In some cases, parallel execution is considered, which tremendously reduces the total analysis turnaround time. Typically, this occurs when virtual experiments for a simulation code take a long time to complete, as is often the case with a crash analysis or CFD computation. Therefore, the most important point to making Optimus methods applicable to real industry problems is the ability to parallelize the computation of the simulation workflow.
Optimus allows for parallelization on both Simulation Workflow Level and Method level. Super-linear gains in execution time can be achieved. This is achieved in a way that is uniquely transparent for you and allows for an optimal and consistent execution of robustness and reliability, optimization and design exploration methods. Interfaces to all Resource Management Systems are available, including LSF™ from platform Computing, Condor, OpenPBS, TORQUE, PBSPro™ from Altair, Grid Engine™ from Oracle and many more. Execution of the analysis programs can take place on any computer platform and operating system irrelevant to the location where Optimus is installed, or where the analysis process has been defined.
Optimus offers a unique workflow calculation parallelization capability. This capability combines experiment-level parallelism, which launches different experiments on different CPUs, and workflow-level parallelism, which launches the different simulation codes in a workflow on different CPUs, to maximize overall CPU usage.