File Download

There are no files associated with this item.

  Links for fulltext
     (May Require Subscription)
Supplementary

Conference Paper: Decentralized remapping of data parallel computations with the generalized dimension exchange method

TitleDecentralized remapping of data parallel computations with the generalized dimension exchange method
Authors
Xu, ChengZhongLau, Francis CM
Issue Date1994
Citation
Proceedings Of The Scalable High-Performance Computing Conference, 1994, p. 414-421 How to Cite?
AbstractThe Generalized Dimension Exchange (GDE) method is a fully distributed load balancing method that is most suitable for multicomputers with a direct communication network. It is extremely easy to implement and can yield optimal performance given a proper tuning. We propose a decentralized remapping method that uses the GDE algorithm periodically to balance (remap) the system's load. We implemented this remapping method in two data parallel applications and found it to be effective in reducing the computation time. The gains in performance (5 - 15%) due to remapping are reasonably substantial given the fact that the two applications by their very nature do not necessarily favor remapping.
Persistent Identifierhttp://hdl.handle.net/10722/151802

 

DC FieldValueLanguage
dc.contributor.authorXu, ChengZhongen_US
dc.contributor.authorLau, Francis CMen_US
dc.date.accessioned2012-06-26T06:29:44Z-
dc.date.available2012-06-26T06:29:44Z-
dc.date.issued1994en_US
dc.identifier.citationProceedings Of The Scalable High-Performance Computing Conference, 1994, p. 414-421en_US
dc.identifier.urihttp://hdl.handle.net/10722/151802-
dc.description.abstractThe Generalized Dimension Exchange (GDE) method is a fully distributed load balancing method that is most suitable for multicomputers with a direct communication network. It is extremely easy to implement and can yield optimal performance given a proper tuning. We propose a decentralized remapping method that uses the GDE algorithm periodically to balance (remap) the system's load. We implemented this remapping method in two data parallel applications and found it to be effective in reducing the computation time. The gains in performance (5 - 15%) due to remapping are reasonably substantial given the fact that the two applications by their very nature do not necessarily favor remapping.en_US
dc.languageengen_US
dc.relation.ispartofProceedings of the Scalable High-Performance Computing Conferenceen_US
dc.titleDecentralized remapping of data parallel computations with the generalized dimension exchange methoden_US
dc.typeConference_Paperen_US
dc.identifier.emailLau, Francis CM:fcmlau@cs.hku.hken_US
dc.identifier.authorityLau, Francis CM=rp00221en_US
dc.description.naturelink_to_subscribed_fulltexten_US
dc.identifier.scopuseid_2-s2.0-0028563612en_US
dc.identifier.spage414en_US
dc.identifier.epage421en_US
dc.identifier.scopusauthoridXu, ChengZhong=8701888000en_US
dc.identifier.scopusauthoridLau, Francis CM=7102749723en_US

Export via OAI-PMH Interface in XML Formats

OR

Export to Other Non-XML Formats