Performance gain and energy savings using multi-threaded software on an embedded asymmetric chip multiprocessor

Research output: Chapter in Book/Report/Conference proceedingsConference contributionpeer-review

Abstract

More powerful computers, as well as the rising costs for energy require modern computing systems to be more and more energy efficient. A modern approach in order to solve this issue is the usage of multi-core processors instead of single-core processing units. This paper focuses on modern ARM multi-core processors, which combine powerful and energy efficient processor cores on a single chip, in order to reduce power consumption and increase performance. To verify how much energy can be saved and how much performance can be gained by using multithreaded software, we have simulated a various number of different calculation scenarios for a modern ARM System-on-Chip. The results of these simulations show notably, how advantageous the usage of multi-core processors and power saving processors together on a single chip is. By using intelligent multi-threading concepts, the processor is able to reduce its power consumption by up to 60% compared to a single-threaded execution.

Original languageEnglish
Title of host publication27th European Modeling and Simulation Symposium, EMSS 2015
EditorsMichael Affenzeller, Francesco Longo, Lin Zhang, Agostino G. Bruzzone, Yuri Merkuryev, Emilio Jimenez
PublisherDIME UNIVERSITY OF GENOA
Pages207-212
Number of pages6
ISBN (Electronic)9788897999485
Publication statusPublished - 2015
Event27th European Modeling and Simulation Symposium, EMSS 2015 - Bergeggi, Italy
Duration: 21 Sep 201523 Sep 2015

Publication series

Name27th European Modeling and Simulation Symposium, EMSS 2015

Conference

Conference27th European Modeling and Simulation Symposium, EMSS 2015
CountryItaly
CityBergeggi
Period21.09.201523.09.2015

Keywords

  • Embedded asymmetric processor
  • Multi-core
  • Parallel software
  • Power consumption

Fingerprint Dive into the research topics of 'Performance gain and energy savings using multi-threaded software on an embedded asymmetric chip multiprocessor'. Together they form a unique fingerprint.

Cite this