Donald Bren School of Information and Computer Sciences

In interactive behavioral synthesis, the designer can control the design process at every stage, including modifying the schedule of the design to improve its performance. In this report, we present a methodology for performance optimization in interactive behavioral synthesis. Also proposed in this report are several quality metrics and hints that can assist the user in utilizing the proposed methodology. When the user is optimizing the performance of the design, one important decision is the selection of a clock period. We have developed an algorihm to estimate the effect of different clock periods on the execution time of the design. This algorithm can be used to facilitate clock period selection by the user in order to optimize the performance of the design. We have tested our methodology on several benchmarks. The experimental results support the proposed methodology on several benchmarks. The experimental results support the proposed methodology by demonstrating an average improvement of 46.2% in design performance.