Scalable loop self-scheduling schemes for heterogeneous clusters
International Journal of Computational Science and Engineering
Future Generation Computer Systems
Dynamic multi phase scheduling for heterogeneous cluste
IPDPS'06 Proceedings of the 20th international conference on Parallel and distributed processing
Hi-index | 0.00 |
Abstract: Load imbalance is a serious impediment to achieving good performance in parallel processing. Global load balancing schemes cannot adequately man-age to balance parallel tasks generated from a single application. Dynamic loop scheduling methods are known to be useful in balancing parallel loops on shared-memory multiprocessor machines. However, their centralized nature causes a bottleneck even for the relatively small number of processors in workstation clusters because of order-of-magnitude differences in communications overheads. Moreover, improvements of basic loop scheduling methods have not dealt effectively with irregularly distributed workloads in parallel loops, which commonly occur in applications for work-station clusters. In this paper, we present a new decentralized balancing method for parallel loops on workstation clusters.