Experiences In Parallel/Distributed Computing For Undergraduate Computer Science Majors

Abstract

The increasing prevalence of multiprocessor and distributed systems in modern society is making it imperative to introduce the underlying principles of parallel/distributed computing to students at the undergraduate level. In order to meet the needs of our students for training in this critical area, the Computer Science Department at Southern Connecticut State University (SCSU) is currently in the process of implementing a curricular and laboratory development project that integrates key concepts and practical experiences in parallel computing throughout the undergraduate curriculum. The goal of this project is to build a strong foundation in parallel computing which would optionally culminate in advanced, senior-level specialized courses in parallel computing and/or senior research projects.

This paper describes the laboratory facility we developed to support instruction in parallel and distributed computing and the parallel computing modules which were incorporated into three of our core undergraduate courses: data structures, operating systems, and programming languages. The laboratory facility enables us to provide our students with “hands-on” experiences in shared memory, distributed memory, and network parallelism. The modules and laboratory exercises give students the opportunity to experiment with a wide array of software and hardware environments and to gain a systematic exposure to the principles and techniques of parallel programming.

The major objective of the curricular and laboratory development project in parallel/distributed computing at Southern Connecticut State University is to make every computer science major aware of the basic hardware models, fundamental concepts, and programming languages and tools relevant to parallel/distributed computing. Support for this project from the National Science Foundation has enabled the Computer Science Department to obtain dedicated parallel computer systems and proprietary parallel computing software. The computer network provides students with the ability to experiment with various parallel computing models both locally and remotely.

At the present time, parallel/distributed computing laboratories have been designed and assigned to students in three courses: data structures, operating systems, and programming languages. Typically, students obtain their first experience in parallel computing in the data structures course. This is followed by a more in-depth experience in the operating systems course in which students \’, ”xc,,, ~~~~” 1996 ASEE Annual Conference Proceedings ‘.@lye,: