Parallel and Distributed Processing
Lecture Notes in Computer Science, 2000, Volume 1800/2000, 435-442, DOI: 10.1007/3-540-45591-4_58

Efficient Parallelization of Unstructured Reductions on Shared Memory Parallel Architectures

Siegfried Benkner and Thomas Brandes

View Related Documents

Abstract

This paper presents a new parallelization method for an efficient implementation of unstructured array reductions on shared memory parallel machines with OpenMP. This method is strongly related to parallelization techniques for irregular reductions on distributed memory machines as employed in the context of High Performance Fortran. By exploiting data locality, synchronization is minimized without introducing severe memory or computational overheads as observed with most existing shared memory parallelization techniques.
The work described in this paper was supported by NEC Europe Ltd. as part of the ADVICE project in cooperation with the NEC C&C Research Laboratories and by the Special Research Program SFB F011 AURORA of the Austrian Science Fund.

Fulltext Preview

Image of the first page of the fulltext document