Machine Learning based Temperature Estimation for Test Scheduling of 3D ICs

Abstract

Three-dimensional integrated circuit (3DIC) is an emerging technology with significant benefits over conventional 2DIC. However, 3D ICs face several challenges. This work presents a thermal-aware test scheduling technique for 3DICs. A machine learning based thermal estimation method for 3D ICs has been proposed in this work. The proposed temperature prediction technique has been used to generate thermally safe test schedules for 3DICs. The proposed techniques implemented on ITC’02 benchmark circuits has shown promising results.