Parallel Performance of CFD Applications and the Ubiquitous Need for HPC with High Fidelity, Multidisciplinary Analysis and Optimization (MDO)
Mark D. Kremenetsky, Srinivas Kodiyalam · 2010
We discuss possible ways to improve CFD application software performance by highlighting key features of the HPC clusters, including, multicore processor performance, host channel adapters (HCA), multi-rail networks, and message passing (MPI) implementations and more as well as briefly highlight the ever-growing need for HPC for solving computing intensive MDO problems.