Energy and Thermal Aware Mapping for NoC Using Non-dominated Sorting Genetic Algorithm Ⅱ

摘要

In order to overcome the hotspot problem caused by local high temperature, based on the establishment of the hotspot temperature assessment model, a mapping method was proposed which could reduce communication energy and weaken hotspot temperature. A Multi-objective optimization model of communication energy and hotspot temperature is constructed, then insertion positioning code method and non-dominated sorting genetic algorithm Ⅱ (NSGA-Ⅱ) with elite strategy are applied to find its Pareto optimal set. The algorithm is evaluated for two real traffic scenarios. Experimental results verify correctness of the hotspot temperature evaluation model, confirm effectiveness of the method in terms of both optimizing energy consumption and thermal balance, and show the solution found by the method outperforms other mono-objective methods.