Job Scheduling in a Computational Cluster with Multicore Processors
Trần Thị Xuân, Tien Van Do · Advances in intelligent systems and computing · 2016
In this paper, we investigate a job scheduling problem in a heterogeneous computing cluster built from servers with multicore processors. Dynamic Power Management technique is applied, where the delay to bring a server from the sleep to the active state is taken into account. Numerical results show that the computing resources are utilized more efficiently if multi jobs are executed in parallel. Furthermore, scheduling policy should investigate machine parameters at core-level to achieve the best efficiency.