Backfilling with Guarantees Granted upon Job Submission
In this paper, we present scheduling algorithms that simultaneously support guaranteed starting times and favor jobs with system-desired traits. To achieve the first of these goals, our algorithms keep a profile with potential starting times for every unfinished job and never move these starting times later, just as in Conservative Backfilling. To achieve the second, they exploit previously unrecognized flexibility in the handling of holes opened in this profile when jobs finish early. We find that, with one choice of job selection function, our algorithms can consistently yield a lower average waiting time than Conservative Backfilling while still providing a guaranteed start time to each job as it arrives. In fact, in most cases, the algorithms give a lower average waiting time than the more aggressive EASY backfilling algorithm, which does not provide guaranteed start times. Alternately, with a different choice of job selection function, our algorithms can focus the benefit on the widest submitted jobs, the reason for the existence of parallel systems. In this case, these jobs experience significantly lower waiting time than Conservative Backfilling with minimal impact on other jobs.
KeywordsAverage Waiting Time Priority Function Batch Scheduler User Runtime Estimate Plan Start Time
Unable to display preview. Download preview PDF.
- 2.Das Sharma, D., Pradhan, D.K.: Job scheduing in mesh multicomputers. In: Proc. Intern. Conf. on Parallel Processing Workshops, pp. 251–258 (1994)Google Scholar
- 3.Feitelson, D.: The parallel workloads archive, http://www.cs.huji.ac.il/labs/parallel/workload/index.html
- 6.Leung, V., Sabin, G., Sadayappan, P.: Parallel job scheduling policies to improve fairness - a case study. In: Proc. 6th Intern. Workshop on Scheduling and Resource Management for Parallel and Distributed Syst. (2010)Google Scholar
- 9.Perković, D., Keleher, P.J.: Randomization, speculation, and adaptation in batch schedulers. In: Proc. 2000 ACM/IEEE Conf. on Supercomputing (2000)Google Scholar
- 10.Schwiegelshohn, U., Yahyapour, R.: Analysis of first-come-first-serve parallel job scheduling. In: Proc. 9th ACM/SIAM Symp. on Discrete Algorithms, pp. 629–638 (1998)Google Scholar
- 14.Talby, D., Feitelson, D.G.: Supporting priorities and improving utilization of the IBM SP scheduler using slack-based backfilling. In: Proc. 13th Intern. Parallel Processing Symp., pp. 513–517 (1999)Google Scholar
- 16.Tsafrir, D., Feitelson, D.G.: The dynamics of backfilling: Solving the mystery of why increased inaccuracy may help. In: Proc. IEEE Intern. Symp. on Workload Characterization, pp. 131–141 (2006)Google Scholar