 |
Course Description
Principles and practice of parallel computing. Parallel algorithm development
and analysis. Design, implementation, and evaluation of parallel programs for shared
memory, local memory, and vector architectures. Prerequisites: Admission to MS in CS program,
or consent of department, plus CS 3270 or equivalent.
Course Syllabus
- Overview of parallel computing
- Performance measures
- Parallel architectures
- Problems amenable to parallel programming solution
- Programming languages for parallel programming
- Program portability issues
- Operating system issues
- Tools for parallel programming
- Parallel Algorithms
- Parallelizing serial programs
Knowledge Acquisition Goals After successfully completing CS 5170 a student should
have a conceptual understanding of:
- Flynn's Taxonomy
- Languages for parallel computing, including: MPI and OpenMP
- The Message passing paradigm
- Parallel Processing speedup issues, including Amdahl's and Gustafason's Laws
- Decomposition methodologies for parallel program development
- Load balancing issues for parallel programs
- Parallel Architectures and effect of architecture on design/implementation of a parallel
algorithm
- Quantifying Speedup; empirical tools and theoretical models
- PRAM computational model
- Current issues in parallel processing
Skill Goals After successfully completing this course a student should be able to
- Design, implement, test and debug a parallel application program using MPI
- Design, implement, test and debug a parallel application program using OpenMP
- Parallelize an existing application using an appropriate parallel programming paradigm
- Explain, in writing, the tradeoffs that result from using a specific programming paradigm for a
given problem class
- Develop and analyze a parallel algorithm using the PRAM model.
Course Requirements Examinations, homework assignments, programming assignments.
|
