Library Support for Resource Constrained Accelerators

Laust Brock-Nannestad, Sven Karlsson

Research output: Chapter in Book/Report/Conference proceedingArticle in proceedingsResearchpeer-review

Abstract

Accelerators, and other resource constrained systems, are increasingly being used in computer systems. Accelerators provide power efficient performance and often provide a shared memory model. However, it is a challenge to map feature rich APIs, such as OpenMP, to resource constrained systems. In this paper, we present a lightweight system where an accelerator can remotely execute library functions on a host processor. The implementation takes up 750 bytes but can replace arbitrary library calls leading to significant savings in memory foot print. We evaluate with a set of SPLASH-2 applications and show that the impact on execution time is negligible when compared to GCCs OpenMP implementation.
Original languageEnglish
Title of host publicationUsing and Improving OpenMP for Devices, Tasks, and More : Proceedings of the 10th International Workshop on OpenMP, IWOMP 2014
EditorsLuiz DeRose, Bronis R. de Supinski, Stephen L. Olivier, Barbara M. Chapman, Matthias S. Müller
PublisherSpringer
Publication date2014
Pages187-201
ISBN (Print)978-3-319-11453-8
ISBN (Electronic)978-3-319-11454-5
DOIs
Publication statusPublished - 2014
Event10th International Workshop on OpenMP - Salvador, Brazil
Duration: 28 Sept 201430 Sept 2014
Conference number: 10
http://www.iwomp.org/

Workshop

Workshop10th International Workshop on OpenMP
Number10
Country/TerritoryBrazil
CitySalvador
Period28/09/201430/09/2014
Internet address
SeriesLecture Notes in Computer Science
Number8766
ISSN0302-9743

Fingerprint

Dive into the research topics of 'Library Support for Resource Constrained Accelerators'. Together they form a unique fingerprint.

Cite this