Sierpinski Curves for Parallel Adaptive Mesh Refinement in Finite Element and Finite Volume Methods
Oliver Meister · 2016
The thesis investigates parallel adaptive structured triangular grids that are generated corresponding to the Sierpinski space-filling curve. Stack- and stream-based data structures grant memory efficiency and scalable heuristics for hybrid parallelization. Based on this approach, the software framework sam(oa)² provides fast adaptive mesh refinement on high-performance systems. Excellent performance is shown on up to 8,000 cores for simulations of oil recovery and tsunami wave propagation.