next up previous contents
Next: Multi-Time Step Subcycling Results Up: uiuc_thesis Previous: Conclusions   Contents


Having gained valuable experience from our one-dimensional analysis of the previous chapter, we can now apply the various optimization methods to the more complex, two-dimensional problems. We begin our analysis with the multi-time step nodal subcycling algorithm using a simple mode I test problem where we vary the number of non-subcycled to subcycled nodes. Next, we analyze the dynamic cohesive element insertion algorithm on a simplified 2-D problem. Using two cohesive element selection criteria, the bounding box method and the stress-based selection method, we apply the insertion algorithm to a series of complex problems, from which we are able to gauge the accuracy and timing information of the solutions. Finally, we use the Charm++ code parallelization technique to take advantage of the benefits of multiple processors in solving a given problem.


Mariusz Zaczek 2002-10-13