Achieving Fair Share Objectives via Goal-Oriented Parallel Computer Job Scheduling Policies

Fair share is one of the scheduling objectives supported on many production systems. However, fair share has been shown to cause performance problems for some users, especially the users with difficult jobs. This work is focusing on extending goaloriented parallel computer job scheduling policies to cover the fair share objective. Goal-oriented parallel computer job scheduling policies have been shown to achieve good scheduling performances when conflicting objectives are required. Goal-oriented policies achieve such good performance by using anytime combinatorial search techniques to find a good compromised schedule within a time limit. The experimental results show that the proposed goal-oriented parallel computer job scheduling policy (namely Tradeofffs( Tw:avgX)) achieves good scheduling performances and also provides good fair share performance.

• Sangsuree Vasupongayya

• goal-oriented parallel job scheduling policies
• fairshare.

[1] OpenPBS, http://www.nas.nasa.gov/Software/PBS/
[2] PBS pro, http://www.pbspro.com
[3] LSF, http://www.platform.com/product/ lsffamily.
[4] LSF fair share documentation, http://accl.grc.nasa.gov/
job_schedulers/lsf/ Docs/lsf6.1/lsf6.1_admin /E_fairshare.html
[5] D. Jackson, Q. Snell & M. Clement. "Core algorithms of the MAUI
scheduler". In proceeding of the Workshop on Job Scheduling Strategies
for Parallel Processing, 2001.
[6] Maui scheduler, http://www.supercluster.org/maui
[7] Moab scheduler, http://www.clusterresources.com/products/mwm/
docs/moabadminguide450.pdf
[8] S. Kannan, M. Roberts, P. Mayes, D. Brelsford & J. Skovira. "Workload
management with LoadLeveler". Technical Report, IBM Redbook, 2001.
[9] J. Key & P. Lauder. "A fair share scheduler". Communications of the
ACM, 31(3):44-55, 1988.
[10] S. Vasupongayya, "Impact of fair share and its configurations on parallel
job scheduling algorithms". (to appear). In proceeding of the 2009
WASET International Conference on High Performance Computing,
Venice, Italy, October 2009.
[11] S.-H. Chiang and S. Vasupongayya, "Design and potential performance
of goal-oriented job scheduling policies for parallel computer
workloads". In the IEEE Transaction on Parallel and Distributed
Systems. 19(12):1642-1656, 2009.
[12] S. Vasupongayya, "Goal-oriented parallel job scheduling: A revisit", In
proceeding of the 2nd UBU-Research, Ubonratchathani, Thailand, July
2008.
[13] S. Vasupongayya, S.-H Chiang and B. Massey, "Search-based job
scheduling for parallel computer workloads", In proceeding of the IEEE
Cluster, Boston, MA, 2005.
[14] S. Kleban and S. Clearwater. "Fair share on high performance computing
system: What does fair really mean?" in proceeding of the IEEE
International Symposium on Cluster Computing and the Grid, 2003.
[15] S. Vasupongayya and S.-H. Chiang. "Multi-objective models for
scheduling jobs on parallel computer systems". In proceeding of IEEE
Cluster, Barcelona, Spain, 2006.
[16] S.-H. Chiang, A. Arpaci-Dusseau and M. Vernon. "The impact of more
accurate request runtimes on production job scheduling performance". In
Lecture Notes in Computer Science (2537):103-127, 2002.
[17] S.-H. Chiang and C. Fu. "Benefit of limited time-sharing in the presence
of very large parallel jobs". In proceedings of the IEEE International
Parallel and Distributed Processing Symposium, 2005.
[18] S.-H. Chiang and M. Vernon. "Production job scheduling for parallel
shared memory systems". In proceeding of the IEEE International
Parallel and Distributed Processing Symposium, 2001.
[19] D. Talby and D. Feitelson, "Supporting priorities and improving
utilization of the IBM SP2 scheduler using slack-based backfilling". In
proceeding of the International Parallel Processing Symposium, 1999.
[20] D. Talby and D. Feitelson, "Improving and stabilizing parallel computer
performance using adaptive backfilling". In proceeding of the IEEE
International Parallel and Distributed Processing Symposium, 2005.