Hostname: page-component-77c89778f8-swr86 Total loading time: 0 Render date: 2024-07-16T16:33:39.257Z Has data issue: false hasContentIssue false
  • English
  • Français

Parallelization of SPEC OMP benchmark programs with HPF and evaluation of their performances

Published online by Cambridge University Press:  20 December 2006

HITOSHI SAKAGAMI  and
HIROYUKI MORII
HITOSHI SAKAGAMI
Affiliation:
National Institute for Fusion Science, 322-6 Oroshi-cho, Toki 509-5292, Japan
HIROYUKI MORII
Affiliation:
Computer Engineering, University of Hyogo, 2167 Shosha, Himeji 671-2201, Japan
Get access Rights & Permissions [Opens in a new window]

Abstract

OpenMP is a directive-oriented paradigm for parallel programming applicable to shared memory parallel computers. On the other hand, High Performance Fortron (HPF) is also a directive-oriented paradigm but for distributed memory machines. Thus, it could be very interesting to compare their programming applicability and parallel performances. We have converted four SPEC OMP 2001 benchmark programs, which are written with OpenMP and designed to evaluate the performance of shared memory parallel computers, into HPF, and evaluated parallel performances. We also point out problems that we had faced during conversion, and discuss solutions for them. Finally, we have improved the HPF performance by modifying the original source code. We could get good performance with HPF quite similar to OpenMP with SWIM, and moderate performance with APSI.

Type
Papers
Information
Journal of Plasma Physics , Volume 72 , Issue 6 , December 2006 , pp. 1087 - 1090
Copyright
2006 Cambridge University Press

Access options

Get access to the full version of this content by using one of the access options below. (Log in options will check for institutional or personal access. Content may require purchase if you do not have access.)