SC²S Colloquium - July 27, 2016
|Date:||July 27, 2016|
|Time:||3:00 pm, s.t.|
Julia Hoffmann: Implementation and Integration of an Efficient Data Structure for Parallel Tsunami Simulations
sunamis are amongst the most devastating disasters caused by nature. Because they can’t be avoided, it is of great interest to find ways to foresee their behavior in order to mitigate damage. The propagation of a tsunami wave is modeled with mathematical and numerical methods and is simulated on computers. The simulation should be as precise and as fast as possible, preferably in real-time. Such a computation is not trivial and therefore a lot of computing power is needed which can only be achieved with high performance computing. This bachelor’s thesis presents how an efficient data structure for parallel tsunami simulation with shallow water equations is implemented for shared, distribute memory systems and integrated into an existing tsunami application. The programming language X10 and the APGAS programming model are used and an already existing tsunami application serves as a model for the implementation of the code. In order to achieve efficiency, language constructs of X10 and programing principles of the APGAS model are exploited. The resulting code is integrated into a two dimensional tsunami application to simulate shallow water equations. Because this part is not to be achieved yet, possibilities for corrections and improvements are proposed.