SC²S Colloquium - July 13, 2016

From Sccswiki
Jump to navigation Jump to search
Date: July 13, 2016
Room: 02.07.023
Time: 3:00 pm, s.t.

Dat Duc Nguyen: Python-based simulation of floods on high performance computers

Flood simulation is an important topic in computer science. Its criticality has been proven with the oods in Europe (2013) as well as Germany (2016). A large number of programs in C, C++ as well as FORTRAN have been developed for this task, mainly because of the languages' memory e�ciency and speed. Such programs have enabled researchers to simulate various scenarios with big data sets using high performance computers. Python is a distinguished programming language in the modern computing environment. To overcome the slowness problem of a high level scripting language, Python provides the exibility to extend functions and code with lower level languages like C and C++. Using MPI[27] makes it possible for parallel computing and combination of computers for faster calculation. Computation for large or multiple scenarios is often involved using high performance computing(HPC). Using the power of HPC, researchers can calculate complex, large scenarios in a minimum amount of time. The goal of this thesis is to create a Python framework for ood simulation with good usability and high performance. By experimenting on numerous parameters for a scenario in the study on chapter 3, the framework has been optimized to utilize the power of high performance computers, allowing complex calculations in a minimum amount of time.