Distributed Computing in ModelCenter Using Cloud-Based Virtual Machines | THE BOEING COMPANY
As Multidisciplinary Design Analysis and Optimization (MDAO) models continue to pull fidelity forward in the design process, model run times and computing requirements have increased exponentially. These MDAO models put an incredible strain on the computer resources available to engineers who may not have access to or a model that can operate successfully in a high performance computing (HPC) environment. This presentation will delve into a distributed computing approach used in support of Boeing/AFRL OPTIMUS program using ModelCenter and cloud-based virtual machines. The approach allowed sufficient model parallelization and load balancing to reduce overall run time with limited resources. It also enabled a quick setup time and eliminated the need to use more expensive computing hardware.