University of Houston
Department of Computer Science

In partial fulfillment of the Requirements for the Degree of
Doctor of Philosophy


Oscar Hernandez
will defend his dissertation proposal

A feedback-based tool for shared memory program optimization




Abstract

A program analysis tool can play an important role in helping users understand and improve OpenMP codes. An analysis and feedback based tool that exploits the Open64 compiler can be helpful to provide static as well as dynamic (feedback-based) information which can be used to develop or optimize OpenMP codes. The tool should help the users to visualize and print essential program structures and obtain runtime information on their applications. On the other hand the tool should be helpful to optimize memory accesses by minimizing the data sharing among threads, either by rearranging code or by privatizing data structures where possible. Proposing a system that supports thread level array region analysis feedback can help the programmer restructure their code to maximize data locality, reduce false sharing or achieve aggressive privatization.

Date: Monday, August 30, 2004
Time: 4 PM
Place: 550-PGH



Faculty, students, and the general public are invited.
Thesis Advisor: Dr. Barbara M. Chapman