EUVIMEDEUVIMED entdecken
Crossref

LLMCO 2 : Advancing Accurate Carbon Footprint Prediction for LLM Inferences

Zhenxiao Fu, Fan Chen, Shan Zhou, Haitong Li, Lei Jiang

ACM SIGEnergy Energy Informatics Review · 2025-07 · 5 (2) · 63-68

DOI: 10.1145/3757892.3757901 ↗

Kurz gefasst

Fachartikel in der Zeitschrift ACM SIGEnergy Energy Informatics Review, online erschienen am 06.08.2025. Verfasst von Zhenxiao Fu und 4 weiteren Autor:innen. Die Arbeit wurde bisher 14-mal zitiert (Crossref) und nennt 25 Literaturangaben.

AbstractBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

Throughout its lifecycle, an LLM incurs significantly higher carbon emissions during inference than training. Inference requests vary in batch size, prompt length, and token generation, while cloud providers deploy heterogeneous GPU configurations to meet diverse service-level objectives. Unlike training, inference exhibits lower and highly variable hardware utilization, making equation-based carbon models unreliable. Existing network-based estimators lack accuracy, as they fail to account for the distinct prefill and decode phases, hardware-specific features, and realistic request distributions. We propose LLMCO 2 , a graph neural network (GNN)-based model, to improve the accuracy of LLM inference carbon footprint estimation by ~ 67% over prior approaches. Source code is available at https://github.com/fuzhenxiao/LLMCO 2 .

Zitieren

APA

Fu, Z., Chen, F., Zhou, S., Li, H., & Jiang, L. (2025). LLMCO 2 : Advancing Accurate Carbon Footprint Prediction for LLM Inferences. ACM SIGEnergy Energy Informatics Review, 5(2), 63-68. https://doi.org/10.1145/3757892.3757901

Vancouver

Fu Z, Chen F, Zhou S, Li H, Jiang L. LLMCO 2 : Advancing Accurate Carbon Footprint Prediction for LLM Inferences. ACM SIGEnergy Energy Informatics Review. 2025;5(2):63-68. doi:10.1145/3757892.3757901

BibTeX
@article{Fu2025,
  author = {Fu, Zhenxiao and Chen, Fan and Zhou, Shan and Li, Haitong and Jiang, Lei},
  title = {LLMCO 2 : Advancing Accurate Carbon Footprint Prediction for LLM Inferences},
  journal = {ACM SIGEnergy Energy Informatics Review},
  year = {2025},
  volume = {5},
  number = {2},
  pages = {63--68},
  doi = {10.1145/3757892.3757901}
}
Kostenlos starten →