An Interactive Isosurface Visualization Cluster

摘要

Extraction and rendering of isosurfaces from large volume data is a computationally expensive task. To achieve interactive isosurface visualization, we have developed a parallel isosurface extraction pipeline, which couples the extraction process with a multiresolution strategy to relieve the host PC from the isosurface transport and rendering bottlenecks. Such an isosurface visualization cluster design can scale with the problem size. The paper presents a comprehensive study of the interplay between the compute nodes (a PC cluster), the host node (a graphics-enhanced PC), the data (voxels and triangles), and user actions (isovalue and view changes) to derive a resource-conscious, user-centered metric for designing effective solutions for large-scale isosurface visualization.