Eliminating Serial Bottlenecks Not Just in the Flow Solving But Also in Mesh Generation and Geometry Management

This paper reports progress towards the integration of a Level Set based geometry kernel with an octree-based cut-Cartesian mesh generator, RANS flow solver and post-processing within a single piece of software. Recent research has concentrated on parallelizing the entire process with commodity PC clusters as the target. The aim of this work is to eliminate all serial bottlenecks from the CFD Process — including the geometry kernel and its editing & management during design optimization. The ultimate aim is to allow rapid prototyping design optimization to take place on geometries of arbitrary size in a spirit of a real time computer game.