Usage of the SCALASCA toolset for scalable performance analysis of large-scale parallel applications

Abstract

scalasca is a performance toolset that has been specifically designed to analyze parallel application behavior on large-scale systems, but is also well-suited for small- and medium-scale hpc platforms. scalasca offers an incremental performance-analysis process that integrates runtime summaries with in-depth studies of concurrent behavior via event tracing, adopting a strategy of successively refined measurement configurations. A distinctive feature of scalasca is its ability to identify wait states even for very large processor counts. The current version supports the mpi, Openmp and hybrid programming constructs most widely used in highly-scalable hpc applications.