Parallelizing engineering workflows enables engineering teams to significantly increase simulation throughput, taking maximum advantage of all available computing resources. Making more effective use of software and hardware infrastructure scales up the amount of virtual prototyping that can be completed within the ever shorter time window that is available for product engineering.

Reducing turnaround times

Reducing turnaround times is particularly useful when dealing with computationally intensive engineering disciplines, such as automotive crash analysis or computational fluid dynamics. Our proven simulation workflow parallelization capabilities support engineering teams in resolving industry challenges in a much shorter time frame. Gains in execution time can be achieved in a transparent way. They allow for an optimal and consistent execution of design space exploration, optimization, and robustness & reliability methods.

A straightforward queuing resource management system

Optimus Parallel Services (OPS) is a lightweight queuing system that is easy to install and configure in a network environment without advanced IT infrastructure. It allows an Optimus client to submit simulation jobs to different nodes or workstations in the network, without spending hours to set up a 3rd party resource management system such as IBM Platform LSF, HTCondor, OpenPBS, TORQUE Resource Manager, Altair PBS Professional and Univa Grid Engine, etc. (though Optimus is integrated with these as well). With OPS, users are able to specify which computer masters the cluster and which nodes run a specific analysis tool (Nastran, Abaqus, etc.).

Balancing available resources

Thanks to the integration with various 3rd party resource management systems (RMS), our technologies allow virtual prototyping simulations to be submitted to and balanced over the available computation resources in a heterogeneous ICT infrastructure - without any user intervention. In addition to our OPS lightweight queuing system, interfaces are available for IBM Platform LSF, HTCondor, OpenPBS, TORQUE Resource Manager, Altair PBS Professional and Univa Grid Engine, etc. Simulations can be executed on any computer platform and operating system, irrespective of the physical location where our technologies are installed or where the simulation process has been defined.

Maximizing overall CPU usage

Our parallelization technologies enable engineering teams to go one step further. Combined with our flexible RMS integration, parallelization can be implemented on various levels. To start with, our range of unique workflow parallelization capabilities makes it possible to launch simulation experiments in parallel – typically as part of a Design of Experiments (DOE) method. Next, workflow-level parallelization technologies empower engineering teams to launch the various engineering simulation tasks that are part of a specific workflow on different CPUs. Ultimately, both experiment-level and workflow-level parallelization can be combined into a unique strategy that maximizes overall CPU usage.


Discover how »