Authors:
Juan J. Merelo-Guervós
1
;
Mario García-Valdez
2
and
Pedro Castillo
1
Affiliations:
1
Department of Computer Engineering, Automatics and Robotics, University of Granada, Granada, Spain
;
2
Department of Graduate Studies, National Technological Institute of Mexico, Tijuana, Mexico
Keyword(s):
Green Computing, Metaheuristics, JavaScript, Energy-Aware Computing, Evolutionary Algorithms.
Abstract:
What is known as energy-aware computing includes taking into account many different variables and parameters when designing an application, which makes it necessary to focus on a single one to obtain meaningful results. In this paper, we will look at the energy consumption of three different JavaScript interpreters: bun, node and deno; given their different conceptual designs, we should expect different energy budgets for running the (roughly) same workload, operations related to evolutionary algorithms (EA), a population-based stochastic optimization algorithm. In this paper we will first test different tools to measure per-process energy consumption in a precise way, trying to find the one that gives the most accurate estimation; after choosing the tool by performing different experiments on a workload similar to the one carried out by EA, we will focus on EA-specific functions and operators and measure how much energy they consume for different problem sizes. From this, we will tr
y to draw a conclusion on which JavaScript interpreter should be used in this kind of workloads if energy (or related expenses) has a limited budget.
(More)