NSF/TCPP Curriculum Planning Workshop

1
NSF/TCPP Curriculum Planning Workshop

• Joseph JaJa
• Institute for Advanced Computer Studies
• Department of Electrical and Computer Engineering
• University of Maryland

2
Recent Teaching Experience

• Machine Learning and Data Mining graduate and
  undergraduate
• VLSI Architectures graduate
• Data Structures and Algorithms graduate
• Probability undergraduate
• Parallel Algorithms - graduate

3
Overview of VLSI Architectures

• Techniques for mapping algorithms onto parallel
  architectures with signal processing computations
  as the main driving applications.
• Pipelining and parallelism retiming and loop
  unrolling mapping algorithms into systolic
  architectures etc.
• Target architectures DSPs VLIW FPGAs
  GPUsmulticore etc.

4
Advanced Course in Parallel Models and Algorithms

• Data Parallelism
• Model parallel time and total work
• Examples from image processing prefix sums
  matrix computations FFT graph algorithms etc.
• Multithreading
• Shared Memory Model synchronization costs
• Examples prefix sums matrix computations FFT
  etc.
• Distributed Memory
• Model message passing communication costs
• Broadcast, reduce, all-to-all operations
• Examples matrix computations FFT sorting etc.