Presentation Details |
|||||
Name: | (PP03) ProfiT-HPC - A Profiling Toolkit for HPC in Tiers 2 & 3 | ||||
Time: | Tuesday, June 20, 2017 03:15 pm - 03:45 pm |
||||
Room: | Booth #L-212 | ||||
Breaks: | 03:15 pm - 03:45 pm Coffee Break | ||||
Presenter: | Mohammad Siahatgar, Leibniz Universität Hannover | ||||
Abstract: | High Performance Computing (HPC) has been established as a standard method in many scientific fields – especially in the natural and engineering sciences. More and more scientific disciplines, e.g., bio-informatics or social studies, newly discover the possibilities of HPC for their research. As a direct consequence the number of researchers writing and using HPC applications without a profound background knowledge of HPC systems is rising. At the same time HPC resources gain complexity, thus further increasing the knowledge gap regarding their efficient usage.
Often enough researchers are content as soon as they receive the results of their computations in an acceptable time frame, without recognising the possibilities of optimising their resource usage through performance engineering. A higher number of or more detailed computations could often already be achieved by changes in the set-up of applications on a given system. A lack of one-on-one support for users in Tier-3 centres aggravates the problem.
The main goal of this project is to increase the awareness of performance issues in all user communities. To achieve this, automated run-time profiles of their applications will be provided to the users. These profiles will be presented in a way that is useful to newcomers and experienced users alike and is aimed to motivate to investigate possibilities to optimise applications. In conjunction with additional support, e.g., by the consultancy network of the HLRN, scientific result generation could be improved. At the same time, HPC providers could benefit from better insight into the usage of their systems. Authors Jörg Benke, Universität Hamburg Sven Bingert, Gesellschaft für wissenschaftliche Datenverarbeitung mbH Göttingen Chrisitan Boehme, Gesellschaft für wissenschaftliche Datenverarbeitung mbH Göttingen Vanessa End, Gesellschaft für wissenschaftliche Datenverarbeitung mbH Göttingen Nikolai Kornev, Universität Rostock Guido Laubender, Zuse Institute Berlin Stephan Olbrich, Universität Hamburg Fabian Pflug, Leibniz Universität Hannover - Computational Health Informatics Alexander Reinefeld, Zuse Institute Berlin Mohammad Siahatgar, Leibniz Universität Hannover - Computational Health Informatics Thomas Steinke, Zuse Institute Berlin Hinnerk Stüben, Universität Hamburg Matthias Walter, Universität Rostock Tobias Watermannn, Zuse Institute Berlin Ramin Yahyapour, Gesellschaft für wissenschaftliche Datenverarbeitung mbH Göttingen Gabriele von Voigt, Leibniz Universität Hannover - Computational Health Informatics
|
||||
Download | PP03_Siahatgar.pdf (311 KB) |
||||