Anzeige
Erschienen in:
2012 | OriginalPaper | Buchkapitel
Automatic Performance Analysis of OpenMP Codes on a Scalable Shared Memory System Using Periscope
verfasst von : Shajulin Benedict, Michael Gerndt
Erschienen in: Applied Parallel and Scientific Computing
Verlag: Springer Berlin Heidelberg
Aktivieren Sie unsere intelligente Suche, um passende Fachinhalte oder Patente zu finden.
Wählen Sie Textabschnitte aus um mit Künstlicher Intelligenz passenden Patente zu finden. powered by
Markieren Sie Textabschnitte, um KI-gestützt weitere passende Inhalte zu finden. powered by
OpenMP is a successful interface for programming parallel applications on shared memory systems. It is widely applied on small scale shared memory systems such as multicore processors, but also in hybrid programming on large supercomputers. This paper presents performance properties for OpenMP and their automatic detection by Periscope. We evaluate Periscope’s OpenMP analysis strategy in the context of the Altix 4700 supercomputer at Leibniz Computing Center (LRZ) in Garching. On this unique machine OpenMP scales up to 500 cores, one partition of in total 19 partitions. We present results for the NAS parallel benchmarks and for a large hybrid scientific application.