Advances of optimization algorithm via neural network computing for EDA
In response to the increasing complexity of chip design,EDA tools and methods are also evolving.However,EDA needs to be coordinated to achieve optimal power,performance,and area,and it does not always guarantee an optimal solu-tion.The application of EDA tools in the circuit design stage,including logic synthesis,layout and verification,belongs to the nonlinear programming solution process with multiple objectives and constraints.To better address the uncertainties of the solu-tion and the problems such as the easy to appear local extreme values,optimization algorithms based on neural network had been integrated into the design process of EDA tools.This paper first gave a brief overview of the optimization problem,multi-objective optimization calculation and optimization algorithm based on neural network in EDA,and then sorted out the optimi-zation solution methods of optimization algorithm based on neural network in different design stages such as logic synthesis,layout and verification,and expounded on the challenges and opportunities faced by the current research institute.It hoped to provide reference for automated integrated circuit design and related research.
万方数据
维普
