PowerFlow-DNN: Compiler-Directed Fine-Grained Power Orchestration for End-to-End Edge AI Inference

2026-03-24

AI Summary

Edge AI systems face significant challenges in efficiently applying fine-grained power management due to stringent energy and form-factor constraints, and inter-layer power decisions often exhibit strong coupling that leads to combinatorial explosion. This work proposes PowerFlow-DNN, the first compiler-guided framework that formulates cross-layer power scheduling for end-to-end DNN inference as a unified optimization problem, explicitly accounting for voltage rail constraints and state-switching overheads. By leveraging a structured state graph, integer linear programming modeling, and a lightweight pruning algorithm, the approach efficiently navigates a search space exceeding $10^{160}$ configurations on a TSMC 40nm accelerator. It achieves a near-optimal solution with only 0.68% higher energy consumption than the theoretical optimum, reduces energy by 37% compared to an unscheduled baseline, and accelerates the solving process by 2.14×.

Technology Category

• Machine Learning: Hardware-aware ML
• Constraint Satisfaction and Optimization: Distributed CSP/Optimization
• Planning, Routing, and Scheduling: Optimization of Spatio-temporal Systems

Application Category

• Systems and Infrastructure for Web, Mobile and WoT: Energy management for devices in mobile Web and WoT environments
• Graph Algorithms and Modeling for the Web: Graph neural networks and deep learning approaches for Web-related graphs
• Economics, Online Markets and Human Computation: Cost models of using LLMs in production systems

Abstract

Edge AI systems often operate under stringent energy and volume constraints that demand extreme efficiency under limited battery capacity, with requirements worsening as intelligent capability demands advance. Prior literature suggests that fine-grained power orchestration, including DVFS and power gating, enables significant energy efficiency benefits that cannot be left unexploited, while still exhibiting unexplored challenges. We observe that layer-level approaches incur unintended overheads due to inter-layer coupling of power control decisions, and that jointly managing these mechanisms under practical constraints such as limited voltage rails and transition overheads leads to a rapidly growing combinatorial schedule space. To address this, we propose PowerFlow-DNN, a compiler-directed framework for end-to-end power-state orchestration in ultra-low-power accelerators. By constructing a rigorous problem formulation for deadline-constrained, real-time, periodic inference as a unified inter-layer power-scheduling problem, our framework enables automated discovery of energy-minimal power-state schedules that adhere to a deadline while accounting for end-to-end, inter-layer impacts. We evaluate the framework on a DNN accelerator VLSI implementation in TSMC 40nm technology. Across representative edge networks, we show that PowerFlow-DNN discovers near-optimal solutions under the discretized formulation and achieves energy within 0.68\% of the exact ILP oracle, reducing energy by up to 37\% compared to an aggressive baseline without power orchestration, while reasoning over a combinatorial schedule space of over $10^{160}$ possible power-state assignments, yet operating on a structured layered state graph that enables efficient optimization, achieving up to 2.14$\times$ solver speedup via lightweight pruning.

Problem

• fine-grained power orchestration
• edge AI inference
• energy efficiency
• inter-layer coupling
• combinatorial schedule space

Innovation

• Power orchestration
• Compiler-directed optimization
• Edge AI inference
• Fine-grained DVFS
• Inter-layer scheduling