direkt zum Inhalt springen

direkt zum Hauptnavigationsmenü

Sie sind hier

TU Berlin

Inhalt des Dokuments

Publikationen

Evaluating Thread Placement Based on Memory Access Patterns for Multi-core Processors
Zitatschlüssel Diener2010
Autor Matthias Diener and Felipe L. Madruga and Eduardo R. Rodrigues and Marco A. Z. Alves and Jörg Schneider and Philippe O. A. Navaux and Hans-Ulrich Heiß
Seiten 491–496
Jahr 2010
ISBN 978-0-7695-4214-0
DOI 10.1109/HPCC.2010.114
Journal Proceedings of 12th IEEE International Conference on High Performance Computing and Communications (HPCC-2010)
Herausgeber Guerrero, Juan E
Zusammenfassung Process placement is a technique widely used on parallel machines with heterogeneous interconnections to reduce the overall communication time. For instance, two processes which communicate frequently are mapped close to each other. Finding the optimal mapping between threads and cores in a shared-memory environment (for example, OpenMP and Pthreads) is an even more complex task due to implicit communication. In this work, we examine data sharing patterns between threads in di erent workloads and use those patterns in a similar way as messages are used to map processes in cluster computers. We evaluated our technique on two state-of-the-art multi-core processors and achieved moderate improvements in the common case and considerable improvements in some cases, reducing execution time by up to 45%.
Link zur Publikation Link zur Originalpublikation Download Bibtex Eintrag

Zusatzinformationen / Extras

Direktzugang

Schnellnavigation zur Seite über Nummerneingabe

Ansprechpartner

Jörg Schneider
+49 30 314-73388
Raum EN 357

Webseite