Static and Dynamic Characterization of C Programs
The goal of this project is to develop tools for characterization of C programs. Such tools should analyze one or more actual programs and generate a set of quantitative parameters characterizing their workload.
The measured quantities can later be used for cluster analysis of workloads or as inputs for other tools, such as automatic generators of benchmark programs. One of our goals is to investigate the area of "program cloning" - the generation of artificial programs of arbitrary size that have characteristics similar to some real programs.
Hue-Man Tang received her MS degree in Computer Science from San Francisco State University. Her interests are in computer performance evaluation and software metrics.