Hybrid Machine Learning Model for Cache Performance Prediction and Design Space Exploration
DOI:
https://doi.org/10.22456/2175-2745.146689Keywords:
hybrid machine learning model, L1 data cache, miss rate prediction, design space exploration, cache configurations, applications’ profilingAbstract
This work proposes a hybrid machine-learning model to predict a level 1 data (L1D) cache miss rate value of diversified applications with different cache configurations and application characteristics. The proposed model quickly predicts a miss rate of the L1D cache. Moreover, the proposed model can simultaneously predict the miss rate value of a new(unseen) application for twenty-eight cache configurations as precisely as possible without extra simulations. Thus, it quickly provides a miss rate value faster than a simulator. Additionally, we could perform an application’s cache design space exploration by getting predicted miss rate values from the proposed model, like the simulator. The proposed model’s mean absolute and root mean squared errors are 0.6% and 1.1% by ten-fold cross-validation. Speed up was achieved 12x to 148x compared to the architectural simulator for a miss rate of twenty-eight cache configurations.
Downloads
References
[1] ALVAREZ, D. et al. Optimizing L1 cache for embedded systems through grammatical evolution. Soft Computing, v. 20, n. 6, p. 2451–2465, 2016.
[2] NAWINNE, I.; PARAMESWARAN, S. A survey on exact cache design space exploration methodologies for application specific SoC memory hierarchies. In: 8th International Conference on Industrial and Information Systems. [S.l.]: IEEE, 2013. p. 332–337.
[3] DAVE, H.; KOTAK, N. Critical analysis of cache memory performance concerning miss rate and power consumption. International Journal of Embedded Systems, v. 15, n. 6, p. 516–524, 2022.
[4] BALASUBADRA, K.; SHANTHI, A.; SRINIVASAN, V. Hybrid design space exploration methodology for application specific system design. International Journal of New Computer Architectures and Their Applications, v. 7, n. 3, p. 102–111, 2017.
[5] DAVE, H.; KOTAK, N. An analysis of cache configuration’s impacts on the miss rate of big data applications using gem5. Serbian Journal of Electrical Engineering, v. 21, n. 2, p. 217–234, 2024.
[6] PIMENTEL, A. Methodologies for design space exploration. In: CHATTOPADHYAY, A. (Ed.). Handbook of Computer Architecture. Singapore: Springer Nature, 2022. p. 1–31.
[7] AKRAM, A.; SAWALHA, L. A survey of computer architecture simulation techniques and tools. IEEE Access, v. 7, p. 78120–78145, 2019.
[8] SHAMIMUL HASAN, S. et al. Design space exploration of emerging memory technologies for machine learning applications. In: International Parallel and Distributed Processing Symposium Workshops. [S.l.]: IEEE, 2021. p. 439–448.
[9] WU, N.; XIE, Y. A survey of machine learning for computer architecture and systems. ACM Computing Surveys, v. 55, n. 3, 2022.
[10] INJADAT, M. et al. Machine learning towards intelligent systems: applications, challenges, and opportunities. Artificial Intelligence Review, v. 54, p. 3299–3348, 2021.
[11] NTI, I. et al. A mini-review of machine learning in big data analytics: Applications, challenges, and prospects. Big Data Mining and Analytics, v. 5, n. 2, p. 81–97, 2022.
[12] DOPPA, J.; ROSCA, J.; BOGDAN, P. Autonomous design space exploration of computing systems for sustainability: Opportunities and challenges. IEEE Design & Test, v. 36, n. 5, p. 35–43, 2019.
[13] AJANI, T.; IMOIZE, L.; ATAYERO, A. An overview of machine learning within embedded and mobile devices–optimizations and applications. Sensors, v. 21, n. 13, 2021.
[14] KIM, R.; DOPPA, J.; PANDE, P. Machine learning for design space exploration and optimization of manycore systems. In: International Conference on Computer-Aided Design (ICCAD). [S.l.]: IEEE Press, 2018. p. 1–6.
[15] ELAKKUMANAN, P. et al. Chiddam: a data mining based technique for cache hierarchy determination in commercial applications. In: 48th Midwest Symposium on Circuits and Systems. [S.l.: s.n.], 2005. v. 2, p. 1888–1891.
[16] ENGIN, I. et al. Efficiently exploring architectural design spaces via predictive modeling. ACM SIGOPS Operating Systems Review, v. 40, n. 5, 2006.
[17] KHAKHAENG, S.; CHANTRAPORNCHAI, C. On the finding proper cache prediction model using neural network. In: 28th International Conference on Knowledge and Smart Technology. [S.l.]: IEEE, 2016. p. 146–151.
