Irregular data redistribution is used to enhance data locality and algorithm performance on heterogeneous processor systems. In this paper, we present an efficient scheduling algorithm based on convex bipartite communications for irregular GEN_BLOCK transformations. The proposed technique consists of two phases: degree reduction phase, schedules communications involved in processors with degree greater than two; and coloring phase, schedules remaining communications of all processors with degree-2 and degree-1. To evaluate the performance of our algorithm, we have implemented the proposed technique along with three scheduling methods. The simulation results show improvement of total communication costs by the proposed algorithm. © Springer-Verlag Berlin Heidelberg 2005.
CITATION STYLE
Hsu, C. H., Chen, S. C., Lan, C. Y., Yang, C. T., & Li, K. C. (2005). Scheduling convex bipartite communications toward efficient GEN_BLOCK transformations. In Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics) (Vol. 3758 LNCS, pp. 419–424). https://doi.org/10.1007/11576235_44
Mendeley helps you to discover research relevant for your work.