作者: Lois Curfman McInnes , Tamara Dahlgren , Jarek Nieplocha , David Bernholdt , Ben Allan
DOI: 10.1088/1742-6596/78/1/012046
关键词: Component-based software engineering 、 Simulation software 、 Computational science 、 Software engineering 、 Implementation 、 Control reconfiguration 、 Component (UML) 、 Common Component Architecture 、 Plug and play 、 Computer science 、 Massively parallel
摘要: This paper introduces three component technology initiatives within the SciDAC Center for Technology Advanced Scientific Component Software (TASCS) that address ever- increasing productivity challenges in creating, managing, and applying simulation software to scientific discovery. By leveraging Common Architecture (CCA), a new standard high-performance computing, these tackle diculties at dierent but related levels development of component-based software: (1) deploying applications on massively parallel heterogeneous architectures, (2) investigating approaches runtime enforcement behavioral semantics, (3) developing tools facilitate dynamic composition, substitution, reconfiguration implementations parameters, so application scientists can explore tradeos among factors such as accuracy, reliability, performance.