A Power-aware Job Scheduling Algorithm

Author

Haitao Chen ; Yutong Lu ; QingHua Zhu

Author_Institution

Sch. of Comput., Nat. Univ. of Defense Technol., Changsha, China

fYear

2012

fDate

22-24 Nov. 2012

Firstpage

8

Lastpage

11

Abstract

A power-aware job scheduling algorithm is proposed to increase the energy efficiency of high performance computer by selectively hibernating some idle nodes during periods of low load. The proposed algorithm records the switching logs of nodes´ power states, dynamic setting idle nodes to sleeping status. Detailed experimentation using traces from the Parallel Workloads Archive indicates that the proposed algorithm can achieve effective energy savings with good control on the switching frequency of nodes´power states.

Keywords

power aware computing; processor scheduling; Exascale system; HPC technology; dynamic setting idle nodes; effective energy savings; energy efficiency; high performance computer; parallel workloads archive; power efficiency ratio; power states; power-aware job scheduling algorithm; sleeping status; switching frequency; switching logs; Algorithm design and analysis; Computers; Energy consumption; Heuristic algorithms; Optimization; Switches; Switching frequency; job scheduling; power-aware; resource allocation; sleep; wakeup;

fLanguage

English

Publisher

ieee

Conference_Titel

Cloud and Service Computing (CSC), 2012 International Conference on

Conference_Location

Shanghai

Print_ISBN

978-1-4673-4724-2

Type

conf

DOI

10.1109/CSC.2012.9

Filename

6414471