File Download
  Links for fulltext
     (May Require Subscription)
Supplementary

Conference Paper: Adaptive sampling-based profiling techniques for optimizing the distributed JVM runtime

TitleAdaptive sampling-based profiling techniques for optimizing the distributed JVM runtime
Authors
KeywordsAccess locality
Correlation tracking
Distributed Java virtual machine
Distributed shared memory systems
Dynamic load balancing
Object sharing
Profiling
Sampling
Thread affinity
Thread migration
Thread stack
Issue Date2010
PublisherIEEE, Computer Society.
Citation
The 24th IEEE International Symposium on Parallel & Distributed Processing (IPDPS 2010), Atlanta, GA., 19-23 April 2010. In Proceedings of the 24th IPDPS, 2010, p. 1-11 How to Cite?
AbstractExtending the standard Java virtual machine (JVM) for cluster-awareness is a transparent approach to scaling out multithreaded Java applications. While this clustering solution is gaining momentum in recent years, efficient runtime support for fine-grained object sharing over the distributed JVM remains a challenge. The system efficiency is strongly connected to the global object sharing profile that determines the overall communication cost. Once the sharing or correlation between threads is known, access locality can be optimized by collocating highly correlated threads via dynamic thread migrations. Although correlation tracking techniques have been studied in some page-based sof Tware DSM systems, they would entail prohibitively high overheads and low accuracy when ported to fine-grained object-based systems. In this paper, we propose a lightweight sampling-based profiling technique for tracking inter-thread sharing. To preserve locality across migrations, we also propose a stack sampling mechanism for profiling the set of objects which are tightly coupled with a migrant thread. Sampling rates in both techniques can vary adaptively to strike a balance between preciseness and overhead. Such adaptive techniques are particularly useful for applications whose sharing patterns could change dynamically. The profiling results can be exploited for effective thread-to-core placement and dynamic load balancing in a distributed object sharing environment. We present the design and preliminary performance result of our distributed JVM with the profiling implemented. Experimental results show that the profiling is able to obtain over 95% accurate global sharing profiles at a cost of only a few percents of execution time increase for fine- to medium- grained applications. © 2010 IEEE.
Persistent Identifierhttp://hdl.handle.net/10722/125697
ISBN
References

 

DC FieldValueLanguage
dc.contributor.authorLam, KTen_HK
dc.contributor.authorLuo, Yen_HK
dc.contributor.authorWang, CLen_HK
dc.date.accessioned2010-10-31T11:46:40Z-
dc.date.available2010-10-31T11:46:40Z-
dc.date.issued2010en_HK
dc.identifier.citationThe 24th IEEE International Symposium on Parallel & Distributed Processing (IPDPS 2010), Atlanta, GA., 19-23 April 2010. In Proceedings of the 24th IPDPS, 2010, p. 1-11en_HK
dc.identifier.isbn978-1-4244-6442-5-
dc.identifier.urihttp://hdl.handle.net/10722/125697-
dc.description.abstractExtending the standard Java virtual machine (JVM) for cluster-awareness is a transparent approach to scaling out multithreaded Java applications. While this clustering solution is gaining momentum in recent years, efficient runtime support for fine-grained object sharing over the distributed JVM remains a challenge. The system efficiency is strongly connected to the global object sharing profile that determines the overall communication cost. Once the sharing or correlation between threads is known, access locality can be optimized by collocating highly correlated threads via dynamic thread migrations. Although correlation tracking techniques have been studied in some page-based sof Tware DSM systems, they would entail prohibitively high overheads and low accuracy when ported to fine-grained object-based systems. In this paper, we propose a lightweight sampling-based profiling technique for tracking inter-thread sharing. To preserve locality across migrations, we also propose a stack sampling mechanism for profiling the set of objects which are tightly coupled with a migrant thread. Sampling rates in both techniques can vary adaptively to strike a balance between preciseness and overhead. Such adaptive techniques are particularly useful for applications whose sharing patterns could change dynamically. The profiling results can be exploited for effective thread-to-core placement and dynamic load balancing in a distributed object sharing environment. We present the design and preliminary performance result of our distributed JVM with the profiling implemented. Experimental results show that the profiling is able to obtain over 95% accurate global sharing profiles at a cost of only a few percents of execution time increase for fine- to medium- grained applications. © 2010 IEEE.en_HK
dc.languageengen_HK
dc.publisherIEEE, Computer Society.-
dc.relation.ispartofProceedings of the IEEE International Symposium on Parallel and Distributed Processing, IPDPS 2010en_HK
dc.rightsCreative Commons: Attribution 3.0 Hong Kong License-
dc.rightsIPDPS Proceedings. Copyright © IEEE, Computer Society.-
dc.rights©2010 IEEE. Personal use of this material is permitted. However, permission to reprint/republish this material for advertising or promotional purposes or for creating new collective works for resale or redistribution to servers or lists, or to reuse any copyrighted component of this work in other works must be obtained from the IEEE.-
dc.subjectAccess localityen_HK
dc.subjectCorrelation trackingen_HK
dc.subjectDistributed Java virtual machineen_HK
dc.subjectDistributed shared memory systemsen_HK
dc.subjectDynamic load balancingen_HK
dc.subjectObject sharingen_HK
dc.subjectProfilingen_HK
dc.subjectSamplingen_HK
dc.subjectThread affinityen_HK
dc.subjectThread migrationen_HK
dc.subjectThread stacken_HK
dc.titleAdaptive sampling-based profiling techniques for optimizing the distributed JVM runtimeen_HK
dc.typeConference_Paperen_HK
dc.identifier.openurlhttp://library.hku.hk:4550/resserv?sid=HKU:IR&issn=978-1-4244-6442-5&volume=&spage=1&epage=11&date=2010&atitle=Adaptive+sampling-based+profiling+techniques+for+optimizing+the+distributed+JVM+runtime-
dc.identifier.emailWang, CL:clwang@cs.hku.hken_HK
dc.identifier.authorityWang, CL=rp00183en_HK
dc.description.naturepublished_or_final_version-
dc.identifier.doi10.1109/IPDPS.2010.5470461en_HK
dc.identifier.scopuseid_2-s2.0-77954027160en_HK
dc.identifier.hkuros179413en_HK
dc.relation.referenceshttp://www.scopus.com/mlt/select.url?eid=2-s2.0-77954027160&selection=ref&src=s&origin=recordpageen_HK
dc.identifier.spage1-
dc.identifier.epage11-
dc.description.otherThe 24th IEEE International Symposium on Parallel & Distributed Processing (IPDPS 2010), Atlanta, GA., 19-23 April 2010. In Proceedings of the 24th IPDPS, 2010, p. 1-11-
dc.identifier.scopusauthoridLam, KT=26031004100en_HK
dc.identifier.scopusauthoridLuo, Y=35759395100en_HK
dc.identifier.scopusauthoridWang, CL=7501646188en_HK

Export via OAI-PMH Interface in XML Formats


OR


Export to Other Non-XML Formats