CPINN-ABPI: Physics-Informed Neural Networks for Accurate Power Estimation in MPSoCs

2025-05-28

AI Summary

To address the severe inaccuracy of the ABPI method in power estimation on real-world MPSoCs—exhibiting errors of 47%–81%—this paper proposes CPINN-ABPI, a novel framework integrating physics-informed learning with hardware-aware constraints. First, it empirically validates ABPI’s limitations on commercial hardware. Second, it synergistically combines a customized Physics-Informed Neural Network (CPINN) with ABPI’s thermal-dynamic model, incorporating hardware-aware physical constraint losses to enforce thermodynamic consistency. Third, it employs a multi-objective genetic algorithm to jointly optimize estimation accuracy and inference overhead. Experimental results demonstrate that CPINN-ABPI reduces mean absolute error by 84.7% for CPU and 73.9% for GPU compared to ABPI; achieves a weighted mean absolute percentage error of ~12%; and incurs only 195.3 μs inference latency—meeting real-time deployment requirements on heterogeneous SoCs.

Technology Category

• Machine Learning: Hardware-aware ML
• Constraint Satisfaction and Optimization: Distributed CSP/Optimization
• Cognitive Modeling & Cognitive Systems: Conceptual Inference and Reasoning

Application Category

• Systems and Infrastructure for Web, Mobile and WoT: Applied ML and AI for Web-based mobile applications
• User Modeling, Personalization and Recommendation: On-Device user modeling, personalization, and recommendation
• Graph Algorithms and Modeling for the Web: Efficient manipulation of static and dynamic Web-related graphs

Abstract

Efficient thermal and power management in modern multiprocessor systems-on-chip (MPSoCs) demands accurate power consumption estimation. One of the state-of-the-art approaches, Alternative Blind Power Identification (ABPI), theoretically eliminates the dependence on steady-state temperatures, addressing a major shortcoming of previous approaches. However, ABPI performance has remained unverified in actual hardware implementations. In this study, we conduct the first empirical validation of ABPI on commercial hardware using the NVIDIA Jetson Xavier AGX platform. Our findings reveal that, while ABPI provides computational efficiency and independence from steady-state temperature, it exhibits considerable accuracy deficiencies in real-world scenarios. To overcome these limitations, we introduce a novel approach that integrates Custom Physics-Informed Neural Networks (CPINNs) with the underlying thermal model of ABPI. Our approach employs a specialized loss function that harmonizes physical principles with data-driven learning, complemented by multi-objective genetic algorithm optimization to balance estimation accuracy and computational cost. In experimental validation, CPINN-ABPI achieves a reduction of 84.7% CPU and 73.9% GPU in the mean absolute error (MAE) relative to ABPI, with the weighted mean absolute percentage error (WMAPE) improving from 47%--81% to $sim$12%. The method maintains real-time performance with 195.3~$mu$s of inference time, with similar 85%--99% accuracy gains across heterogeneous SoCs.

Problem

• Accurate power estimation in MPSoCs for thermal management
• Empirical validation of ABPI method on commercial hardware
• Improving ABPI accuracy with physics-informed neural networks

Innovation

• Integrates CPINNs with ABPI thermal model
• Uses specialized loss function combining physics
• Employs genetic algorithm for accuracy optimization