Abstract
Process migration is an useful mechanism to offer load balancing. In this context, we developed a model called MigBSP that controls processes rescheduling on BSP applications. MigBSP is especially pertinent to obtain performance on this type of applications, since they are composed by supersteps which always wait for the slowest process. In this paper, we focus on the BSP-based modeling of the widely used LU Decomposition algorithm as well as its execution with MigBSP. The use of multiple metrics to decide migrations and adaptations on rescheduling frequency turn possible gains up to 19% over our cluster-of-clusters architecture. Finally, our final idea is to show the possibility to get performance in LU effortlessly by using novel migration algorithms. © 2011 Springer-Verlag Berlin Heidelberg.
Cite
CITATION STYLE
Da Rosa Righi, R., Lima Pilla, L., Carissimi, A., Navaux, P. O. A., & Heiss, H. U. (2011). Applying process migration on a BSP-based LU decomposition application. In Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics) (Vol. 6449 LNCS, pp. 314–326). https://doi.org/10.1007/978-3-642-19328-6_30
Register to see more suggestions
Mendeley helps you to discover research relevant for your work.
