On the development of an efficient coscheduling system

202. B. B. Zhou and R. P. Brent, On the development of an efficient coscheduling system, in Job Scheduling Strategies for Parallel Processing (Proc. Seventh JSSPP Workshop, MIT, Cambridge, Mass., June 2001), Lecture Notes in Computer Science, Vol. 2221, Springer-Verlag, Berlin, 2001, 103-115.

Paper: pdf (287K), ps (120K).

Abstract

Applying gang scheduling can alleviate the blockade problem caused by exclusively space-sharing scheduling. To simply allow jobs to run simultaneously on the same processors as in conventional gang scheduling may, however, introduce a large number of time slots in the system. In consequence the cost of context switches is greatly increased, each running job can only obtain a small portion of resources including memory space and processor utilisation, and no jobs can finish their computations quickly.

In this paper we present some experimental results to show that properly dividing jobs into different classes and applying different scheduling strategies to jobs of different classes can greatly reduce the average number of time slots in the system and significantly improve the performance in terms of average slowdown.

Comments

Related papers are [169, 180, 181, 189, 192, 194, 198].

Go to next publication

Return to Richard Brent's index page