KIT | KIT-Bibliothek | Impressum | Datenschutz

Malleable Sorting

Flick, Patrick; Sanders, Peter; Speck, Jochen

Abstract (englisch):
Malleable jobs can adapt to varying degrees of available parallelism. This is an interesting approach to more flexible usage of parallel resources. For example, malleable jobs can be scheduled optimally and efficiently where more restricted forms of parallel jobs are NP-hard to handle. However, little work has been done on how to make fundamental computations malleable. We study how this can be done for sorting. Our algorithm is an adaptive version of Multiway Merge Sort and outperforms a state-of-the art implementation in the multi core STL when the number of available cores fluctuates.

DOI: 10.1109/IPDPS.2013.90
Zitationen: 5
Zugehörige Institution(en) am KIT Institut für Theoretische Informatik (ITI)
Publikationstyp Proceedingsbeitrag
Publikationsmonat/-jahr 05.2013
Sprache Englisch
Identifikator ISBN: 978-0-7695-4971-2
KITopen-ID: 1000097635
Erschienen in 27th IEEE International Parallel and Distributed Processing Symposium (IPDPS 2013), Boston, MA, May 20-24, 2013
Verlag IEEE, Piscataway, NJ
Seiten 418–426
Nachgewiesen in Scopus
KIT – Die Forschungsuniversität in der Helmholtz-Gemeinschaft
KITopen Landing Page