This paper discusses the development of an automatic mesh generation t
echnique designed to effectively operate on multiple instruction, mult
iple data (MIMD) parallel computers. The paper first provides an overv
iew of a new octree-based procedure designed for the parallel generati
on of meshes of large numbers of elements directly from solid model re
presentations. The techniques required to implement such a procedure u
sing message passing on MIMD computers are emphasized. Technical issue
s considered include the region meshing algorithm, multiple octant mig
ration and incremental load balancing.