CarbonScaling: Extending Neural Scaling Laws for Carbon Footprint in Large Language Models

Fuente: arXiv
Salvato in:
Dettagli Bibliografici
Autori principali: Jiang, Lei, Chen, Fan
Natura: Preprint
Pubblicazione: 2025
Soggetti:
Accesso online:
Tags: Aggiungi Tag
Nessun Tag, puoi essere il primo ad aggiungerne!!
_version_ 1866911689548496896
author Jiang, Lei
Chen, Fan
author_facet Jiang, Lei
Chen, Fan
contents Large language models (LLMs) increasingly follow neural scaling laws that tie performance gains to rapidly expanding computational budgets, raising concerns about the sustainability of frontier-scale training. Existing carbon-estimation methods largely depend on regression over historical runs and fail to capture critical system-level factors, including hardware heterogeneity, distributed parallelism, communication overhead, and architectural sparsity. We present \textit{CarbonScaling}, a hardware-aware analytical framework for modeling the carbon scaling behavior of frontier LLM training. The framework integrates neural scaling laws, distributed training strategies, accelerator and interconnect modeling, and operational and embodied carbon accounting to estimate feasible hardware configurations and associated emissions. CarbonScaling jointly models tensor, pipeline, data, and expert parallelism while incorporating memory, bandwidth, utilization, and runtime constraints. Experimental validation demonstrates substantially higher fidelity than regression-based baselines and highlights the growing importance of embodied carbon at trillion-parameter scales. Source code: \url{https://github.com/UnchartedRLab/CarbonScaling}.
format Preprint
id arxiv_https___arxiv_org_abs_2508_06524
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle CarbonScaling: Extending Neural Scaling Laws for Carbon Footprint in Large Language Models
Jiang, Lei
Chen, Fan
Computation and Language
Artificial Intelligence
Computers and Society
Distributed, Parallel, and Cluster Computing
Machine Learning
Large language models (LLMs) increasingly follow neural scaling laws that tie performance gains to rapidly expanding computational budgets, raising concerns about the sustainability of frontier-scale training. Existing carbon-estimation methods largely depend on regression over historical runs and fail to capture critical system-level factors, including hardware heterogeneity, distributed parallelism, communication overhead, and architectural sparsity. We present \textit{CarbonScaling}, a hardware-aware analytical framework for modeling the carbon scaling behavior of frontier LLM training. The framework integrates neural scaling laws, distributed training strategies, accelerator and interconnect modeling, and operational and embodied carbon accounting to estimate feasible hardware configurations and associated emissions. CarbonScaling jointly models tensor, pipeline, data, and expert parallelism while incorporating memory, bandwidth, utilization, and runtime constraints. Experimental validation demonstrates substantially higher fidelity than regression-based baselines and highlights the growing importance of embodied carbon at trillion-parameter scales. Source code: \url{https://github.com/UnchartedRLab/CarbonScaling}.
title CarbonScaling: Extending Neural Scaling Laws for Carbon Footprint in Large Language Models
topic Computation and Language
Artificial Intelligence
Computers and Society
Distributed, Parallel, and Cluster Computing
Machine Learning
url https://arxiv.org/abs/2508.06524