Performance of Hierarchical Processor Scheduling in Shared-Memory Multiprocessor Systems
IEEE Transactions on Computers
An Efficient Adaptive Scheduling Scheme for Distributed Memory Multicomputers
IEEE Transactions on Parallel and Distributed Systems
Scalable loop self-scheduling schemes for heterogeneous clusters
International Journal of Computational Science and Engineering
Future Generation Computer Systems
Hi-index | 0.00 |
Processor scheduling policies for distributed-memorysystems can be divided into space-sharing or time-sharingpolicies. In space sharing, the set of processors in thesystem is partitioned and each partition is assigned for theexclusive use of a job. In time sharing policies, on the otherhand, none of the processors is given exclusively to jobs; instead,several jobs share the processors (for example, in a round robinfashion). There are advantages and disadvantages associated witheach type of policies. Typically, space-sharing polices are good atlow to moderate system loads and when jobs parallelism do notvary much. However, at high system loads and widely varyingjob parallelism, time sharing policies provide a better performance.In this paper we propose a new policy that is based on ahierarchical organization that incorporates the merits of these twotypes of policies. The new policy is a hybrid policy that uses bothspace-sharing as well as time-sharing to achieve better performance.We demonstrate that, at most system loads of interest, theproposed policy outperforms both space-sharing and time-sharingpolicies by a wide margin.