Energy Scaling Laws for Diffusion Models: Quantifying Compute in Image Generation

Fuente: arXiv
Guardado en:
Detalles Bibliográficos
Autores principales: Iyengar, Aniketh, Han, Jiaqi, Ruf, Boris, Grari, Vincent, Detyniecki, Marcin, Ermon, Stefano
Formato: Preprint
Publicado: 2025
Materias:
Acceso en línea:
Etiquetas: Agregar Etiqueta
Sin Etiquetas, Sea el primero en etiquetar este registro!
_version_ 1866911676475899904
author Iyengar, Aniketh
Han, Jiaqi
Ruf, Boris
Grari, Vincent
Detyniecki, Marcin
Ermon, Stefano
author_facet Iyengar, Aniketh
Han, Jiaqi
Ruf, Boris
Grari, Vincent
Detyniecki, Marcin
Ermon, Stefano
contents The rapidly growing computational demands of diffusion models for image generation have raised significant concerns about energy consumption and environmental impact. While existing approaches to energy optimization focus on architectural improvements or hardware acceleration, there is a lack of principled methods to predict energy consumption across different model configurations and hardware setups. We propose an adaptation of Kaplan scaling laws to predict GPU energy consumption for diffusion models based on computational complexity (FLOPs). Our approach decomposes diffusion model inference into text encoding, iterative denoising, and decoding components, with the hypothesis that denoising operations dominate energy consumption due to their repeated execution across multiple inference steps. We conduct comprehensive experiments across four state-of-the-art diffusion models (Stable Diffusion 2, Stable Diffusion 3.5, Flux, and Qwen) on three GPU architectures (NVIDIA A100, A4000, A6000), spanning various inference configurations including resolution ($256^2$--$1024^2$), precision (fp16/fp32), step counts (10--50), and classifier-free guidance settings. Our energy scaling law achieves high predictive accuracy within individual architectures ($R^2 > 0.9$) and exhibits strong cross-architecture generalization, maintaining high rank correlations across models and enabling reliable energy estimation for unseen model--hardware combinations. These results validate the compute-bound nature of diffusion inference and establish energy consumption estimation as a necessary foundation for sustainable AI deployment planning and subsequent carbon footprint assessment.
format Preprint
id arxiv_https___arxiv_org_abs_2511_17031
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle Energy Scaling Laws for Diffusion Models: Quantifying Compute in Image Generation
Iyengar, Aniketh
Han, Jiaqi
Ruf, Boris
Grari, Vincent
Detyniecki, Marcin
Ermon, Stefano
Machine Learning
Computer Vision and Pattern Recognition
Computers and Society
The rapidly growing computational demands of diffusion models for image generation have raised significant concerns about energy consumption and environmental impact. While existing approaches to energy optimization focus on architectural improvements or hardware acceleration, there is a lack of principled methods to predict energy consumption across different model configurations and hardware setups. We propose an adaptation of Kaplan scaling laws to predict GPU energy consumption for diffusion models based on computational complexity (FLOPs). Our approach decomposes diffusion model inference into text encoding, iterative denoising, and decoding components, with the hypothesis that denoising operations dominate energy consumption due to their repeated execution across multiple inference steps. We conduct comprehensive experiments across four state-of-the-art diffusion models (Stable Diffusion 2, Stable Diffusion 3.5, Flux, and Qwen) on three GPU architectures (NVIDIA A100, A4000, A6000), spanning various inference configurations including resolution ($256^2$--$1024^2$), precision (fp16/fp32), step counts (10--50), and classifier-free guidance settings. Our energy scaling law achieves high predictive accuracy within individual architectures ($R^2 > 0.9$) and exhibits strong cross-architecture generalization, maintaining high rank correlations across models and enabling reliable energy estimation for unseen model--hardware combinations. These results validate the compute-bound nature of diffusion inference and establish energy consumption estimation as a necessary foundation for sustainable AI deployment planning and subsequent carbon footprint assessment.
title Energy Scaling Laws for Diffusion Models: Quantifying Compute in Image Generation
topic Machine Learning
Computer Vision and Pattern Recognition
Computers and Society
url https://arxiv.org/abs/2511.17031