NBPE: Neural network based power estimation simulator for specification design

Abstract

This paper forwards a neural network based VLSI power estimation Simulator (NBPE) for VLSI specification design with a graphical user interface developed. The user can enter parameters from VLSI specification such as IO number, frequency, flash depth and parameters on neural network structure such as layer number, learning algorithm etc. This simulator then estimate VLSI's power based on given information. Power estimation results can be used to adjust VLSI specification design before further implementation. Available as an executable code, the simulator can run on various platforms.