Locality-conscious process scheduling in embedded systems

In many embedded systems, existence of a data cache might influence the effectiveness of process scheduling policy significantly. Consequently, a scheduling policy that takes inter-process data reuse into account might result in large performance benefits. In this paper, we focus on array-intensive...

Full description

Saved in:
Bibliographic Details
Published inInternational Conference on Hardware Software Codesign: Proceedings of the tenth international symposium on Hardware/software codesign; 06-08 May 2002 pp. 193 - 198
Main Authors Kadayif, I., Kandemir, M., Kolcu, I., Chen, G.
Format Conference Proceeding
LanguageEnglish
Published New York, NY, USA ACM 06.05.2002
IEEE
SeriesACM Conferences
Subjects
Online AccessGet full text

Cover

Loading…
More Information
Summary:In many embedded systems, existence of a data cache might influence the effectiveness of process scheduling policy significantly. Consequently, a scheduling policy that takes inter-process data reuse into account might result in large performance benefits. In this paper, we focus on array-intensive embedded applications and present a locality-conscious scheduling strategy where we first evaluate the potential data reuse between processes, and then, using the results of this evaluation, select an order for process executions. We also show how process codes can be transformed by an optimizing compiler for increasing inter-process data reuse, thereby making locality-conscious scheduling more effective. Our experimental results obtained using two large, multi-process application codes indicate significant runtime benefits.
Bibliography:SourceType-Conference Papers & Proceedings-1
ObjectType-Conference Paper-1
content type line 25
ISBN:1581135424
9781581135428
DOI:10.1145/774789.774829