Decoupled Access-Execute enabled DVFS for tinyML deployments on STM32 microcontrollers

TL;DR

This paper presents a novel approach for deploying CNNs on STM32 microcontrollers that optimizes power consumption through decoupled access-execute kernels and DVFS, achieving significant energy savings for edge ML inference.

Contribution

It introduces a new methodology combining decoupled access-execute kernels with DVFS for power-efficient CNN deployment on STM32 MCUs, outperforming existing solutions.

Findings

Up to 25.2% reduction in energy consumption compared to state-of-the-art methods.

Effective clocking and configuration exploration enhances power efficiency.

DVFS optimization under latency constraints improves energy-performance trade-offs.

Abstract

Over the last years the rapid growth Machine Learning (ML) inference applications deployed on the Edge is rapidly increasing. Recent Internet of Things (IoT) devices and microcontrollers (MCUs), become more and more mainstream in everyday activities. In this work we focus on the family of STM32 MCUs. We propose a novel methodology for CNN deployment on the STM32 family, focusing on power optimization through effective clocking exploration and configuration and decoupled access-execute convolution kernel execution. Our approach is enhanced with optimization of the power consumption through Dynamic Voltage and Frequency Scaling (DVFS) under various latency constraints, composing an NP-complete optimization problem. We compare our approach against the state-of-the-art TinyEngine inference engine, as well as TinyEngine coupled with power-saving modes of the STM32 MCUs, indicating that we can achieve up to 25.2% less energy consumption for varying QoS levels.