Evaluation of next-generation high-order compressible fluid dynamic solver on cloud computing for complex industrial flows

Industrially relevant computational fluid dynamics simulations frequently require vast resources that are only available to governments, wealthy corporations, and institutions. Thus, in many contexts realities, high-performance computing grids cloud on demand should be evaluated as viable alternatives conventional clusters. In this work, we present the analysis of time-to-solution cost an entropy stable collocated discontinuous Galerkin (SSDC) compressible framework Ibex, on-premises cluster at KAUST, Amazon Web Services Elastic Compute Cloud for complex flows. SSDC is a prototype next generation frameworks developed following road map established by NASA CFD vision 2030. We simulate flow problems using high-order accurate fully-discrete algorithms. terms time-to-solution, delivers best performance, with Graviton2 processors based Arm architecture being fastest. However, results also indicate Ibex nodes AMD Rome deliver good close those observed Cloud. Furthermore, computations performed currently less expensive than cloud. Our findings could used develop guidelines selecting realistic problems.