Potentials and Limitations for Energy Efficiency Auto-Tuning

Research output: Contribution to book/conference proceedings/anthology/reportChapter in book/anthology/reportContributedpeer-review

Abstract

With an objective target of 20 MW for an Exascale system, the power budget for future super computers has a severe restriction. Fortunately, current hardware and runtime environments feature various adjustable parameters that can be changed at runtime in order to confine the power usage. We present an overview of the potentials and limitations of several power saving mechanisms and describe the integration of energy efficiency optimizations in existing performance analysis tools. We also introduce an energy efficiency tuning cycle and exemplify its usage.

Details

Original languageEnglish
Title of host publicationParallel Computing: Accelerating Computational Sciene and Engineering (CSE)
EditorsMichael Bader, Arndt Bode, Hans-Joachim Bungartz, Michael Gerndt, Gerhard R. Joubert, Frans Peters
PublisherIOS Press, Amsterdam [u. a.]
Pages678-687
Number of pages10
Volume25
ISBN (print)978-1-61499-381-3
Publication statusPublished - 1 Dec 2013
Peer-reviewedYes

Conference

TitleInternational Conference on Parallel Computing 2013
Abbreviated titleParCo 2013
Duration10 - 13 September 2013
Website
Degree of recognitionInternational event
LocationTechnische Universität München
CityMünchen
CountryGermany

External IDs

Scopus 84902284275
ORCID /0009-0003-0666-4166/work/151475604

Keywords

Sustainable Development Goals