Considerations of computational optimality in parallel algorithms for grid generation

A parallel algorithm for structured grid generation is described. Theoretical and practical considerations of various definitions of parallel speed-ups are given. The objective of the paper is to quantify gains realized from the use of various parallel computing modes and to suggest an optimal computing paradigm in the context of grid generation. It is established that computational efficiency is a more realistic indicator of the capability of a parallel grid generation algorithm than speed-up alone.