Energy-efficient Backfill-based Scheduling Approach for SLURM Resource Manager

With the rapid development of information technologies and growing amount of data that needs to be processed, the problem of energy efficiency of computing systems is becoming more and more topical. Energy aware scheduling is one of the possible ways to reducing power consumption of such systems. In this paper it is proposed to enhance one of the most popular workload managers SLURM with energy aware scheduling approach PCPB (Power Consumption and Performance Balance). Enhanced approach allows to find the balance between energy efficiency and performance of data processing, while saving hardware resources. An approach shows the gain in power saving and data processing performance as well in comparison with widely used scheduling algorithms that were considered.