Deakin home > Deakin University Library > Deakin Research Online > The impact of under-estimated length of jobs on EASY-backfill scheduling

The impact of under-estimated length of jobs on EASY-backfill scheduling

Wong, Adam K. L. and Goscinski, Andrzej M. 2008, The impact of under-estimated length of jobs on EASY-backfill scheduling, in PDP 2008 : Proceedings of the 16th Euromicro Conference on Parallel, Distributed and Networked-based Processing, IEEE Computer Society, Piscataway, N. J., pp. 343-350.

Attached Files (Some files may be inaccessible until you login with your Deakin Research Online credentials)
Name Description MIMEType Size Downloads

Title The impact of under-estimated length of jobs on EASY-backfill scheduling
Author(s) Wong, Adam K. L.
Goscinski, Andrzej M.
Conference name Euromicro Conference on Parallel, Distributed and Network-based Processing (16th : 2008 : Toulouse, France)
Conference location Toulouse, France
Conference dates February 13-15 2008
Title of proceedings PDP 2008 : Proceedings of the 16th Euromicro Conference on Parallel, Distributed and Networked-based Processing
Editor(s) El Baz, Didier
Bourgeois, Julien
Spies, Francois
Publication date 2008
Conference series Euromicro Conference on Parallel, Distributed and Network-based Processing
Start page 343
End page 350
Publisher IEEE Computer Society
Place of publication Piscataway, N. J.
Summary The issue of under-estimated length of jobs (parallel applications) on backfill-based scheduling is ignored in the current literature because users want to avoid their jobs to be killed when the requested time expires. Therefore, users prefer to over-estimate the length of their jobs. This paper shows the impact of underestimated length of jobs on their execution performance in an EASY-backfill scheduling-based system. We have developed a batch job scheduler for Linux clusters that implements an enhanced EASY- backfilling algorithm in such a way that a job with an under-estimated execution time would not be killed unless it would delay other jobs. We have carried out performance evaluation by scheduling static workloads of well known MPI parallel applications on a real cluster. Our results show that most of the jobs do not have to be aborted even though their job lengths are under-estimated whereas the slowdown of jobs and the throughput of the system are only slightly degraded.
ISBN 9780769530895
0769530893
ISSN 1066-6192
Language eng
Field of Research 080307 Operating Systems
HERDC Research category E1 Full written paper - refereed
Copyright notice ©2008, IEEE
Persistent URL http://hdl.handle.net/10536/DRO/DU:30018074

Document type: Conference Paper
Collection: School of Engineering and Information Technology
Connect to link resolver
 
Unless expressly stated otherwise, the copyright for items in Deakin Research Online is owned by the author, with all rights reserved.

Versions
Version Filter Type
Access Statistics: 338 Abstract Views, 0 File Downloads  -  Detailed Statistics
Created: Fri, 14 Aug 2009, 14:04:32 EST