Saved in:
Bibliographic Details
Main Authors: Wilkins, Grant, Keshav, Srinivasan, Mortier, Richard
Format: Preprint
Published: 2024
Subjects:
Online Access:https://arxiv.org/abs/2407.04014
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866909242742538240
author Wilkins, Grant
Keshav, Srinivasan
Mortier, Richard
author_facet Wilkins, Grant
Keshav, Srinivasan
Mortier, Richard
contents The rapid adoption of large language models (LLMs) has led to significant advances in natural language processing and text generation. However, the energy consumed through LLM model inference remains a major challenge for sustainable AI deployment. To address this problem, we model the workload-dependent energy consumption and runtime of LLM inference tasks on heterogeneous GPU-CPU systems. By conducting an extensive characterization study of several state-of-the-art LLMs and analyzing their energy and runtime behavior across different magnitudes of input prompts and output text, we develop accurate (R^2>0.96) energy and runtime models for each LLM. We employ these models to explore an offline, energy-optimal LLM workload scheduling framework. Through a case study, we demonstrate the advantages of energy and accuracy aware scheduling compared to existing best practices.
format Preprint
id arxiv_https___arxiv_org_abs_2407_04014
institution arXiv
publishDate 2024
record_format arxiv
spellingShingle Offline Energy-Optimal LLM Serving: Workload-Based Energy Models for LLM Inference on Heterogeneous Systems
Wilkins, Grant
Keshav, Srinivasan
Mortier, Richard
Distributed, Parallel, and Cluster Computing
The rapid adoption of large language models (LLMs) has led to significant advances in natural language processing and text generation. However, the energy consumed through LLM model inference remains a major challenge for sustainable AI deployment. To address this problem, we model the workload-dependent energy consumption and runtime of LLM inference tasks on heterogeneous GPU-CPU systems. By conducting an extensive characterization study of several state-of-the-art LLMs and analyzing their energy and runtime behavior across different magnitudes of input prompts and output text, we develop accurate (R^2>0.96) energy and runtime models for each LLM. We employ these models to explore an offline, energy-optimal LLM workload scheduling framework. Through a case study, we demonstrate the advantages of energy and accuracy aware scheduling compared to existing best practices.
title Offline Energy-Optimal LLM Serving: Workload-Based Energy Models for LLM Inference on Heterogeneous Systems
topic Distributed, Parallel, and Cluster Computing
url https://arxiv.org/abs/2407.04014