High-performance scientific applications require more and more compute power.The concurrent use of multiple distributed compute resources is vital formaking scientific progress. The resulting distributed system, a so-calledJungle Computing System, is both highly heterogeneous and hierarchical,potentially consisting of grids, clouds, stand-alone machines, clusters,desktop grids, mobile devices, and supercomputers, possibly with acceleratorssuch as GPUs. One striking example of applications that can benefit greatly of JungleComputing Systems are Multi-Model / Multi-Kernel simulations. In thesesimulations, multiple models, possibly implemented using different techniquesand programming models, are coupled into a single simulation of a physicalsystem. Examples include the domain of computational astrophysics and climatemodeling. In this paper we investigate the use of Jungle Computing Systems for suchMulti-Model / Multi-Kernel simulations. We make use of the software developedin the Ibis project, which addresses many of the problems faced when runningapplications on Jungle Computing Systems. We create a prototype Jungle-awareversion of AMUSE, an astrophysical simulation framework. We show preliminaryexperiments with the resulting system, using clusters, grids, stand-alonemachines, and GPUs.
展开▼
