Abstract
Software application when executed on computer systems having disparate hardware features, result in dissimilar performance and power. Therefore, selecting systems with optimal performance and power values for application execution is an essential problem to address. However, access to many physical systems is required to collect performance power, which is a demanding task. Therefore, our first objective is to build an accurate prediction model for physical systems to achieve the second objective of computer system selection. To achieve the first objective, we propose a novel model, “cross performance and power prediction with scaling.” We develop a cross prediction model for physical systems by training a decision tree machine learning algorithm on performance and power datasets obtained from a large number of simulation systems built in the Gem5 simulator using emulation mode. However, design differences between Gem5 systems and physical systems lead to large prediction inaccuracies. We determine the application-specific “scaling factor” to compensate for the prediction inaccuracies and apply it to the predicted values for accurate physical systems predictions. We evaluate our model on well-known applications from SD-VBS and MiBench benchmarks achieving errors of 10–25% and 6–40% for performance and power for general-purpose systems. With accurate predictions for physical systems from our model, we achieve the second goal of computer system selection.
Similar content being viewed by others
References
Sifakis J (2011) A vision for computer science-the system perspective. https://doi.org/10.2478/s13537-011-0008-y. http://www.mathworks.com/products/simulink/. Accessed 23 Aug 2020
Ranganathan P (2010) Recipe for efficiency: principles of power-aware computing. Commun ACM 53(4):60. https://doi.org/10.1145/1721654.1721673
Aldossary M, Djemame K, Alzamil I, Kostopoulos A, Dimakis A, Agiatzidou E (2019) Energy-aware cost prediction and pricing of virtual machines in cloud computing environments. Future Gener Comput Syst 93:442. https://doi.org/10.1016/j.future.2018.10.027
Ma J, Yan G, Han Y, Li X (2016) An analytical framework for estimating scale-out and scale-up power efficiency of heterogeneous Manycores. IEEE Trans Comput 65(2):367. https://doi.org/10.1109/TC.2015.2419655
Lai Z, Lam KT, Wang CL, Su J (2017) PoweRock: power modeling and flexible dynamic power management for many-core architectures. IEEE Syst J 11(2):600. https://doi.org/10.1109/JSYST.2015.2499307
Sirbu A, Babaoglu O (2016) In: Proceedings of the 2016 international conference on high performance computing and simulation (HPCS 2016), Innsbruck, Austria (2016)
Ardalani N, Lestourgeon C, Sankaralingam K, Zhu X (2015) In: Proceedings of the annual international symposium on microarchitecture, MICRO. IEEE Computer Society, pp 725–737. https://doi.org/10.1145/2830772.2830780
Malakar P, Balaprakash P, Vishwanath V, Morozov V, Kumaran K (2019) In: Proceedings of PMBS 2018: performance modeling, benchmarking and simulation of high performance computer systems, held in conjunction with SC 2018: the international conference for high performance computing, networking, storage and analysis. Institute of Electrical and Electronics Engineers Inc., pp 33–44. https://doi.org/10.1109/PMBS.2018.8641686
Zheng X, John LK, Gerstlauer A (2017) LACross: learning-based analytical cross-platform performance and power prediction. Int J Parallel Program 45(6):1488. https://doi.org/10.1007/s10766-017-0487-0
Kim Y, Mercati P, More A, Shriver E, Rosin T In: 2017 IEEE/ACM international conference on computer-aided design (ICCAD) digest of technical papers November 13–16, 2017, Irvine Marriot, Irvine CA
Xu D, Shi Y, Tsang IW, Ong YS, Gong C, Shen X (2019). http://arxiv.org/abs/1901.00248
Binkert N, Beckmann B, Black G, Reinhardt SK, Saidi A, Basu A, Hestness J, Hower DR, Krishna T, Sardashti S, Sen R, Sewell K, Shoaib M, Vaish N, Hill MD, Wood DA, The gem5 simulator. ACM SIGARCH Comput Archit News 39(2)1 (2011). https://dl.acm.org. https://doi.org/10.1145/2024716.2024718. Accessed 4 May 2020
Cano-Cano J, Andújar FJ, Alfaro FJ, Sánchez JL (2019) Speeding up exascale interconnection network simulations with the VEF3 trace framework. J Parallel Distrib Comput 133:124. https://doi.org/10.1016/j.jpdc.2019.06.013
