An Improved Energy-Efficient Scheduling for Precedence Constrained Tasks in Multiprocessor Clusters

Excessive energy consumption has become a critical issue in high performance computing. Task scheduling algorithms affect not only schedule length but also energy consumption. To shorten schedule length of parallel tasks with precedence constraints, scheduling algorithms could duplicate tasks on critical paths to avoid communication delay caused by inter-task dependence. However, task duplications incur more energy consumption. In this paper, we propose a heuristic Processor Reduction Optimizing (PRO) method to reduce the number of processors used to run parallel tasks, thereby decreasing system energy consumption. The PRO method can find appropriate time slots to accommodate tasks immigrated from low-utilized processors. The PRO method can be combined with existing duplication-based scheduling algorithms, such as Task Duplication Scheduling (TDS), Energy-Aware Duplication (EAD) scheduling and Performance-Energy Balanced Duplication (PEBD) scheduling. Experimental results show that the proposed PRO method can effectively decrease the number of used processors and save energy while maintaining schedule length.