TOWARDS AN AUTONOMOUS FRAMEWORK FOR HPC OPTIMIZATION: A STUDY OF PERFORMANCE PREDICTION USING HARDWARE COUNTERS AND MACHINE LEARNING

TOWARDS AN AUTONOMOUS FRAMEWORK FOR HPC OPTIMIZATION: A STUDY OF PERFORMANCE PREDICTION USING HARDWARE COUNTERS AND MACHINE LEARNING

Gritz, Matheus; Silva, Gabrieli; Klôh, Vinicius; Schulze, Bruno; Ferro, Mariza

Artigo Completo:

As the high processing computing becomes even more critical for scientific research across various fields, increasing performance without raising the energy consumption levels becomes an essential task in order to warrant the financial viability of exascale systems. This work presents the first step towards understanding how the many computational requirements of benchmark applications relate to the overall runtime through a machine learning model and how that can be used for the development of an autonomous framework capable of scaling applications to have an optimal trade-off- between performance and energy consumption.

As the high processing computing becomes even more critical for scientific research across various fields, increasing performance without raising the energy consumption levels becomes an essential task in order to warrant the financial viability of exascale systems. This work presents the first step towards understanding how the many computational requirements of benchmark applications relate to the overall runtime through a machine learning model and how that can be used for the development of an autonomous framework capable of scaling applications to have an optimal trade-off- between performance and energy consumption.

Palavras-chave:

