Efficient Translation of OpenMP to Distributed Memory

Abstract

Abstract. The shared memory paradigm provides many benefits to the parallel programmer, particular with respect to applications that are hard to parallelize. Unfortunately, there are currently no efficient implementations of OpenMP for distributed memory platforms and this greatly diminishes its usefulness for real world parallel application development. In this paper we introduce a basic strategy for implementing OpenMP on distributed memory systems via a translation to Global Arrays. Global Arrays is a library of routines that provides many of the same features as OpenMP yet targets distributed memory platforms. Since it enables a reasonable translation strategy and also allows precise control over the movement of data within the resulting code, we believe it has the potential to provide higher levels of performance than the traditional translation of OpenMP to distributed memory via software distributed shared memory.

The SCore Cluster Enabled OpenMP Environment: Performance Prospects for Computational Science

“…Given the cost advantages of clusters, but the programming advantages of OpenMP and the existence of a large body of OpenMP code, it is not surprising that various groups have been attempting to develop OpenMP programming environments for clusters [2][3][4][5]. Most of these attempts have been based on layering OpenMP on top of some existing Software Distributed Shared Memory (SDSM) environment, although we note interesting recent work by Huang et al to implement OpenMP over Global Arrays [6]. In either case, to obtain reasonable performance from an OpenMP code on a cluster it is likely that the application programmer will require some knowledge of the implementation.…”