Gang Scheduling for Multiprocessor Systems

        This paper presents the design and underlying algorithms for a job scheduling system which efficiently supports gang scheduling in distributed multiprocessor systems. Our scheduler is highly scalable, modular in its design, and has low overhead. It is composed of three interacting components, global scheduler, local scheduler and a context switching agent. The system is architected to allow the time sharing of a parallel machine while preserving the advantages of high performance communication subsystems often present in today's MPPs.

By: Hubertus Franke, Fang Wang (Yale Univ.), Pratap Pattnaik and Larry Rudolph (Hebrew Univ., Jerusalem)

Published in: RC20450 in 1996

This Research Report is not available electronically. Please request a copy from the contact listed below. IBM employees should contact ITIRC for a copy.

Questions about this service can be mailed to reports@us.ibm.com .