High performance computing using ProActive environment and the asynchronous iteration model

This paper presents a new library for the ProActive environment, called AIL-PA (asynchronous iterative library for ProActive). This new library allows to execute programs for solving large scale problems on various architectures. Two models of algorithm can be used: the synchronous iteration model which is efficient on single clusters; the asynchronous iteration model which is more efficient on distributed clusters. Both approaches are tested on both architectures, using Kernel CG of the NAS Parallel Benchmarks on the Grid'5000 platform. These tests also allow us to compare ProActive with AIL-PA and with the Jace programming environment. The results show that the asynchronous iteration model with AIL-PA is more efficient on distributed clusters than the synchronous iteration model. Moreover, these experiments also show that AIL-PA does not involve additional overhead to ProActive.