Green Algorithms:
Quantifying the Carbon Footprint of Computation
Source: Lannelongue—2021—Advanced Science—Wiley Online Library
This is a scientific publication published in the magazine Advanced Science (Advanced Science, Volume 8, Number 12, 23 June 2021) which tackles the subject of calculating the carbon footprint of AI projects.
The publication explains how researchers Loïc Lannelongue, Jason Grealey and Michael Inouye have developed a method for estimating the carbon footprint of an AI project. To obtain this estimate, the researchers took into account a number of parameters such as the sources of energy consumption (CPU, memory), the time allocated to data processing, etc.
The researchers also provided a website (available by clicking here) with a graphical interface. As well as displaying the cost in CO2 of carrying out the AI project, the Green Algorithms site shows the energy that will be consumed in kWh, the number of trees needed to absorb the CO2 emitted, the number of km driven by car and an estimate (as a percentage) of the cost of a one-way flight between Paris and London.
Figure 1 - website of Greens Algorithms.
Source: http://calculator.green-algorithms.org