Note: Hal-04662683v1.
Keywords: Green, data center, bare metal server, cloud computing.
@InProceedings{ InProceedingsAgusti.ACFLNO_24,
address = {Copenhagen, Denmark.},
author = {Agusti, Maxime and Caron, Eddy and Fichel, Benjamin and Lef{\`e}vre, Laurent and Nicol, Olivier and Orgerie, Anne-C{\'e}cile},
booktitle = {{GreenCom 2024 - 20th IEEE International conference on Green Computing and Communications}},
editor = {IEEE},
keywords = {Green, data center, bare metal server, cloud computing},
month = {August 19-22},
note = {hal-04662683v1},
pdf = {https://hal.science/hal-04662683/file/Greencom_2024_PowerHeat.pdf},
title = {{PowerHeat: A Non-Intrusive Approach for Estimating the Power Consumption}},
url = {https://inria.hal.science/hal-04662683v1},
year = {2024}
}
Note: Hal-04346102.
Keywords: Green.
| Abstract: | As digital services are increasingly being deployed and used in a variety of domains, the environmental impact of Information and Communication Technologies (ICTs) is a matter of concern. Artificial intelligence is driving some of this growth but its environmental cost remains scarcely studied. A recent trend in large-scale generative models such as ChatGPT has especially drawn attention since their training requires intensive use of a massive number of specialized computing resources. The inference of those models is made accessible on the web as services, and using them additionally mobilizes end-user terminals, networks, and data centers. Therefore, those services contribute to global warming, worsen metal scarcity, and increase energy consumption. This work proposes an LCA-based methodology for a multi-criteria evaluation of the environmental impact of generative AI services, considering embodied and usage costs of all the resources required for training models, inferring from them, and hosting them online. We illustrate our methodology with Stable Diffusion as a service, an open-source text-to-image generative deep-learning model accessible online. This use case is based on an experimental observation of Stable Diffusion training and inference energy consumption. Through a sensitivity analysis, various scenarios estimating the influence of usage intensity on the impact sources are explored. |
@InProceedings{ InProceedingsBerthelot.BCJL_24,
abstract = {As digital services are increasingly being deployed and used in a variety of domains, the environmental impact of Information and Communication Technologies (ICTs) is a matter of concern. Artificial intelligence is driving some of this growth but its environmental cost remains scarcely studied. A recent trend in large-scale generative models such as ChatGPT has especially drawn attention since their training requires intensive use of a massive number of specialized computing resources. The inference of those models is made accessible on the web as services, and using them additionally mobilizes end-user terminals, networks, and data centers. Therefore, those services contribute to global warming, worsen metal scarcity, and increase energy consumption. This work proposes an LCA-based methodology for a multi-criteria evaluation of the environmental impact of generative AI services, considering embodied and usage costs of all the resources required for training models, inferring from them, and hosting them online. We illustrate our methodology with Stable Diffusion as a service, an open-source text-to-image generative deep-learning model accessible online. This use case is based on an experimental observation of Stable Diffusion training and inference energy consumption. Through a sensitivity analysis, various scenarios estimating the influence of usage intensity on the impact sources are explored. },
address = {Turin. Italia},
author = {Berthelot, Adrien and Caron, Eddy and Jay, Mathilde and Lef{\`e}vre, Laurent},
institution = {Inria and Octo Technology},
keywords = {Green},
month = {19-21 June},
note = {hal-04346102},
organization = {CIRP LCE 2024},
pages = {707-712},
pdf = {https://www.sciencedirect.com/science/article/pii/S2212827124001173/pdf?md5=8b9c2eca3ef60ba2bebac981d47e9225&pid=1-s2.0-S2212827124001173-main.pdf},
publisher = {Procedia CIRP},
series = {Special issue 31st CIRP Conference on Life Cycle Engineering.},
title = {Estimating the environmental impact of {G}enerative-{AI} services using an {LCA}-based methodology},
url = {https://www.sciencedirect.com/science/article/pii/S2212827124001173},
volume = {122},
year = {2024}
}
Note: Hal-04797304.
Keywords: Cloud Computing, Green.
@InProceedings{ InProceedingsCaron.CGLL_24,
address = {Khalifa University, Abu Dhabi, UAE},
author = {Caron, Eddy and Grivel, R{\'e}mi and Lambert, Simon and Lef{\`e}vre, Laurent},
booktitle = {IEEE CloudCom 2024 (International Conference on Cloud Computing Technology and Science)},
keywords = {Cloud Computing, Green},
month = {December 9-11},
note = {hal-04797304},
pdf = {https://inria.hal.science/hal-04797304v1/preview/CloudCom2024SL-FINAL.pdf},
publisher = {IEEE},
title = {{S-ORCA: a social-based consolidation approach to reduce Cloud infrastructures energy consumption}},
url = {https://inria.hal.science/hal-04797304},
year = {2024}
}