Studying the impact of synchronization frequency on scheduling tasks with dependencies in heterogeneous systems

Andronikos, T; Ciorba, FM; Riakiotakis, I; Papakonstantinou, G; Chronopoulos, AT

dc.contributor.author	Andronikos, T	en
dc.contributor.author	Ciorba, FM	en
dc.contributor.author	Riakiotakis, I	en
dc.contributor.author	Papakonstantinou, G	en
dc.contributor.author	Chronopoulos, AT	en
dc.date.accessioned	2014-03-01T02:47:02Z
dc.date.available	2014-03-01T02:47:02Z
dc.date.issued	2010	en
dc.identifier.issn	0166-5316	en
dc.identifier.uri	https://dspace.lib.ntua.gr/xmlui/handle/123456789/32985
dc.subject	Dynamic load balancing	en
dc.subject	Heterogeneous systems	en
dc.subject	Inter-processor communication	en
dc.subject	Loops with dependencies	en
dc.subject	Performance evaluation	en
dc.subject	Self-scheduling algorithms	en
dc.subject	Synchronization frequency	en
dc.subject.classification	Computer Science, Hardware & Architecture	en
dc.subject.classification	Computer Science, Theory & Methods	en
dc.subject.other	Dynamic load balancing	en
dc.subject.other	Heterogeneous systems	en
dc.subject.other	Inter processor communication	en
dc.subject.other	Loops with dependencies	en
dc.subject.other	Performance evaluation	en
dc.subject.other	Self-scheduling algorithms	en
dc.subject.other	Dynamic loads	en
dc.subject.other	Image processing	en
dc.subject.other	Optimization	en
dc.subject.other	Parallel architectures	en
dc.subject.other	Scheduling algorithms	en
dc.subject.other	Synchronization	en
dc.title	Studying the impact of synchronization frequency on scheduling tasks with dependencies in heterogeneous systems	en
heal.type	conferenceItem	en
heal.identifier.primary	10.1016/j.peva.2010.08.020	en
heal.identifier.secondary	http://dx.doi.org/10.1016/j.peva.2010.08.020	en
heal.language	English	en
heal.publicationDate	2010	en
heal.abstract	In this work, we develop and evaluate a theoretical model, which we then use to study the impact of the synchronization frequency on the performance of dynamic self-scheduling algorithms. These algorithms are used to parallelize loops with data dependencies on heterogeneous systems. The proposed model uses a formula to estimate the parallel time as a function of the synchronization frequency. Inter-node communication has been proven to be the dominant factor for the performance degradation of applications containing loops with data dependencies. The synchronization mechanism therefore requires careful fine-tuning in order to give the best possible performance. The proposed model determines the optimal synchronization frequency that results in the minimum parallel time. We use this model to study the impact of the synchronization frequency on the parallel execution of a computational kernel from image processing. For this kernel, the synchronization frequency giving the minimum parallel time predicted by our theoretical model was very close to the synchronization frequency giving the least parallel time in practice. We validate our model by extensive comparisons of the theoretically predicted parallel time and synchronization frequency against those obtained from practical experiments. The comparisons show that the proposed model is highly accurate, its predictions for the optimal synchronization frequency being within 0.0250% of the experimentally optimal synchronization frequency in the best case, and within 0.1750% of the experimentally optimal synchronization frequency in the worst case. Finally, the comparisons show that the proposed model improves on a previously existing model in heterogeneous systems, whereas it gives similar results in homogeneous systems. © 2010 Elsevier B.V. All rights reserved.	en
heal.publisher	ELSEVIER SCIENCE BV	en
heal.journalName	Performance Evaluation	en
dc.identifier.doi	10.1016/j.peva.2010.08.020	en
dc.identifier.isi	ISI:000284965500004	en
dc.identifier.volume	67	en
dc.identifier.issue	12	en
dc.identifier.spage	1324	en
dc.identifier.epage	1339	en