Generating Multibillion Element Unstructured Meshes on Distributed Memory Parallel Machines

Figure 6

The dependence of speedup on number of processors and coarse mesh size to fine mesh size ratio.