Characterization of backfilling strategies for parallel job scheduling
- 25 June 2003
- conference paper
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- No. 15302016,p. 514-519
- https://doi.org/10.1109/icppw.2002.1039773
Abstract
Although there is wide agreement that backfilling produces significant benefits in scheduling of parallel jobs, there is no clear consensus on which backfilling strategy is preferable e.g. should conservative backfilling be used or the more aggressive EASY backfilling scheme; should a first-come first-served (FCFS) queue-priority policy be used, or some other such as shortest job first (SF) or expansion factor (XF); In this paper we use trace-based simulation to address these questions and glean new insights into the characteristics of backfilling strategies for job scheduling. We show that by viewing performance in terms of slowdowns and turnaround times of jobs within various categories based on their width (processor request size), length (job duration) and accuracy of the user's estimate of run time, some consistent trends may be observed.Keywords
This publication has 7 references indexed in Scilit:
- Distributed job scheduling on computational Grids using multiple simultaneous requestsPublished by Institute of Electrical and Electronics Engineers (IEEE) ,2003
- Selective preemption strategies for parallel job schedulingPublished by Institute of Electrical and Electronics Engineers (IEEE) ,2003
- Job-length estimation and performance in backfilling schedulersPublished by Institute of Electrical and Electronics Engineers (IEEE) ,2003
- Utilization, predictability, workloads, and user runtime estimates in scheduling the IBM SP2 with backfillingIEEE Transactions on Parallel and Distributed Systems, 2001
- Randomization, Speculation, and Adaptation in Batch SchedulersPublished by Institute of Electrical and Electronics Engineers (IEEE) ,2000
- Attacking the bottlenecks of backfilling schedulersCluster Computing, 2000
- A comparative study of online scheduling algorithms for networks of workstationsCluster Computing, 2000