Considerations for LS-DYNA Workflow Efficiencies in an HPC Linux Environment

Manufacturing industry and research organizations continue to increase their investments in structural analysis and impact simulations such that the growing number of LS-DYNA users continues to demand more from computer system environments. These LS-DYNA workflow demands typically include rapid single job turnaround and multi- job throughput capability for users with diverse application requirements in a high-performance computing (HPC) environment. Additional complexity arises from the need for many LS-DYNA HPC environments to coexist with other computer aided engineering (CAE) software for a variety of multi-physics and multi-scale structural and computational fluid dynamics analyses that all compete for the same HPC workflow resources. For today’s economics of HPC, these resources such as CPU cycles, memory, system bandwidth and scalability, storage and I/O, and file and data management – must deliver the highest levels of CAE productivity and HPC reliability that is possible from a Linux platform environment. This presentation examines workflow efficiencies of CAE simulations for relevant applications in LS-DYNA for an HPC Linux platform developed by SGI. LS-DYNA modeling parameters such as model size, element types, schemes of implicit and explicit (and coupled), and a variety of simulation conditions can produce a wide range of computational behavior and data management requirements, such that careful consideration should be given to how system resources are configured, deployed, and allocated to meet increasing user demands. The HPC system technology of the SGI® AltixTM clusters and servers, based on Linux® and Itanium® 2 from Intel®, have demonstrated both LS-DYNA turnaround and throughput achievement that includes industrial-sized examples. In addition, SGI simulation data management technology of HPC file systems and data storage tools, are providing the LS-DYNA workflow management necessary to maximize user productivity, and enable a user roadmap of increasing LS-DYNA modeling fidelity.