
CHINeSE Proverb

• Blessed with great teammates
• My wife and family
• Students and colleagues at UC Berkeley
• Friends in industry and architecture community
• Wrestlers and coaches

25
Backup Slides
26
GRAYS List OF 12 GRAND Challenges

• Devise an architecture that scales up by 106.
• The Turing test win the impersonation game 30
  of the time.
• 3.Read and understand as well as a human.
• 4.Think and write as well as a human.
• Hear as well as a person (native speaker) speech
  to text.
• Speak as well as a person (native speaker) text
  to speech.
• See as well as a person (recognize).
• Remember what is seen and heard and quickly
  return it on request.
• Build a system that, given a text corpus, can
  answer questions about the text and summarize it
  as quickly and precisely as a human expert.
  Then add sounds conversations, music. Then add
  images, pictures, art, movies.
• Simulate being some other place as an observer
  (Tele-Past) and a participant (Tele-Present).
• Build a system used by millions of people each
  day but administered by a ½ time person.
• Do 9 and prove it only services authorized users.
• Do 9 and prove it is almost always available
  (out less than 1 second per 100 years).
• Automatic Programming Given a specification,
  build a system that implements the spec. Prove
  that the implementation matches the spec. Do it
  better than a team of programmers.