A Historical Application Pro ler for Use by Parallel Schedulers

Abstract. Scheduling algorithms that use application and system knowl-edge have been shown to be more eective at scheduling parallel jobs on a multiprocessor than algorithms that do not. This paper focuses on obtaining such information for use by a scheduler in a network of work-stations environment. The log les from three parallel systems are examined to determine both how to categorize parallel jobs for storage in a job database and what job information would be useful to a scheduler. A Historical Proler is proposed that stores information about programs and users, and manip-ulates this information to provide schedulers with execution time pre-dictions. Several preemptive and non-preemptive versions of the FCFS, EASY and Least Work First scheduling algorithms are compared to eval-uate the utility of the proler. It is found that both preemption and the use of application execution time predictions obtained from the Histori-cal Proler lead to improved performance.