DOI: 10.5151/spolm2019-196

Referências bibliográficas

• [1] SIMON, H. D. Barriers to Exascale Computing. Berlin, Heidelberg: Springer Berlin Heidelberg, 2013. 1{3 p. ISBN 978-3-642-38718-0. Disponvel em: <https://doi.org/10.1007/978-3-642-38718-0 1>. 1 [2] BERGMAN, K. et al. Exascale computing study: Technology challenges in achieving exascale systems. Defense Advanced Research Projects Agency Information Processing Techniques Ofice (DARPA IPTO), Tech. Rep, v. 15, 2008. 1 [3] ALVIN, K. et al. On the path to exascale. International Journal of Distributed Systems and Technologies, IGI Global, v. 1, n. 2, p. 1{22, 2010. 1 [4] RAJOVIC, N. et al. Supercomputing with commodity cpus: Are mobile socs ready for hpc? In: Proceedings of the International Conference on High Per- formance Computing, Networking, Storage and Analysis. New York, NY, USA: ACM, 2013. (SC '13), p. 40:1{40:12. ISBN 978-1-4503-2378-9. Disponvel em: <http://doi.acm.org/10.1145/2503210.2503281>. 1 [5] MESSINA, P. The exascale computing project. Computing in Science Engineering, v. 19, n. 3, p. 63{67, May 2017. ISSN 1521-9615. 1 [6] SILVA, G. et al. Smcis: um sistema para o monitoramento de aplicac oes cientficas em ambientes hpc. SBC, S~ao Paulo-SP, p. 277{288, Outubro 2018. Http://www2.sbc.org.br/wscad/current/anais/anais-wscad-2018.pdf. 2 [7] POVOA, L. V. et al. Predic~ao de consumo energetico com base na utilizac~ao de recursos computacionais. 2013. 2, 4 [8] FERREIRA, A. R. et al. Um modelo analtico para estimar o consumo de energia de sistemas multi-camadas no nvel de transac~ao. Universidade Federal de Goias, 2017. 2, 4 [9] SIEGMUND, N. et al. Performance-in uence models for highly configurable systems. In: ACM. Proceedings of the 2015 10th Joint Meeting on Foundations of Software En- gineering. [S.l.], 2015. p. 284{294. 2, 4 [10] JAIANTILAL, A.; JIANG, Y.; MISHRA, S. Modeling cpu energy consumption for energy eficient scheduling. In: ACM. Proceedings of the 1st Workshop on Green Com- puting. [S.l.], 2010. p. 10{15. 2, 4 [11] BALLADINI, J. et al. Metodologa para predecir el consumo energetico de checkpoints en sistemas de hpc. In: XX Congreso Argentino de Ciencias de la Computacion (Buenos Aires, 2014). [S.l.: s.n.], 2014. 2, 4 [12] WU, X. et al. Using performance-power modeling to improve energy eficiency of hpc applications. Computer, IEEE, v. 49, n. 10, p. 20{29, 2016. 2, 4 [13] ASANOVIC, K. et al. A View of the Parallel Computing Landscape. Commun. ACM, ACM, New York, NY, USA, v. 52, n. 10, p. 56{67, oct 2009. ISSN 0001-0782. Disponvel em: <http://doi.acm.org/10.1145/1562764.1562783>. 2, 12 [14] FERRO, M. Avaliao de Sistemas de Computao Massivamente Paralela e Distribuda: Uma metodologia voltada aos requisitos das aplicaes cientficas. Tese (Tese de Doutorado) | Laboratrio Nacional de Computao Cientfica, Petrpolis - RJ, Maio 2015. 3 [15] FERRO, M.; EVOY, G. M.; SCHULZE, B. Analysis of high performance applications using workload requirements. In: . High Performance Computing for Computa- tional Science { VECPAR 2016: 12th International Conference, Porto, Portugal, June 28-30, 2016, Revised Selected Papers. Cham: Springer International Publishing, 2017. p. 7{10. ISBN 978-3-319-61982-8. Disponvel em: <https://doi.org/10.1007/978-3-319- 61982-82>:3 [16] KALTENECKER, C. Comparison of analytical and empirical performance models: A case study on multigrid systems. 2016. 4 [17] MARTINEZ, V. et al. Performance improvement of stencil computations for multicore architectures based on machine learning. Procedia Computer Science, Elsevier, v. 108, p. 305{314, 2017. 4, 5 [18] SIEGMUND, N. et al. Performance-in uence models for highly configurable systems. In: ACM. Proceedings of the 2015 10th Joint Meeting on Foundations of Software En- gineering. [S.l.], 2015. p. 284{294. 4 [19] BERRAL, J. L.; GAVALDA, R.; TORRES, J. Power-aware multi-data center management using machine learning. In: IEEE. 2013 42nd International Conference on Parallel Processing. [S.l.], 2013. p. 858{867. 4 [20] BERRAL, J. L.; GAVALDA, R.; TORRES, J. Adaptive scheduling on power-aware managed data-centers using machine learning. In: IEEE COMPUTER SOCIETY. Pro- ceedings of the 2011 IEEE/ACM 12th International Conference on Grid Computing. [S.l.], 2011. p. 66{73. 4 [21] BERRAL, J. L. et al. Towards energy-aware scheduling in data centers using machine learning. In: ACM. Proceedings of the 1st International Conference on energy-Eficient Computing and Networking. [S.l.], 2010. p. 215{224. 4 [22] BERRAL, J. L.; GAVALDA, R.; TORRES, J. Empowering automatic data-center management with machine learning. In: ACM. Proceedings of the 28th Annual ACM Symposium on Applied Computing. [S.l.], 2013. p. 170{172. 4 [23] BHIMANI, J. et al. Fim: performance prediction for parallel computation in iterative data processing applications. In: IEEE. 2017 IEEE 10th International conference on cloud computing (CLOUD). [S.l.], 2017. p. 359{366. 5 [24] MELO, A. C. D. The new linux'perf'tools. In: Slides from Linux Kongress. [S.l.: s.n.], 2010. v. 18. 5 [25] MITCHELL, T.; UTGOFF, P.; BANERJI, R. Learning by Experimentation: Acquiring and Refining Problem-Solving Heuristics. In: MICHALSKI, R.; CARBONNEL, J.; MITCHELL, T. (Ed.). Machine Learning: An Artificial Intelligence Approach. Palo Alto, CA: Tioga, 1983. 7 [26] FERRO, M.; LEE, H. D.; ESTEVES, S. C. Intelligent Data Analysis: A Case Study of the Diagnostic Sperm Processing. In: Proceedings ACIS International Conference on Computer Science, Software Engineering, Information Technology, e-Business, and Applications (CSITeA-02). Foz do Iguau, Brazil: [s.n.], 2002. p. 116{120. 7 [27] WITTEN, I. H.; FRANK, E. Data Mining: Practical Machine Learning Tools and Techniques. Second. [S.l.]: Morgan Kaufmann, 2005. 525 p. (Morgan Kaufmann Series in Data Management Systems). ISBN 0120884070. 7 [28] PEDREGOSA, F. et al. Scikit-learn: Machine learning in Python. Journal of Machine Learning Research, v. 12, p. 2825{2830, 2011. 7 [29] KOTSIANTIS, S.; KANELLOPOULOS, D.; PINTELAS, P. Data preprocessing for supervised leaning. International Journal of Computer Science, Citeseer, v. 1, n. 2, p. 111{117, 2006. 7 [30] MOTODA, H.; LIU, H. Feature Selection, Extraction and Construction. In: Proceed- ings of Foundation of Data Mining, PAKDD. [S.l.: s.n.], 2002. p. 67{72. 8

Como citar:

Gritz, Matheus; Silva, Gabrieli; Klôh, Vinicius; Schulze, Bruno; Ferro, Mariza; "TOWARDS AN AUTONOMOUS FRAMEWORK FOR HPC OPTIMIZATION: A STUDY OF PERFORMANCE PREDICTION USING HARDWARE COUNTERS AND MACHINE LEARNING", p-2721-2734. In: Anais do XIX Simpósio de Pesquisa Operacional & Logística da Marinha. São Paulo: Blucher, 2020.
ISSN 21756295, DOI 10.5151/spolm2019-196