Dynamic GPU Energy Optimization for Machine Learning Training Workloads

TL;DR

This paper introduces GPOEO, an online framework that optimizes GPU energy consumption during machine learning training by dynamically balancing energy use and training time, leading to significant energy savings.

Contribution

GPOEO is a novel online energy optimization framework that employs multi-objective prediction and search techniques to reduce GPU energy consumption during training workloads.

Findings

Achieves 16.2% average energy savings

Increases training time by only 5.1% on average

Effective across diverse machine learning workloads

Abstract

GPUs are widely used to accelerate the training of machine learning workloads. As modern machine learning models become increasingly larger, they require a longer time to train, leading to higher GPU energy consumption. This paper presents GPOEO, an online GPU energy optimization framework for machine learning training workloads. GPOEO dynamically determines the optimal energy configuration by employing novel techniques for online measurement, multi-objective prediction modeling, and search optimization. To characterize the target workload behavior, GPOEO utilizes GPU performance counters. To reduce the performance counter profiling overhead, it uses an analytical model to detect the training iteration change and only collects performance counter data when an iteration shift is detected. GPOEO employs multi-objective models based on gradient boosting and a local search algorithm to find a trade-off between execution time and energy consumption. We evaluate the GPOEO by applying it to 71 machine learning workloads from two AI benchmark suites running on an NVIDIA RTX3080Ti GPU. Compared with the NVIDIA default scheduling strategy, GPOEO delivers a mean energy saving of 16.2% with a modest average execution time increase of 5.1%.