CarbonCall: Sustainability-Aware Function Calling for Large Language Models on Edge Devices

Fuente: arXiv
Guardado en:
Detalles Bibliográficos
Autores principales: Paramanayakam, Varatheepan, Karatzas, Andreas, Anagnostopoulos, Iraklis, Stamoulis, Dimitrios
Formato: Preprint
Publicado: 2025
Materias:
Acceso en línea:
Etiquetas: Agregar Etiqueta
Sin Etiquetas, Sea el primero en etiquetar este registro!
_version_ 1866913816840765440
author Paramanayakam, Varatheepan
Karatzas, Andreas
Anagnostopoulos, Iraklis
Stamoulis, Dimitrios
author_facet Paramanayakam, Varatheepan
Karatzas, Andreas
Anagnostopoulos, Iraklis
Stamoulis, Dimitrios
contents Large Language Models (LLMs) enable real-time function calling in edge AI systems but introduce significant computational overhead, leading to high power consumption and carbon emissions. Existing methods optimize for performance while neglecting sustainability, making them inefficient for energy-constrained environments. We introduce CarbonCall, a sustainability-aware function-calling framework that integrates dynamic tool selection, carbon-aware execution, and quantized LLM adaptation. CarbonCall adjusts power thresholds based on real-time carbon intensity forecasts and switches between model variants to sustain high tokens-per-second throughput under power constraints. Experiments on an NVIDIA Jetson AGX Orin show that CarbonCall reduces carbon emissions by up to 52%, power consumption by 30%, and execution time by 30%, while maintaining high efficiency.
format Preprint
id arxiv_https___arxiv_org_abs_2504_20348
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle CarbonCall: Sustainability-Aware Function Calling for Large Language Models on Edge Devices
Paramanayakam, Varatheepan
Karatzas, Andreas
Anagnostopoulos, Iraklis
Stamoulis, Dimitrios
Performance
Artificial Intelligence
Systems and Control
Large Language Models (LLMs) enable real-time function calling in edge AI systems but introduce significant computational overhead, leading to high power consumption and carbon emissions. Existing methods optimize for performance while neglecting sustainability, making them inefficient for energy-constrained environments. We introduce CarbonCall, a sustainability-aware function-calling framework that integrates dynamic tool selection, carbon-aware execution, and quantized LLM adaptation. CarbonCall adjusts power thresholds based on real-time carbon intensity forecasts and switches between model variants to sustain high tokens-per-second throughput under power constraints. Experiments on an NVIDIA Jetson AGX Orin show that CarbonCall reduces carbon emissions by up to 52%, power consumption by 30%, and execution time by 30%, while maintaining high efficiency.
title CarbonCall: Sustainability-Aware Function Calling for Large Language Models on Edge Devices
topic Performance
Artificial Intelligence
Systems and Control
url https://arxiv.org/abs/2504.20348