AutoPower: Automated Few-Shot Architecture-Level Power Modeling by Power Group Decoupling

2025-08-17

AI Summary

Architecture-level power modeling faces challenges of low accuracy in analytical models and high dependency on labeled configurations in data-driven approaches. Method: This paper proposes an automated few-shot modeling paradigm. Its core innovation is the first introduction of a power-group decoupling mechanism, decomposing total power into independently modelable subsystems (e.g., clock, SRAM), each equipped with a lightweight hierarchical submodel—significantly reducing data requirements. The method integrates architectural feature analysis with few-shot machine learning, requiring only two known configurations for training. Results: Experiments on mainstream processors achieve a mean absolute error of 4.36% and an R² of 0.96—outperforming McPAT-Calib by 5% lower error and 0.09 higher R²—demonstrating both high accuracy and strong generalization across diverse microarchitectural configurations.

Technology Category

• Machine Learning: Hardware-aware ML
• Cognitive Modeling & Cognitive Systems: (Computational) Cognitive Architectures
• Computer Vision: Large Vision Models

Application Category

• User Modeling, Personalization and Recommendation: On-Device user modeling, personalization, and recommendation
• Graph Algorithms and Modeling for the Web: Graph neural networks and deep learning approaches for Web-related graphs
• Systems and Infrastructure for Web, Mobile and WoT: Applied ML and AI for Web-based mobile applications

Abstract

Power efficiency is a critical design objective in modern CPU design. Architects need a fast yet accurate architecture-level power evaluation tool to perform early-stage power estimation. However, traditional analytical architecture-level power models are inaccurate. The recently proposed machine learning (ML)-based architecture-level power model requires sufficient data from known configurations for training, making it unrealistic. In this work, we propose AutoPower targeting fully automated architecture-level power modeling with limited known design configurations. We have two key observations: (1) The clock and SRAM dominate the power consumption of the processor, and (2) The clock and SRAM power correlate with structural information available at the architecture level. Based on these two observations, we propose the power group decoupling in AutoPower. First, AutoPower decouples across power groups to build individual power models for each group. Second, AutoPower designs power models by further decoupling the model into multiple sub-models within each power group. In our experiments, AutoPower can achieve a low mean absolute percentage error (MAPE) of 4.36% and a high $R^2$ of 0.96 even with only two known configurations for training. This is 5% lower in MAPE and 0.09 higher in $R^2$ compared with McPAT-Calib, the representative ML-based power model.

Problem

• Automated power modeling with few known configurations
• Improving accuracy in architecture-level power estimation
• Decoupling power groups for precise individual modeling

Innovation

• Automated few-shot power modeling
• Power group decoupling technique
• Accurate low-configuration training model