Final Study Guide
Parallel Computing I - Topics 20032
- General
- Why parallel computing now? (Reference Lester/Wilson/Barry and others)
- Hardware Issues
- Flynn's Taxonomy
- Global vs Distributed Memory
- Communication/memory networks
- Granularity (# of processors)
- Software Issues
- Mapping Algorithms to Architecture
- Languages
- Compilers (Automatic)
- Operating Systems
- General Purpose
- New Methods (algorithm design)
- Correctness/debugging
- Granularity (problem decomposition)
- Performance Issues
- Granularity
- Load Balancing
- Parallel Constructs vs Architecture
- Computation vs Communication
- Parallel Languages
- Parallaxis
- Linda
- MPI
- FORTRAN (automatic vs subroutine calls vs compiler directives)
- OpenMP
- Issues
- Implicit Parallelism
- Explicit Parallelism
- Parallel Languages
- Portability vs Efficiency
- Metrics (references Quinn/Lester/Wilson)
- Speedup
- Efficiency
- Amdahl's Law
- Cost
- Network Topologies
- Topologies
- Why hypercube was so popular
- Diameter
- Grey Codes
- Dynamically Configurable Networks
- How relates to issues with memory
- Parallel Algorithms/Applications
- Designing Parallel Algorithms (Reference Quinn)
- From SIMD to MIMD
- Odd-Even Transposition Sort
- Parallel Prefix (Didn't cover)
- 1-D Wave Equation (Reference - Fox etal.)
- Communication Template
- Communication Overhead
- Problem Decomposition
- Problem Dimensionality and Performance Metrics
- Matrix Multiplication (Reference - Quinn)
- Parallelization of iterative techniques (Reference - Bertsikas, etc.)
- Parallel Sorts
- Parallel graphics
- FFT's (didn't cover)
- Heterogeneous and Cluster Computing
- Future Parallel Computers?
- Quantum Computing
- DNA Computing
- Optical Computing
Nan C. Schaller
Rochester Institute of Technology
Computer Science Department
102 Lomb Memorial Dr.
Rochester, NY 14623-5608
telephone: +1.585.475.2139
fax: +1.585.475.7100
e-mail:
ncs@cs.rit.edu
November 24, 2003