Abstract
Algorithms are presented for extracting synchronization-free parallelism available in arbitrarily nested parameterized loops. The parallelism is represented with synchronization-free trees of dependent operations. Sets representing trees can be described with non-linear expressions. The main idea is to firstly extract sources of synchronization-free trees and next to generate parallel code based on a while loop. Experimental results are presented exposing speed-up and efficiency of parallel programs written in the OpenMP standard on the basis of code generated by the algorithms proposed. © 2008 Springer-Verlag Berlin Heidelberg.
Cite
CITATION STYLE
Bielecki, W., Beletska, A., Palkowski, M., & San Pietro, P. (2008). Finding synchronization-free parallelism represented with trees of dependent operations. In Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics) (Vol. 5022 LNCS, pp. 185–195). https://doi.org/10.1007/978-3-540-69501-1_20
Register to see more suggestions
Mendeley helps you to discover research relevant for your work.
