High Performance Computing Benchmark Tool for Parallel Processing of Large Models

Benchmarks for parallel processing of large models is an urgent need for High Performance Computing (HPC) as today’s model size reaches millions of degrees of freedom. Explicit solvers as in the case of crash dynamics or fluid dynamics do not require matrix based equation solvers and inherently exhibit good scalability on large numbers of processors. Where as analysis requiring implicit solvers in thermally driven structural response utilize large matrix equation solvers. The thermal radiation exchange problem matrices also may be dense and unsymmetric. Publicly available benchmarks such as Standard Performance Evaluation Corporation (SPEC) benchmarks do not address performance in areas of large matrix equation solvers for large number of nodes or for dense and unsymmetric matrix solvers. This study attempts to develop a benchmark tool that can be applied to dense and unsymmetrical matrix solvers for avionics heat transfer problems, thermally driven structural response problems and radiation exchange problems. In particular it addresses implicit solution algorithms with production models and benchmarks for indefinite matrices and configurations scaling for large dense matrix problems in distributed high performance system in shared, distributed and mixed memory conditions.