Scalability Issues of In-situ Visualization in Parallel Simulation of Unsteady Flows

Grand challenge applications of 3-dimensional high resolution unsteady computational fluid dynamics result in huge amounts of data. To avoid significant bottlenecks of the storage and communication resources, efficient techniques for data extraction and preprocessing at the s ource have been realized in the parallel, network-distributed process chain called DSVR. Here the 3D data extraction is implemented as a parallel library and can be done in-situ during the numerical simulations, which avoids the storage of raw data for visualization. In this work we present, evaluate, and compare three techniques of parallel in-situ pathline extraction in distributed memory architectures. The gain in parallel scalability is achieved by an innovative trade-off betw een parallelization of partial tasks and asynchronous execution of suited serialized tasks. It has been shown that advanced parallelization schemes increase the scalability significantly