Contemporary Computer Architectures
605.711

Course Description
This course provides a foundation for understanding and evaluating the design principles incorporated in contemporary computer systems. Although the field of computer architecture is constantly changing, this course stresses design ideas embodied in many machines, the techniques to evaluate these ideas, and the means to achieve balance and efficiency in the context of any device technology. The first part of the course provides the principles incorporated in single processor systems. Topics include memory system design, instruction and arithmetic pipelining, reduced instruction set versus complex instruction set architectures, and storage systems. The second part of the course presents basic techniques for designing multiprocessor and multicomputer systems. Lectures cover a variety of interconnection networks and the application of parallel architectures. The course concludes with new architectural ideas such as systems based on biological concepts and fuzzy set theory. Throughout the semester, basic principles are presented and illustrated with examples of existing computer systems and real applications with their architectural requirements.

Syllabus

1. Classification of Architectures
2. The Structural Levels of a Computer System
3. Memory System Design
4. Instruction Pipelining, RISC versus CISC Architectures
5. Arithmetic Pipelining
6. Storage Systems
7. Mid-term Exam
8. Interconnection Networks
9. Multiprocessor and Multicomputer System Overview
10. Multiprocessor and Multicomputer Architectures
11. Systolic Arrays
12. Data Flow Architecture
13. Future Horizons for Computer Architecture
14. Final Exam

Instructor
Robert Martino is Acting Scientific Director of the Division of Computer Research and Technology (DCRT) at the National Institutes of Health and Chief of the Computational Bioscience and Engineering Laboratory of the DCRT. In these positions, he is responsible for the application of high-performance computing and communications to biomedical problems including the areas of image processing, structural biology, computational chemistry, medical imaging, telemedicine, scientific visualization, signal processing, genetic database searching, and genetic linkage analysis. In addition, he serves as the NIH representative to the national, multiagency High Performance Computing and Communications program. He has been a member of the graduate school faculty of the Part-Time Programs of the Whiting School of Engineering at the Johns Hopkins University since 1982.

Dr. Martino received a B.S. from Northeastern University, Boston, in 1971 and an M.S. and a Ph.D. from the University of Maryland, College Park, in 1973 and 1982, respectively. All of his degrees are in electrical engineering. His recent publications include a paper in Science magazine entitled "Parallel Computing in Biomedical Research" and a book published by Kluwer Academic Publishers entitled High Performance Computational Methods For Biological Sequence Analysis.

Computer Lab Requirements
No specific computer requirements are necessary for this course.

Textbook
Computer Architecture, Single and Parallel Systems by Mehdi R. Zargham

Return to Computer Science Courses | Computer Science  | Part-Time Engineering

updated 8/19/1999