Comments on Development-Oriented vs Basic Research

1
Comments on Development-Oriented vs Basic
Research

• Prof. Jason Cong
• cong_at_cs.ucla.edu
• Computer Science Department
• University of California, Los Angeles

2
Is There Basic Research In Computer Science

• What is basic research?
• Understanding the nature and discovering the
  ultimate truth
• Without consideration of practical implication or
  commercial impact
• With exception of theory, most of computer
  science is applied research, not basic research
• But there is a lot to leverage from basic
  research
• Examples from my research areas in computer-aided
  design

3
Characterization of Programmable Logic Blocks for
Lookup-Table (LUT) based FPGAs (Field-Programmable
Gate-Arrays)
Each box is a k-input look-up table, can
implement of any Boolean function up to k inputs
4
Use of Theory of Functional Decomposition

• Wed like to answer the questions
• f(x1,x2,,xn) g(y1(x1,,xk),,ym(x1,,xk),xk
  1,xk2,,xn))
• Ashenhurst Decomposition (1959)
• Roth-Karp Decomposition (1962)
• Recent improvement using Binary Decision Diagram
  (BDD)

5
mPL produced best result at ISPD2006 Placement
Contest
6
Multilevel Optimization Framework

• Successful in many applications
• Originally developed for partial differential
  equations (PDEs)
• Recent success in VLSI CAD partitioning,
  placement, routing
• Collaboration of Tony Chan (UCLA Math)

7
How to Best Leverage Basic Research ?

• Inter-discipline research is very important
• Are we aware whats going on in basic research?
• How and who will bridge the gap?
• A broad education of students is very important
• Be able to understand and appreciate basic
  research

8
Short-Term vs. Long-Term Applied Research

• Need to focus on long-term applied research
• 5 -15 years beyond immediate application
• Pre-competitive
• Example the invention of Internet
• 1st node of arpa-net built at UCLA in 1969
• It is not easy to identify long-term research in
  an exponential growing field
• Law of exponential power Every generation of
  advancement is larger than the sum of all
  previous generations
• What takes to do long-term applied research
• Vision and good taste
• Sufficient funding
• Reward and understanding of risk taking