REOH: Using Probabilistic Network for Runtime Energy Optimization of Heterogeneous Systems

TL;DR

REOH is a probabilistic network-based method that predicts energy-efficient configurations for heterogeneous systems, significantly reducing sampling effort while maintaining near-optimal energy consumption, and is implemented in an open-source runtime framework.

Contribution

This paper introduces REOH, a novel holistic tuning approach using probabilistic networks for energy optimization in heterogeneous systems, extending prior homogeneous-focused methods.

Findings

REOH achieves near-optimal energy consumption close to brute-force methods.

REOH reduces sampling runs by 17% compared to previous homogeneous approaches.

The approach is validated on systems with CPUs and GPUs, demonstrating effectiveness.

Abstract

Significant efforts have been devoted to choosing the best configuration of a computing system to run an application energy efficiently. However, available tuning approaches mainly focus on homogeneous systems and are inextensible for heterogeneous systems which include several components (e.g., CPUs, GPUs) with different architectures. This study proposes a holistic tuning approach called REOH using probabilistic network to predict the most energy-efficient configuration (i.e., which platform and its setting) of a heterogeneous system for running a given application. Based on the computation and communication patterns from Berkeley dwarfs, we conduct experiments to devise the training set including 7074 data samples covering varying application patterns and characteristics. Validating the REOH approach on heterogeneous systems including CPUs and GPUs shows that the energy consumption by the REOH approach is close to the optimal energy consumption by the Brute Force approach while saving 17% of sampling runs compared to the previous (homogeneous) approach using probabilistic network. Based on the REOH approach, we develop an open-source energy-optimizing runtime framework for selecting an energy efficient configuration of a heterogeneous system for a given application at runtime.