CodeCarbon: for a cleaner technology
Originally published by: Find your digital self | FySelf Blog
Computing today represents 0.5% of energy consumption globally. This figure will grow by more than 2% in the coming years, representing a significant increase in global CO2 emissions. Therefore, it is vitally important to track carbon emissions that go into the atmosphere.
As a solution, CodeCarbon, an open source software package, has came to light. This makes it possible to track and reduce carbon dioxide emissions that computer technology generates.
What is CodeCarbon?
CodeCarbon is a Mila software, which is a Montreal-based world leader in Artificial Intelligence (AI) research. They worked on the code in conjunction with BCG GAMMA’s AI and Data Science team, Pennsylvania-based Haverford College and solution provider MLOps Comet.ml (machine learning operations).
The essential purpose of this software package is basically to call on the environmental impact of digital technology. Therefore, its essential mission is to help reduce the carbon footprint that AI has.
AI is a powerful technology and a force for good, but it is important to be conscious of its growing environmental impact. The CodeCarbon project aims to do just that, and hope it inspires the AI community to calculate, disclose, and reduce their carbon footprint.
Yoshua Bengio, founder of Mila Canadian AI expert and winner of the prestigious 2018 Turing award.
Overall CodeCarbon is a Python codebase-built code and can calculate CO2 emissions from computing resources to execute code. In this way, it enables developers to improve their work. It also suggests reducing emissions by selecting cloud infrastructures in regions that use energy sources with a lower carbon footprint.
How does this software package work?
The operation of CodeCarbon is very easy. They just need to add one more line of code. The code makes the calculations on the amount of carbon dioxide generated by the devices that execute the code.
Using public data sources as a basis, CodeCarbon can calculate the amount of CO2 emissions produced. To do this, it takes into account the electrical network to which the equipment is anchored. Later it shows the results compared with data. Fot instance, the kilometers driven or the hours of television seen in an average home. This way programmers can visualize how much their code contaminates.
If recent history is any indicator, the use of computing in general, and AI computing in particular, will continue to expand exponentially around the world. As this happens, CodeCarbon can help organizations ensure that their collective carbon footprint increases as little as possible.
Sylvain Duranton, Managing Director and Senior Partner of Boston Consulting Group (BCG) and Global Director of BCG GAMMA H2 Environmental approach
The Zero Waste approach to technology is becoming a trend. In this context, artificial intelligence is a way to solve environmental problems. The world can no longer live without technology.
The solution proposed by CodeCarbon is, without a doubt, a step forward in the development of a sustainable planet. When its results on the carbon trace of a certain algorithm are evident, it is possible to work on it to maintain the advantages of the technology, but to do so in a friendly way with our environment.
