Data-Driven Power Modeling and Monitoring via Hardware Performance Counter Tracking

TL;DR

This paper presents a novel, accurate, and low-overhead power modeling method using hardware performance counters, integrated into the Linux kernel for real-time power monitoring and management in embedded systems.

Contribution

The paper introduces a new PMC-based power modeling approach that does not rely on microarchitectural details and achieves high accuracy with minimal overhead.

Findings

Average power estimation error of 7.5%

Energy estimation error of 1.3%

Runtime power measurement with 0.7% overhead

Abstract

Energy-centric design is paramount in the current embedded computing era: use cases require increasingly high performance at an affordable power budget, often under real-time constraints. Hardware heterogeneity and parallelism help address the efficiency challenge, but greatly complicate online power consumption assessments, which are essential for dynamic hardware and software stack adaptations. We introduce a novel power modeling methodology with state-of-the-art accuracy, low overhead, and high responsiveness, whose implementation does not rely on microarchitectural details. Our methodology identifies the Performance Monitoring Counters (PMCs) with the highest linear correlation to the power consumption of each hardware sub-system, for each Dynamic Voltage and Frequency Scaling (DVFS) state. The individual, simple models are composed into a complete model that effectively describes the power consumption of the whole system, achieving high accuracy and low overhead. Our evaluation reports an average estimation error of 7.5% for power consumption and 1.3% for energy. We integrate these models in the Linux kernel with Runmeter, an open-source, PMC-based monitoring framework. Runmeter manages PMC sampling and processing, enabling the execution of our power models at runtime. With a worst-case time overhead of only 0.7%, Runmeter provides responsive and accurate power measurements directly in the kernel. This information can be employed for actuation policies in workload-aware DVFS and power-aware, closed-loop task scheduling.