一种基于可测性和低功耗的高层次综合方法

High-level synthesis method based on testability and low power

提出了一种基于遗传算法的多目标优化高层次综合方法。该方法在时间和面积约束下,通过高层次调度和模块分配,对可测性和功耗问题进行研究。给出一种可同时进行调度和模分配的编码方法,并设计了相应的遗传算子,避免了进化过程中不可行解的产生。实验证明了该方法在可测性改善和功耗优化方面的有效性。

This paper proposed a high-level synthesis method for multi-objective optimization based on genetic algorithm. This method studied the testability and power problems through high-level scheduling and module allocating under time and area constrained. The main contributions were that it proposed a novel coding method could be used for scheduling and module allocating simultaneously and designed corresponding genetic operators avoiding the generation of infeasible solutions. The efficiency of testability improvement and power optimization is demonstrated by experimental results.