Fast optimal load balancing algorithms for 1D partitioning

dc.citation.epage996en_US
dc.citation.issueNumber8en_US
dc.citation.spage974en_US
dc.citation.volumeNumber64en_US
dc.contributor.authorPınar, A.en_US
dc.contributor.authorAykanat, Cevdeten_US
dc.contributor.bilkentauthorAykanat, Cevdet
dc.date.accessioned2016-02-08T10:26:22Z
dc.date.available2016-02-08T10:26:22Zen_US
dc.date.issued2004en_US
dc.departmentDepartment of Computer Engineeringen_US
dc.description.abstractThe one-dimensional decomposition of nonuniform workload arrays with optimal load balancing is investigated. The problem has been studied in the literature as the "chains-on-chains partitioning" problem. Despite the rich literature on exact algorithms, heuristics are still used in parallel computing community with the "hope" of good decompositions and the "myth" of exact algorithms being hard to implement and not runtime efficient. We show that exact algorithms yield significant improvements in load balance over heuristics with negligible overhead. Detailed pseudocodes of the proposed algorithms are provided for reproducibility. We start with a literature review and propose improvements and efficient implementation tips for these algorithms. We also introduce novel algorithms that are asymptotically and runtime efficient. Our experiments on sparse matrix and direct volume rendering datasets verify that balance can be significantly improved by using exact algorithms. The proposed exact algorithms are 100 times faster than a single sparse-matrix vector multiplication for 64-way decompositions on the average. We conclude that exact algorithms with proposed efficient implementations can effectively replace heuristics. © 2004 Elsevier Inc. All rights reserved.en_US
dc.description.provenanceMade available in DSpace on 2016-02-08T10:26:22Z (GMT). No. of bitstreams: 1 bilkent-research-paper.pdf: 70227 bytes, checksum: 26e812c6f5156f83f0e77b261a471b5a (MD5) Previous issue date: 2004en_US
dc.identifier.doi10.1016/j.jpdc.2004.05.003en_US
dc.identifier.issn0743-7315
dc.identifier.issn1096-0848
dc.identifier.urihttp://hdl.handle.net/11693/24250en_US
dc.language.isoEnglishen_US
dc.publisherAcademic Pressen_US
dc.relation.isversionofhttp://dx.doi.org/10.1016/j.jpdc.2004.05.003en_US
dc.source.titleJournal of Parallel and Distributed Computingen_US
dc.subjectChains-on-chains partitioningen_US
dc.subjectDynamic programmingen_US
dc.subjectImage-space parallel volume renderingen_US
dc.subjectIterative refinementen_US
dc.subjectOne-dimensional partitioningen_US
dc.subjectOptimal load balancingen_US
dc.subjectParallel sparse matrix vector multiplicationen_US
dc.subjectParametric searchen_US
dc.subjectAlgorithmsen_US
dc.subjectDynamic programmingen_US
dc.subjectGraph theoryen_US
dc.subjectHeuristic methodsen_US
dc.subjectMatrix algebraen_US
dc.subjectVectorsen_US
dc.subjectOne-dimensional partitioningen_US
dc.subjectOptimal load balancingen_US
dc.subjectParallel sparse matrix vector multiplicationen_US
dc.subjectParallel processing systemsen_US
dc.titleFast optimal load balancing algorithms for 1D partitioningen_US
dc.typeArticleen_US
Files
Original bundle
Now showing 1 - 1 of 1
Loading...
Thumbnail Image
Name:
Fast optimal load balancing algorithms for 1D partitioning.pdf
Size:
564.22 KB
Format:
Adobe Portable Document Format
Description:
Full printable version