CS 545 Parallel Computing System
11:00 - 11:50 MWF
EPS 350
Instructor: Year-Back Yoo
EPS 359, x3541, yoo@cs.montana.edu
Office Hour: 9-11 MW
Text: Lecture notes. (No textbook)
References: Introduction to Parallel Computing, 2nd Ed
Grama, Gupta, Karypis, and Kumar, Addison Wesley, 2003
Parallel Computing: Theory and Practice, 2nd Ed
Quinn, McGraw Hill, 1994
The Design and Analysis of Parallel Algorithms
Akl, Prentice-Hall, 1989
Algorithms: Sequential, Parallel, and Distributed
Berman and Paul, Thomson, 2005
Requirements
· Assignments 40
o Seminar (Tutorial) 30
· Project 80'
· Tests 100
Project (80' pts)
Implementation of a parallel algorithm for the problem you choose.
Follow the following guideline.
1. Journal Survey
2. Design of parallel algorithm for the problem.
3. Implementation of the algorithm on a parallel computing system
4. Experiments with various input sizes (Use benchmark, if available)
5. Report from he the experiment (Include discussion and conclusion).
Submit the report by November 30, and give a presentation in December.
The evaluation criteria:
1. Extent of the journal survey
2. Efficiency of the parallel algorithm designed.
3. Effectiveness of the experimental design
4. Computational results
5. Report quality
6. Presentation
Part I. Foundation
1. Introduction
Motivation
History
Performance metric – speedup, efficiency
Model of parallel computation
RAM (Shared memory)
Distributed Memory (Message passing)
Design of Parallel Algorithm
2. Parallel Computer Architecture
o Taxonomy
o Processor Interconnection - Goodness measures
o Benchmarking - Top500
3. Mapping and Scheduling
o Mapping
o Scheduling
4. Parallel Programming Languages
o Parallel Program Constructs
o Parallel Programming Languages
Part II. Parallel Algorithms
5. Elementary Parallel Algorithms: sum, prefix sum, surfix sum, broadcast
6. Sorting and Merging
7. Matrix Algorithms
8. Graph and Network Algorithms
Part III. Special Topics (Seminar)
o Team A: Ant Colony Optimization
o Team B: Tabu Search
o Team C: Systolic Array
o Team D: Dataflow Machine
o Team E: Optical Computer/DNA Computer/Quantum Computer