OpenMP extensions for master-slave message passing computing

摘要

This paper presents a directive-based programming environment for master-slave message passing applications that enables the efficient execution of the same code on both shared and distributed memory multiprocessors. The environment exports an extension of the OpenMP workqueuing model, supports multiple levels of task parallelism and more than one master and provides transparent load balancing with a combination of static and dynamic scheduling of tasks. In addition, it operates exclusively through the available hardware on shared-memory machines and exploits MPI for explicit communication on clusters. Experimental results on a Linux-cluster demonstrate the successful combination of ease of programming with the performance of message passing.