A prototype framework for parallel visualization of large flow data

•In this paper, we present a prototype integrated framework for parallel visualization of large flow data. Still with direct access to MPI functionalities, it adopts DIY (Do it yourself) as the primary parallel computing platform, which is a layer on top of MPI to offer a set of components and mechanisms amenable to data visualization. This framework takes OSUFlow, a suite of geometry-based flow visualization algorithms, as the vector field visualization engine. The DIY-OSUFlow combination supports parallel visualization of flow data defined on Cartesian, rectilinear, and curvilinear grids, specifically in creating streamlines and pathlines. Furthermore, the framework exploits VTK (Visualization Toolkit) for versatile data input, advanced graphics rendering, and flexible scene interaction. In this way, the broad user community of VTK can take advantage of a variety of parallel flow visualization capabilities of DIY-OSUFlow (e.g., a stream surface as a triangulated connection of a collection of streamlines originating from seeds along a rake line). Preliminary results show that this framework is capable of exploiting the horsepower of processors to accelerate data processing and visualization for explorative analysis of massive steady/unsteady volume flows.