Identification of Energy Hotspots: A Case Study of the Very Fast Decision Tree

Abstract

Large-scale data centers account for a significant share of the energy consumption in many countries. Machine learning technology requires intensive workloads and thus drives requirements for lots of power and cooling capacity in data centers. It is time to explore green machine learning. The aim of this paper is to profile a machine learning algorithm with respect to its energy consumption and to determine the causes behind this consumption. The first scalable machine learning algorithm able to handle large volumes of streaming data is the Very Fast Decision Tree (VFDT), which outputs competitive results in comparison to algorithms that analyze data from static datasets. Our objectives are to: (i) establish a methodology that profiles the energy consumption of decision trees at the function level, (ii) apply this methodology in an experiment to obtain the energy consumption of the VFDT, (iii) conduct a fine-grained analysis of the functions that consume most of the energy, providing an understanding of that consumption, (iv) analyze how different parameter settings can significantly reduce the energy consumption. The results show that by addressing the most energy intensive part of the VFDT, the energy consumption can be reduced up to a 74.3%.

Fine-grained energy profiling for deep convolutional neural networks on the Jetson TX1

“…Despite this massive scale of efforts towards developing energyefficient solutions to deep learning problems, there are surprisingly very few studies that measure energy for deep learning workloads [4, 22,28,45,52]. We consolidate our observations from these works and attribute the lack of adoption of energy-use as an evaluation criteria to the following reasons:…”

Hoeffding Trees with Nmin Adaptation

“…In another publication, the authors [3] already confirmed the high energy impact of the functions involved in calculating the best attributes. This matches with our energy model, and motivates the reasons and objectives for nmin adaptation:…”

“…concept drift adaptation, the parameters of such algorithms are fixed from the beginning of the execution. We have observed that having fixed parameters leads to the algorithm making unnecessary computations, thus increasing its energy consumption [3].…”

Em Busca de uma Inteligência Artificial Ecologicamente Viável: Um estudo de caso do Consumo Energético de Algoritmos de íArvore de Decisão

“…Ou ainda, trabalhos que usam algoritmos de AM para predizer o desempenho e o consumo de energia sobre a execução de uma aplicação científica [Siegmund et al 2015, Wu et al 2016, Ferreira et al 2017, Klôh et al 2019. Porém, ainda existem poucos trabalhos que avaliam o consumo de energia dos algoritmos de AM [Li et al 2016, Garcia-Martin et al 2017, Yang et al 2017, Abdelhafez et al 2019, García-Martín et al 2019.…”

“…Em [Garcia-Martin et al 2017] são avaliados os hotspots de energia do algoritmo Very Fast Decision Tree (VFDT). A implementação, o treinamento e teste do algoritmo são realizados no framework Massive Online Analysis (MOA) [Bifet et al 2010] e a energiaé medida com a ferramenta Jalen [Noureddine et al 2015].…”