In partial fulfillment of the Requirements for the Degree of
PhD
Hatem Ltaief
will defend his thesis
Designing Efficient Parallel PDE Solvers in Heterogeneous Grid Environments
Abstract
In the last decade, the explosive interests and needs of the scientific community, in term of computing power, for solving large problems have emerged. The resources on a single cluster are often very limited and shared among users. The idea of coupling distributed computers became a paramount key to help scientists facing the new coming challenges. But this direction has some drawbacks. One of them is that most of the applications running on a single machine are quite hardware/software dependent and porting them to distributed computers is not straightforward. Moreover, while standard parallel applications based on master/slave architecture will overcome the network restrictions, scientific simulations solving PDEs will encounter serious difficulties to get reasonable performances. Furthermore, many computational science programs are now being designed to run for days or even for months, and as a matter of fact, scalability and fault tolerance can be a bottleneck. This becomes more critical when dealing with applications such as weather nowcasting or biomedical problems where close to real-time and accuracy are highly solicited.
Therefore, the objective of this talk is to demonstrate how parallel PDE solvers can efficiently run on a computational grid. The main idea is to minimize the overhead due to the grid constraints using an approach based on parallel numerical algorithms and tools.