Exploiting SPMD Horizontal Locality

In this paper, we analyze a particular spatial locality case (called horizontal locality) inherent to manycore accelerator architectures employing barrel execution of SPMD kernels, such as GPUs. We then propose an adaptive memory access granularity framework to exploit and enforce the horizontal loc...

Ausführliche Beschreibung

Gespeichert in:
Bibliographische Detailangaben
Veröffentlicht in:IEEE computer architecture letters 2011-01, Vol.10 (1), p.20-23
Hauptverfasser: Chunyang Gou, Gaydadjiev, G N
Format: Artikel
Sprache:eng
Schlagworte:
Online-Zugang:Volltext bestellen
Tags: Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
Beschreibung
Zusammenfassung:In this paper, we analyze a particular spatial locality case (called horizontal locality) inherent to manycore accelerator architectures employing barrel execution of SPMD kernels, such as GPUs. We then propose an adaptive memory access granularity framework to exploit and enforce the horizontal locality in order to reduce the interferences among accelerator cores memory accesses and hence improve DRAM efficiency. With the proposed technique, DRAM efficiency grows by 1.42X on average, resulting in 12.3% overall performance gain, for a set of representative memory intensive GPGPU applications.
ISSN:1556-6056
1556-6064
DOI:10.1109/L-CA.2011.5