Parallel components for PDEs and optimization: some issues and experiences

摘要: High-performance simulations in computational science often involve the combined software contributions of multidisciplinary teams scientists, engineers, mathematicians, and computer scientists. One goal component-based engineering large-scale scientific is to help manage such complexity by enabling better interoperability among codes developed different groups. This paper discusses recent work on building component interfaces implementations parallel numerical toolkits for mesh manipulations, discretization, linear algebra, optimization. We consider several motivating applications involving partial differential equations unconstrained minimization demonstrate this approach evaluate performance.