[18] SEN, S.; IMAM, N. Machine learning based design space exploration for hybrid main-memory design. In: Proceedings of the International Symposium on Memory Systems. [S.l.]: ACM, 2019. p. 480–489.
[19] SINGH, G. et al. Napel: Near-memory computing application performance prediction via ensemble learning. In: 56th Annual Design Automation Conference. [S.l.]: ACM, 2019. p. 1–6.
[20] VAZQUEZ, R.; GORDON-ROSS, A.; STITT, G. Machine learning-based prediction for dynamic architectural optimizations. In: 10th International Green and Sustainable Computing Conference. [S.l.]: IEEE, 2019. p. 1–6.
[21] NAVARRO, O. et al. A machine learning methodology for cache memory design based on dynamic instructions. ACM Transactions on Embedded Computing Systems, v. 19, n. 2, 2020.
[22] THASNEEMA, V.; BHASKARAN, S. A classification model for predicting suitable cache level in a multi-core architecture. In: International Conference on Communication, Control and Information Sciences. [S.l.]: IEEE, 2021. v. 1, p. 1–6.
[23] TRIPATHI, A.; RAJAWAT, A. Early area and power estimation model for rapid system level design and design space exploration. Advances in Electrical and Electronic Engineering, v. 20, n. 1, p. 66–72, 2022.
[24] KECHENG, J. et al. An artificial neural network model of LRU-cache misses on out-of-order embedded processors. Microprocessors and Microsystems, v. 50, p. 66–79, 2017.
[25] LOWE-POWER, J.; AL, M. A. et al. The gem5 simulator: Version 20.0+. CoRR, abs/2007.03152, 2020. Disponível em: ⟨https://arxiv.org/abs/2007.03152⟩.
[26] LUK, C. et al. Pin: building customized program analysis tools with dynamic instrumentation. In: ACM SIGPLAN Conference on Programming Language Design and Implementation. [S.l.]: ACM, 2005. p. 190–200.
[27] WEKA tool. Weka-3.8.6. Disponível em: ⟨https://sourceforge.net/projects/weka/files/weka-3-8/3.8.6/⟩.
[28] NAI, L. et al. Graphbig: understanding graph computing in the context of industrial solutions. In: International Conference for High Performance Computing, Networking, Storage and Analysis. [S.l.]: ACM, 2015.
[29] CHE, S. et al. Rodinia: A benchmark suite for heterogeneous computing. In: International Symposium on Workload Characterization (IISWC). [S.l.]: IEEE, 2009. p. 44–54.
[30] ABELLA-GONZÁLEZ, M. et al. Polybench/python: benchmarking python environments with polyhedral optimizations. In: 30th ACM SIGPLAN International Conference on Compiler Construction. [S.l.]: ACM, 2021. p. 59–70.
[31] THOMAS, S. et al. Cortexsuite: A synthetic brain benchmark suite. In: International Symposium on Workload Characterization. [S.l.]: IEEE, 2014. p. 76–79.
[32] GUTHAUS, M. et al. Mibench: A free, commercially representative embedded benchmark suite. In: Proceedings of the Fourth Annual IEEE International Workshop on Workload Characterization. WWC-4 (Cat. No.01EX538). [S.l.]: IEEE, 2001. p. 3–14.
[33] PIN tool. Pin 3.31 - A dynamic binary instrumentation. Disponível em: ⟨http://Software.Intel.Com/En-Us/Articles/Pintool-Downloads⟩.
[34] BINKERT, N. et al. The gem5 simulator. SIGARCH Computer Architecture News, v. 39, n. 2, p. 1–7, 2011.
[35] HALL, M. et al. The Weka data mining software: an update. ACM SIGKDD Explorations Newsletter, v. 11, n. 1, p. 10–18, 2009.
[36] CUTLER, A.; CUTLER, D.; STEVENS, J. Random forests. In: ZHANG, C.; M. Y. (Ed.). Ensemble Machine Learning. New York: Springer, 2012. p. 157–175.
[37] JADAMA, A.; TORAY, M. Ensemble learning: Methods, techniques, application. 2024.
Downloads
Published
How to Cite
Issue
Section
License
Copyright (c) 2025 Hetal V. Dave, Dr. Nirali A. Kotak, Dr. Jay B. Teraiya
This work is licensed under a Creative Commons Attribution-NonCommercial 4.0 International License.
Autorizo aos editores a publicação de meu artigo, caso seja aceito, em meio eletrônico de acordo com as regras do Public Knowledge Project.
