The Fatigue Damage Software Parallelization

The purpose of this paper is to describe the parallelization of PragTic - a fatigue damage simulator. There is a discussion about the former version inefficiency (caused by frequent I/O operations), parallel implementation and parallelized PragTic scalability testing. The parallelization is done at the node level where the nodes are defined in the context of the finite element mesh. We also ran a simple model example (hundreds of nodes), real world example (millions of nodes) and our approach is to test using numerical experiments where we achieve weak parallel scalability. At the end of the article, the results of numerical experiments are depicted in tables and graphs.