Greathouse JL, Loh GH (2018) In: IEEE/ACM international conference on computer-aided design, digest of technical papers, ICCAD. Institute of Electrical and Electronics Engineers Inc. https://doi.org/10.1145/3240765.3243484
Yang L, Wang L, Zhang X, Wang DL (2016) An approach to build cycle accurate full system VLIW simulation platform. Simul Model Pract Theory 67:14. https://doi.org/10.1016/j.simpat.2016.06.006
Yassin YH, Jahre M, Kjeldsberg PG, Aunet S, Catthoor F (2021) Fast and Accurate Edge Computing Energy Modeling and DVFS Implementation in GEM5 Using System Call Emulation Mod. J Signal Process Syst 93(1):33. https://doi.org/10.1007/S11265-020-01544-Z/TABLES/7
Zou Q, Zhu Y, Tan Y, Deng Y, Chen W (2021) Temporal characterization of memory access behaviors in SPEC CPU2017 workloads: Analysis and synthesis. Future Gener Comput Syst. https://doi.org/10.1016/J.FUTURE.2021.12.009
Cebrian JM, Barredo A, Caminal H, Moretó M, Casas M, Valero M (2020) Semi-automatic validation of cycle-accurate simulation infrastructures: the case for gem5-x86. Future Gener Comput Syst 112:832. https://doi.org/10.1016/J.FUTURE.2020.06.035
Li S, Ahn JH, Strong RD, Brockman JB, Tullsen DM, Jouppi NP (2013) The McPAT framework for multicore and manycore architectures simultaneously modeling power, area, and timing. Trans Archit Code Optim 10:11. https://doi.org/10.1145/2445572.2445577
Butko A, Bruguier F, Gamatié A, Sassatelli G, Novo D, Torres L, Robert M (2016) In: Proceedings - IEEE 10th international symposium on embedded multicore/many-core systems-on-Chip, MCSoC 2016. Institute of Electrical and Electronics Engineers Inc., pp 201–208. https://doi.org/10.1109/MCSoC.2016.20
Mankodi A, Bhatt A, Chaudhury B, Kumar R, Amrutiya A. In: 2020 IEEE international conference on electronics, computing and communication technologies (CONECCT) (IEEE, 2020), pp 1–6. https://doi.org/10.1109/CONECCT50063.2020.9198512
Breiman L, Friedman JH, Olshen RA, Stone CJ (2017). Classification and regression trees. https://doi.org/10.1201/9781315139470
Li B, Peng L, Ramadass B (2009) Accurate and efficient processor performance prediction via regression tree based modeling. J Syst Archit 55(10–12):457. https://doi.org/10.1016/j.sysarc.2009.09.004
Pedregosa F, Varoquaux G, Gramfort A, Michel V, Thirion B, Grisel O, Blondel M, Prettenhofer P, Weiss R, Dubourg V, Vanderplas J, Passos A, Cournapeau D, Brucher M, Perrot M, Duchesnay É, J Mach Learn Res 12, 2825 (2011). http://scikit-learn.sourceforge.net. Accessed 4 May 2020
Venkata S, Ahn I, Jeon D, Gupta A, Louie C, Garcia S, Belongie S, Taylor M (2009) In: Proceedings of the 2009 IEEE international symposium on workload characterization, IISWC : October 4–6 2009, Austin, TX, USA, IEEE
Guthaus MR, Ringenberg JS, Ernst D, Austin TM, Mudge T, Brown RB (2001) In: 2001 IEEE international workshop on workload characterization, WWC 2001. Institute of Electrical and Electronics Engineers Inc., pp. 3–14. https://doi.org/10.1109/WWC.2001.990739
Yu Z, Wang J, Eeckhout L, Xu C (2018) QIG: quantifying the importance and interaction of GPGPU architecture parameters. IEEE Trans Comput-Aided Des Integrated Circuits Syst 37(6):1211. https://doi.org/10.1109/TCAD.2017.2698026
Wikipedia contributors. Pearson correlation coefficient—Wikipedia, The free encyclopedia. https://en.wikipedia.org/w/index.php?title=Pearson_correlation_coefficient&oldid=922293481. Accessed 29 Nov 2019
Mankodi A, Bhatt A, Chaudhury B (2020) In: 2020 19th international symposium on parallel and distributed computing (ISPDC). IEEE , pp. 150–157. https://doi.org/10.1109/ISPDC51135.2020.00029. https://ieeexplore.ieee.org/document/9201765/. Accessed 23 Sept 2020
Terpstra Dan H, Jagode YH, Jack D (2010) In: Müller MS, Resch MM, Alexander S, Nagel WE (eds) Tools for high performance computing 2009, Springer, Berlin, pp 157–173
Terpstra (2013) PAPITopics: accessing the intel RAPL registers. https://icl.cs.utk.edu/projects/papi/wiki/PAPITopics:RAPL_Access. Accessed 2 Mar 2020
Author information
Authors and Affiliations
Corresponding author
Additional information
Publisher's Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Rights and permissions
About this article
Cite this article
Mankodi, A., Bhatt, A. & Chaudhury, B. Predicting physical computer systems performance and power from simulation systems using machine learning model. Computing 105, 935–953 (2023). https://doi.org/10.1007/s00607-022-01066-5
Received:
Accepted:
Published:
Issue Date:
DOI: https://doi.org/10.1007/s00607-022-01066-5