Recent developments in genetics, molecular biology, and biomedical sciences do make a tremendous amount of data available to scientists. The required analyses (i.e., post processing) is increasingly done on homogenous high performance computing clusters, the need for computing power of these kind of applications is very demanding. But focusing just on homogenous systems has a major drawback: the exceeding costs. Our research will focus on the development of a resource facility that utilizes heterogeneous hardware infrastructure to achieve an optimal cost to performance ratio. Emphasis is laid on - but not restricted to - alternative system and processor topologies like general purpose graphic processors (GPGPUs) as well as the IBM CELL Broadband Engine. Besides the hardware aspects, our research will also concentrate on how to bring this computational power without any loss to the user and make it thus available for solving highly specific problems of domain experts, especially molecular biologists. Finally, our research will also concentrate on how to develop intelligent scheduling and priority management for highly demanding compute tasks.
|Titel||Proceedings of 21st European Modeling and Simulation Symposium EMSS 2009|
|Herausgeber (Verlag)||DIPTEM University of Genova|
|Publikationsstatus||Veröffentlicht - 2009|
|Veranstaltung||21st European Modeling and Simulation Symposium 2009 EMSS09 - Tenerife, Spanien|
Dauer: 23 Sep 2009 → 25 Sep 2009
|Konferenz||21st European Modeling and Simulation Symposium 2009 EMSS09|
|Zeitraum||23.09.2009 → 25.09.2009|