A Wafer-scale heterogeneous integration thermal simulator

The resurgence of wafer-scale heterogeneous integration necessitates efficient simulation methodologies to enable thermal-aware design space exploration, particularly for iteration-based algorithms. Existing finite element method based models are computationally prohibitive, and published finite difference method based models face scale limitations. To address these issues, this paper proposes a new finite difference method based thermal simulator, WSHITS (wafer-scale heterogeneous integration thermal simulator), equipped with an automatic wafer edge approximation algorithm. Additionally, the airflow and heatsink design effects are considered to investigate the impact of forced convection on the thermal performance. Compared with finite element method based approaches, WSHITS achieves significant performance improvements, including a 2.78x reduction in cell count and a 25.9x speedup in simulation runtime, while maintaining high accuracy with a relative error of only 0.38%. By offering flexible cell granularity and superior simulation efficiency, WSHITS is promising as an efficient thermal simulation tool in thermal-aware design space exploration, enhancing the thermal design quality of wafer-scale heterogeneous integration.