This title appears in the Scientific Report : 2014 

Assessing Measurement and Analysis Performance and Scalability of Scalasca 2.0
Zhukov, Ilya (Corresponding Author)
Wylie, Brian J. N.
Jülich Supercomputing Center; JSC
Euro-Par 2013: Parallel Processing Workshops
Berlin, Heidelberg Springer Berlin Heidelberg 2014
627 - 636
978-3-642-54419-4 (print)
978-3-642-54420-0 (electronic)
10.1007/978-3-642-54420-0_61
Euro-Par 2013: Parallel Processing Workshops, Aachen (Germany), 2013-08-26 - 2013-08-27
Contribution to a book
Contribution to a conference proceedings
Computational Science and Mathematical Methods
Lecture Notes in Computer Science 8374
Please use the identifier: http://dx.doi.org/10.1007/978-3-642-54420-0_61 in citations.
The Scalasca toolset was developed to provide highly scalable performance measurement and analysis of scientific applications on current HPC platforms, including leadership systems such as IBM BlueGene/Q and more traditional Linux clusters. Its primary focus is support for C/C++/Fortran applications using MPI and OpenMP, and mixed-mode combinations thereof, offering detailed call-path profiles for each process and thread produced by runtime summarization or augmented with wait-state analysis of event traces. A new generation of Scalasca (2.0) uses the community-developed infrastructure comprising of Score-P and associated components, while continuing to provide the previous functionality. By comparing the new version of Scalasca with its predecessor, using the applications from the NPB3.3-MZ-MPI benchmark suite, we validate core functionality and assess overheads and scalability. Although adequate for general use, various aspects are identified for further improvement, particularly for larger scales.