University of Houston
Department of Computer Science

In partial fullfillment of the Requirements for the Degree of
Ph. D. of Science


Ruijian Zhang
will defend his dissertation

A HYBRID BRANCH PREDICTION METHOD TO IMPROVE INSTRUCTION LEVEL PARALLELISM


Abstract

Providing accurate branch prediction is critical to effectively exploit instruction level parallelism. Some of the existing branch prediction schemes are static in that compilers use opcode information and profiling statistics to make predictions. Others of the existing branch prediction schemes are dynamic in that the hardware uses run-time execution history to make predictions. In this research, we propose a hybrid branch prediction method that combines an improved static branch prediction scheme and a new dynamic branch predictor. To implement this hybrid branch prediction scheme, an intelligent compiler applies techniques of artificial intelligence, including knowledge base system, machine learning and evolutionary programming. The dynamic branch predictor consists of a "Switch-Counter" and a modified two-level adaptive branch predictor. Preliminary simulation results indicate that this approach is promising.


Date: January 31, 2000
Time: 10:30 AM
Place: 550-PGH


Faculty, students, and the general public are invited
Thesis Advisor: Dr. W. K. King