Abstract
Developing efficient parallel programs for supercomputers is a challenging task. It requires insight into the application, the parallelization concepts, as well as the parallel architectures. Performance analysis tools such as Periscope, an automatic performance analysis tool currently under development at Technische Universität München, help the programmer in detecting performance bottlenecks. The goal of the ISAR project is to enhance the existing Periscope research prototype and deliver a production version. This paper focuses on the evaluation of Periscope's main features based on two large scale simulation codes. © 2010 Springer-Verlag.
Author supplied keywords
Cite
CITATION STYLE
Benedict, S., Brehm, M., Gerndt, M., Guillen, C., Hesse, W., & Petkov, V. (2010). Automatic performance analysis of large scale simulations. In Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics) (Vol. 6043 LNCS, pp. 199–207). https://doi.org/10.1007/978-3-642-14122-5_24
Register to see more suggestions
Mendeley helps you to discover research relevant for your work.
