作者: J. Davison de St. Germain , Alan Morris , Steven G. Parker , Allen D. Malony , Sameer Shende
关键词:
摘要: The increasing complexity of high-performance computing environments and programming methodologies presents challenges for empirical performance evaluation. Evolving parallel distributed systems require technology that can be flexibly configured to observe different events associated data interest. It must also possible integrate evaluation techniques with the paradigms software engineering methods. This is particularly important tracking on projects involving many code teams over stages development. paper describes integration TAU XPARE tools in Uintah Computational Framework (UCF). Discussed use mapping associate low-level higher levels abstraction UCF regression testing provide a historical portfolio evolution application performance. A scalability study shows benefits integrating building large-scale applications.