A Machine Learning Approach for Power Consumption Forecasting in VLSI Circuits

Abstract

The purpose of this study is to provide a complete investigation on the prediction of power consumption in VLSI  circuits by using techniques of machine learning. By conducting an analysis of previous design data, we develop  and test prediction algorithms that assist designers in making early decisions that are more in line with power  efficiency. Our results, which give large increases in prediction accuracy, reveal promising paths for future  optimization of VLSI designs. These possibilities are emphasized and highlighted by our findings.