Thermal-aware task scheduling for 3D-network-on-chip: A Bottom-to-Top scheme

3D-NoC emerges as a potential multi-core architecture delivering high performance, high energy efficiency and great scalability. However, severe thermal challenges due to high power density continue to be a major hurdle in the development of 3D-NoC. In this paper, we propose a novel thermal-aware task scheduling scheme named Bottom-to-Top (B2T) approach to address this challenge. Incorporating communication overhead into analysis based on task graph, this heuristic-based method judiciously performs task allocation on processing units to efficiently minimize the peak temperature and improve the execution time of the applications. Our simulation results demonstrate that our scheme can achieve significant peak temperature reduction (up to 7.95°C) and performance improvement (up to 4%) when compared to other methods.