Abstract
In this work, we propose an Automatic Performance Tracking System for analyzing the changes in execution performance and finding the source code modifications that cause the degradation of performance portability. The proposed system works in order to support evolving a large-scale numerical application while maintaining its performance portability across multiple target computing systems. By evaluating the performance of an application on every computing system, the proposed system helps application developers find the source code modifications that degrade the execution performance on a computing system. The proposed system also retrieves multiple versions of an application from its code repository, and automatically executes them on a newly added computing system. As a result, application developers are able to analyze how the source code modifications in the past affect the performance on the new computing system. Based on the evaluation results, the application developers can review the source code changes to improve the performance portability of the HPC application through the system.
Access this chapter
Tax calculation will be finalised at checkout
Purchases are for personal use only
References
Zhu, H., Hall, P.A.V., May, J.H.R.: Software unit test coverage and adequacy. ACM Comput. Surv. 29(4), 366–427 (1997)
Kim, S., Zimmermann, T., Pan, K., Whitehead, E.J.: Automatic identification of bug-introducing changes. In: 21st IEEE/ACM International Conference on Automated Software Engineering, 2006. ASE ’06, pp. 81–90 (2006)
Williams, C.C., Hollingsworth, J.K.: Automatic mining of source code repositories to improve bug finding techniques. IEEE Trans. Soft. Eng. 31(6), 466–480 (2005)
http://cvs.nongnu.org/. Cvs - concurrent versions system
http://gitscm.com/. Git - the fast version control system
Shigetam, M., Watanabe, T.: Growth model of binary alloy nanopowders for thermal plasma synthesis. J. Appl. Phys. 108(4), 043306–043306–15 (2010)
Watson, G.R., Rasmussen, C.E., Tibbitts, B.R.: An integrated approach to improving the parallel application development process. In: IEEE International Symposium on Parallel Distributed Processing, 2009. IPDPS 2009, pp. 1–8, May 2009
Acknowledgements
The authors would like to thank Prof. Shigeta of Osaka University for allowing us to use the application. This work is partially supported by JST CREST “An Evolutionary Approach to Construction of a Software Development Environment for Massively-Parallel Heterogeneous Systems.”
Author information
Authors and Affiliations
Corresponding author
Editor information
Editors and Affiliations
Rights and permissions
Copyright information
© 2016 Springer International Publishing AG
About this paper
Cite this paper
Hirasawa, S., Takizawa, H., Kobayashi, H. (2016). An Automatic Performance Tracking System for Large-Scale Numerical Applications. In: Resch, M., Bez, W., Focht, E., Patel, N., Kobayashi, H. (eds) Sustained Simulation Performance 2016. Springer, Cham. https://doi.org/10.1007/978-3-319-46735-1_10
Download citation
DOI: https://doi.org/10.1007/978-3-319-46735-1_10
Published:
Publisher Name: Springer, Cham
Print ISBN: 978-3-319-46734-4
Online ISBN: 978-3-319-46735-1
eBook Packages: Mathematics and StatisticsMathematics and Statistics (R0)