Boosting irregular array Reductions through In-lined Block-ordering on fast processors

Array-type reductions represent a frequently occurring algorithmic pattern in many scientific applications. A special case occurs if array elements are accessed in an irregular, often random manner, making their concurrent and scalable execution difficult. In this work we present a new approach that...

Ausführliche Beschreibung

Gespeichert in:
Bibliographische Detailangaben
Veröffentlicht in:2015 IEEE High Performance Extreme Computing Conference (HPEC) S. 1 - 6
Hauptverfasser: Ciesko, Jan, Mateo, Sergi, Teruel, Xavier, Beltran, Vicenc, Martorell, Xavier, Labarta, Jesus
Format: Tagungsbericht
Sprache:Englisch
Veröffentlicht: IEEE 01.09.2015
Schlagworte:
Array reduction
Arrays
Bandwidth
Instruction sets
OpenMP
PIBOR
Privatization
Resource management
Runtime
Online-Zugang:Volltext
Tags: Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
Beschreibung
Zusammenfassung:Array-type reductions represent a frequently occurring algorithmic pattern in many scientific applications. A special case occurs if array elements are accessed in an irregular, often random manner, making their concurrent and scalable execution difficult. In this work we present a new approach that consists of language and runtime support and targets popular parallel programming models such as OpenMP. Its runtime support implements Privatization with In-lined, Block-Ordered Reductions (PIBOR), a new approach that trades processor cycles to increase locality and bandwidth efficiency for such algorithms. A reference implementation in OmpSs, a task-parallel programming model, shows promising results on current multicore systems.
DOI:10.1109/HPEC.2015.7322443