Experiment-based project in undergraduate computer architecture

Abstract

The project presented in this paper has been developed for an advanced undergraduate computer architecture course specifically aimed at non-engineering students. The project is intended to develop understanding of issues in modern computer architecture through empirical testing and experimentation, using sophisticated execution time simulation and industry standard benchmark programs. The students are lead through the functional design process of a superscalar pipelined processor step by step, in a series of carefully structured experiments. The experiments either provide direct support for the design decisions, or make students explore a specific concept. The main goal of the project is to give students a true insight into the current technology and the driving forces behind modern computer architecture, without the burden of engineering and implementation details.

Core empirical concepts and skills for computer science

“…[12] In recent years, several articles have appeared in the literature advocating the use of experimentation in computer science courses. However, these articles tend to involve applying empirical reasoning to a problem in a particular course, such as operating systems [7,15], organization/architecture [1,4,17], software engineering [3,9], or human-computer interaction [6]. While these examples demonstrate the utility of experimentation, the broader issue that must be addressed is the need for a systematic approach to developing empirical skills across the computer science curriculum.…”