Abstract
Efficient job-scheduling strategies are important to improve the performance and usability of HPC-clusters. In this paper we evaluate job-scheduling strategies (FCFS, SJF, and LJF)used in the resource management system CCS (Computing Center Software). As input for our simulations we use two job sets that are generated from trace files of CCS. Based on the evaluation we introduce the dynP scheduler which combines the three scheduling strategies and dynamically changes between them online. The average estimated runtime of jobs in the waiting queue and two bounds are used as the criterion whether or not dynP switches to a new policy. Obviously the performance of dynP depends on the setting of the two bounds. Diverse parameter pairs were evaluated, and with the proper setting dynP achieves an equal or even better (+9%) performance.
This is a preview of subscription content, log in via an institution.
Buying options
Tax calculation will be finalised at checkout
Purchases are for personal use only
Learn about institutional subscriptionsPreview
Unable to display preview. Download preview PDF.
References
A. Mu’alem and D. Feitelson. Utilization, Predictability, Workloads, and User Runtime Estimates in Scheduling the IBM SP2 with Backfilling. In IEEE Trans. Parallel & Distributed Systems 12(6), pages 529–543, June 2001.
A. Streit. On Job Scheduling for HPC-Clusters and the dynP Scheduler. TR-001-01, PC2-Paderborn Center for Parallel Computing, Paderborn University, July 2001.
D. Feitelson and A. Weil. Utilization and Predictability in Scheduling the IBM SP2 with Backfilling. In Proceedings ofthe 1st Merged International Parallel Processing Symposium and Symposium on Parallel and Distributed Processing (IPPS/SPDP-98), pages 542–547, Los Alamitos, March 1998. IEEE Computer Society.
D.G. Feitelson and L. Rudolph. Metrics and Benchmarking for Parallel Job Scheduling. Lecture Notes in Computer Science, 1459:1–24, 1998.
D.G. Feitelson and M. Naaman. Self-Tuning Systems. In IEEE Software 16(2), pages 52–60, April/Mai 1999.
F. Ramme and T. Romke and K. Kremer. A Distributed Computing Center Software for the Efficient Use of Parallel Computer Systems. Lecture Notes in Computer Science, 797:129–136, 1994.
H. Franke and J. Jann and J. Moreira and P. Pattnaik and M. Jette. An Evaluation of Parallel Job Scheduling for ASCI Blue-Pacific. In Proceedings of SC’ 99, Portland, Oregon, pages 11–18. ACM Press and IEEE Computer Society Press, 1999.
J. Gehring and F. Ramme. Architecture-Independent Request-Scheduling with Tight Waiting-Time Estimations. Lecture Notes in Computer Science, 1162:65–80, 1996.
A. Keller, M. Brune, and A. Reinefeld. Resource Management for High-Performance PC Clusters. Lecture Notes in Computer Science, 1593:270–281, 1999.
A. Keller and A. Reinefeld. CCS Resource Management in Networked HPC Systems. In Proc. of Heterogenous Computing Workshop HCW’98 at IPPS, Orlando, 1998; IEEE Computer Society Press, pages 44–56, 1998.
D.A. Lifka. The ANL/IBM SP Scheduling System. Lecture Notes in Computer Science, 949:295–303, 1995.
J.E. Moreira, H. Franke, W. Chan,and L.L. Fong. A Gang-Scheduling System for ASCI Blue-Pacific. Lecture Notes in Computer Science, 1593, 1999.
J. Skovira, W. Chan, H. Zhou, and D. Lifka. The EASY-LoadLeveler API Project. Lecture Notes in Computer Science, 1162:41–47, 1996.
D. Talby and D.G. Feitelson. Supporting Priorities and Improving Utilization of the IBM SP2 Scheduler Using Slack-Based Backfilling. TR 98-13, Hebrew University, Jerusalem, April 1999.
Author information
Authors and Affiliations
Editor information
Editors and Affiliations
Rights and permissions
Copyright information
© 2001 Springer-Verlag Berlin Heidelberg
About this paper
Cite this paper
Streit, A. (2001). On Job Scheduling for HPC-Clusters and the dynP Scheduler. In: Monien, B., Prasanna, V.K., Vajapeyam, S. (eds) High Performance Computing — HiPC 2001. HiPC 2001. Lecture Notes in Computer Science, vol 2228. Springer, Berlin, Heidelberg. https://doi.org/10.1007/3-540-45307-5_6
Download citation
DOI: https://doi.org/10.1007/3-540-45307-5_6
Published:
Publisher Name: Springer, Berlin, Heidelberg
Print ISBN: 978-3-540-43009-4
Online ISBN: 978-3-540-45307-9
eBook Packages: Springer Book Archive