Thermal Reliability Modeling And Optimization Of 3D IC

In microelectronics,a three-dimensional integrated circuits(3D IC)is an new integrated circuit manufactured by stacking silicon wafers and/or dies and interconnecting them vertically using through-silicon vias(TSV)so that they behave as a single device to achieve performance improvements at reduced power and smaller footprint than conventional two dimensional processes.But every advantage has its disadvantages,the same is true for 3D IC.Heat dissipation and the related thermal-mechanical stress problems are the major obstacles in the development and commercializing process of 3D IC.Dynamic thermal management(DTM)techniques can be used to alleviate such problems and enhance the thermal reliability of 3D IC.However,the time varying stress information is hard to obtain at runtime which limits the effectiveness of DTM.In this thesis,a fast stress analysis method for runtime usage was proposed.The new method builds artificial neural network(ANN)model by training offline using thermal and stress data.Next,the ANN model is used to generate important stress information,such as maximum stress around each TSV,for DTM methods at runtime.In order to improve the stress estimation accuracy and speed,specially designed input selection plans are proposed and implemented for ANN model generation.In addition,a necessary 3D IC model is built and simulated in a finite element software to generate the offline data which is needed for artificial neural network training.Experiments with different configurations of ANN models show that the new method is able to estimate important stress information at extremely fast speed with good accuracy for runtime 3D IC reliability enhancement usage.