Efficiently Scheduling Task Dataflow Parallelism: A Comparison Between Swan and QUARK

Research output: Chapter in Book/Report/Conference proceedingConference contribution

160 Downloads (Pure)

Abstract

Increased system variability and irregularity of parallelism in applications put increasing demands on the ef- ficiency of dynamic task schedulers. This paper presents a new design for a work-stealing scheduler supporting both Cilk- style recursively parallel code and parallelism deduced from dataflow dependences. Initial evaluation on a set of linear algebra kernels demonstrates that our scheduler outperforms PLASMA’s QUARK scheduler by up to 12% on a 16-thread Intel Xeon and by up to 50% on a 32-thread AMD Bulldozer.
Original languageEnglish
Title of host publicationProceedings of the Exascale Applications and Software Conference 2015
Place of PublicationEdinburgh
PublisherThe University of Edinburgh
Pages36-41
Number of pages6
ISBN (Print)978-0-9926615-1-9
Publication statusPublished - Jul 2015
EventExascale Applications and Software Conference 2015 - Edinburgh, United Kingdom
Duration: 21 Apr 201523 Apr 2015

Conference

ConferenceExascale Applications and Software Conference 2015
CountryUnited Kingdom
CityEdinburgh
Period21/04/201523/04/2015

Fingerprint Dive into the research topics of 'Efficiently Scheduling Task Dataflow Parallelism: A Comparison Between Swan and QUARK'. Together they form a unique fingerprint.

  • Cite this

    Vandierendonck, H. (2015). Efficiently Scheduling Task Dataflow Parallelism: A Comparison Between Swan and QUARK. In Proceedings of the Exascale Applications and Software Conference 2015 (pp. 36-41). The University of Edinburgh. http://www.easc2015.ed.ac.uk/home