Avaliação do Consumo de Energia para o Treinamento de Aprendizado de Máquina utilizando Single-board computers baseadas em ARM

• Felipe Bernardo LNCC
• André Yokoyama LNCC
• Bruno Schulze LNCC
• Mariza Ferro LNCC

DOI: https://doi.org/10.5753/wscad.2021.18512

Abstract

In this work, the use of ARM-based single-board computers is evaluated for training Machine Learning (ML) algorithms. For this, an experimental setup was developed, training the algorithm XGBoost with 36 hyperparameter configurations in four different architectures. Furthermore, its efficiency (energy consumption, acquisition cost and execution time) was compared with the main architectures used for the training of ML (x86 and GPU). The results show that this type of architecture can become a viable and greener alternative, not only for inference but also for the training phase of these algorithms.

References

Chen, T., He, T., Benesty, M., Khotilovich, V., Tang, Y., Cho, H., et al. (2015). Xgboost: extreme gradient boosting. R package version 0.4-2, 1(4):1–4.

Ferro, M., Silva, G., Klôh, V., Yokoyama, A. M., Mury, A. R., and Schulze, B. (2017). Challenges in HPC evaluation: Towards a methodology for scientific application requirements. In Grandinetti, L., Mirtaheri, S. L., Shahbazian, R., Sterling, T. L., and Voevodin, V. V., editors, Big Data and HPC: Ecosystem and Convergence, TopHPC 2017, Tehran, Iran, 24-26 April 2017, volume 33 of Advances in Parallel Computing, pages 32–52. IOS Press.

Guo, R., Zhao, Z., Wang, T., Liu, G., Zhao, J., and Gao, D. (2020). Degradation state recognition of piston pump based on iceemdan and xgboost. Applied Sciences, 10(18).

Henderson, P., Hu, J., Romoff, J., Brunskill, E., Jurafsky, D., and Pineau, J. (2020). Towards the systematic reporting of the energy and carbon footprints of machine learning. Journal of Machine Learning Research, 21(248):1–43.

Holt, J. and Sievert, S. (2021). Training machine learning models faster with dask. SciPy Conferences.

Kaewkasi, C. and Srisuruk, W. (2014). A study of big data processing constraints on a low-power hadoop cluster. In 2014 International Computer Science and Engineering Conference (ICSEC), pages 267–272. IEEE.

Kaggle (2020). State of data science and machine learning 2020. Technical report, https://www.kaggle.com/kaggle-survey-2020.

Kanagachidambaresan, G. R., Prakash, K. B., and Mahima, V. (2021). Programming Tensor Flow with Single Board Computers, pages 145–157. Springer International Publishing, Cham.

Khaydarova, R., Fishchenko, V., Mouromtsev, D., Shmatkov, V., and Lapaev, M. (2020). Rock-cnn: a distributed rockpro64-based convolutional In 2020 26th neural network cluster for iot. verification and performance analysis. Conference of Open Innovations Association (FRUCT), pages 174–181.

Kim, J., Galanopoulos, A., Joseph, J. V., and Kwak, J. (2020). A study In 2020 Int. on energy-process-latency tradeoff in embedded artificial intelligence. Conf. on Inf. and Communication Tec. Convergence (ICTC), pages 22–24. IEEE.

Miranda, M. M. d. (2012). Fator de emissão de gases de efeito estufa da geração de energia elétrica no Brasil: implicações da aplicação da Avaliação do Ciclo de Vida. PhD thesis, Universidade de São Paulo.

Mittal, S. (2019). A survey on optimized implementation of deep learning models on the nvidia jetson platform. Journal of Systems Architecture, 97:428–442.

Partel, V., Kakarla, S. C., and Ampatzidis, Y. (2019). Development and evaluation of a low-cost and smart technology for precision weed management utilizing artificial intelligence. Comp. and electronics in agriculture, 157:339–350.

Sapio, A., Canini, M., Ho, C.-Y., Nelson, J., Kalnis, P., Kim, C., Krishnamurthy, A., Moshref, M., Ports, D. R., and Richtárik, P. (2019). Scaling distributed ml with in-network aggregation. arXiv preprint arXiv:1903.06701.

Serpa, M. S., Krause, A. M., Cruz, E. H., Navaux, P. O. A., Pasin, M., and Felber, P. (2018). Optimizing machine learning algorithms on multi-core and many-core architectures using thread and data mapping. In 2018 26th Euromicro Int. Conf. on Parallel, Dist. and Network-based Processing (PDP), pages 329–333. IEEE.

Silva, G., Schulze, B., and Ferro, M. (2021). Performance and energy efficiency analysis of machine learning algorithms towards green ai: a case study of decision tree algorithms. Master’s thesis, National Lab. for Scientific Computing.

Strubell, E., Ganesh, A., and McCallum, A. (2019). Energy and In Proceedings of the 57th Annual policy considerations for deep learning in nlp. Meeting of the Association for Computational Linguistics, pages 3645–3650.

Süzen, A. A., Duman, B., and Sen, B. (2020). Benchmark analysis of jetson tx2, jetson nano and raspberry pi using deep-cnn. In 2020 Int. Cong. on Human-Computer Interaction, Optimization and Robotic Applications, pages 1–5. IEEE.

Tynes, I. L. (2021). Pain recognition performance on a single board computer. Master’s thesis, University of South Florida.

UNESCO, D.-G. (2021). Preliminary report on the first draft of the recommendation on the ethics of artificial intelligence. https://unesdoc.unesco.org/ark:/48223/pf0000374266.locale=en.

Vinuesa, R., Azizpour, H., Leite, I., Balaam, M., Dignum, V., Domisch, S., Felländer, A., Langhans, S., Tegmark, M., and Nerini, F. F. (2020). The role of artificial intelligence in achieving the sustainable development goals. Nature Communications, 11(233).