Momentan haben wir am FG Simulation und Modellierung noch zwei Abschlussarbeiten zu vergeben:
Effective usage of GPU resources in Distributed Discrete Event Simulation (DDES)
In Parallel and Distributed Simulation there are open research issues regarding the exploitation of Graphics Processing Units. One focus of interest is to develop practical, scalable techniques to significantly accelerate real-world simulation applications by using modern graphic cards.
Within our investigation in the field of GPU computation as part of DDES, we are offering a bachelor/master thesis. The student will be introduced into basic programming techniques (e.g. OpenCL, CUDA) related to General Purpose Computation on Graphics Processing Unit (GPGPU). As part of the thesis a DDES will be implemented to emphasize benefits as well as drawbacks that ensue from the GPGPU in the DDES.
Knowledge in C++ is preferred. Skills in simulation are beneficial.
Literature:
Richard M. Fujimoto, Research Challenges in Parallel and Distributed Simulation
Distributed Fluid Animation and Distributed Fluid Simulation using heterogeneous models
Different goals offer different ways to model fluids (such as ocean water). It is common in simulation, that there are multiple simulators available, implementing various models of a system with miscellaneous advantages and disadvantages. For example, let the first model “M1” be beneficial in one part of the scenario. But in the other part, the second model “M2” would have preferable properties. Given the two Models (“M1” or “M2”), which simulator should be chosen?
In Fluid Animation (FA) the goal is to generate visual effects emulating the visual behavior of the fluid. The computational resources in FA are often limited in order to satisfy real-time restrictions or limitations on memory. In contrast the area of Computational Fluid Dynamics (CFD) focuses on models with high physical accuracy. There are many researches in the area of FA and CFD and a huge set of models known, all of them modelling fluids with different characteristics. For a bachelor/master thesis two different simulators, one for each model, should be implemented and a higher level simulator, using these two simulators, should be designed.
Knowledge in C++ is preferred. Skills in simulation are beneficial.
Literature:
Jerry Tessendorf, Simulating Ocean Water
Bei Interesse oder Fragen wendet euch an Marcus O’Connor (F454 /
oconnor@sim.uni-hannover.de)