This paper presents the development of a 2D solver for inviscid and viscous compressible flows using the spectral difference (SD) method for unstructured grids with mixed elements. A mixed quadrilateral and triangular grid is first refined using one-level h-refinement to generate a quadrilateral grid while keeping the curvature of boundary edges. The SD method designed for quadrilateral meshes ...

Here, U ≡ (v,ux,uy,e), superscript T stands for transpose, r is mass density, v ≡ 1/ρ, ux and uy are the components of flow velocity, e is the specific total energy, and Fx and Fy are fluxes in the xand ydirections respectively. A 2-D Riemann problem is Eq.(1) with a set of constant states for each region surrounding a point; for example, four constant states in the four quadrants in a structur...

A parallel volume rendering system for unstructured grid volume data is proposed in this paper. By implementing the mechanism for dynamic load balancing into the system, the authors solve the issue of load imbalance due to the view dependency and run-time features of early ray termination. An experimental implementation of the system achieved a 4.32-times performance

For a large class of irregular grid applications the structure of the mesh changes from one phase of the computation to the next Eventually as the graph evolves the adapted mesh has to be reparti tioned to ensure good load balance If this new graph is partitioned from scratch it will lead to an excessive migration of data among processors In this paper we present a new scheme for computing repa...

Interpolation of field data from unstructured meshes requires the potentially expensive identification of the finite element or volume within which the interpolating point lies. A number of geometric search algorithms have been proposed to reduce this expense at the cost of setting up and storing additional search tables. Using tetrahedral finite element models we show that a structured auxilia...

As the e ciency of parallel software increases it is becoming common to measure near linear speedup for many applications. For a problem size N on P processors then with software running at O(N=P ) the performance restrictions due to le i/o systems and mesh decomposition running at O(N) become increasingly apparent especially for large P . For distributed memory parallel systems an additional l...