Parallel Linear Octree Meshing with Immersed Surfaces

A parallel octree-based mesh generation method is proposed to create reasonable-quality, geometry-adapted unstructured hexahedral meshes automatically from triangulated surface models. We present algorithms for the construction, 2:1 balancing and meshing large linear octrees on distributed memory machines. Our scheme uses efficient computer graphics algorithms for surface detection, allowing us to represent complex geometries. Isogranular analysis is performed on a variety of input surfaces and demonstrates good scalability. Our implementation is able to execute the 2:1 balancing operations over 4.0 × 10⁸ octants on 128 cores in less than 10 seconds per 2 × 10⁵ octants/core.