HPC Considerations for Scalable Multidiscipline CAE Applications on Conventional Linux Platforms
This paper examines HPC workload efficiencies for sample multidiscipline LS-DYNA applications on a conventional HPC Linux platform with proper balance for I/O treatment. Model parameters such as 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 I/O management requirements. Consideration must be given to how HPC resources are configured and deployed, in order to satisfy growing LS-DYNA user requirements for increased fidelity from multidiscipline CAE.
https://www.dynalook.com/conferences/european-conf-2007/hpc-considerations-for-scalable-multidiscipline.pdf/view
https://www.dynalook.com/@@site-logo/DYNAlook-Logo480x80.png
HPC Considerations for Scalable Multidiscipline CAE Applications on Conventional Linux Platforms
This paper examines HPC workload efficiencies for sample multidiscipline LS-DYNA applications on a conventional HPC Linux platform with proper balance for I/O treatment. Model parameters such as 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 I/O management requirements. Consideration must be given to how HPC resources are configured and deployed, in order to satisfy growing LS-DYNA user requirements for increased fidelity from multidiscipline CAE.
3.5.2.pdf
— 111.6 KB