Data dependent energy modelling for worst case energy consumption analysis

TL;DR

This paper introduces a probabilistic energy modeling approach for embedded software, enabling more accurate worst-case energy consumption analysis by considering instruction operand distributions.

Contribution

It proposes a novel method to compose instruction sequences using probabilistic distributions for improved WCEC analysis, addressing limitations of existing models.

Findings

The energy of embedded benchmarks can be characterized as a distribution.

The proposed probabilistic model predicts worst-case energy with statistical analysis.

Comparison shows improved accuracy over traditional estimation methods.

Abstract

Safely meeting Worst Case Energy Consumption (WCEC) criteria requires accurate energy modeling of software. We investigate the impact of instruction operand values upon energy consumption in cacheless embedded processors. Existing instruction-level energy models typically use measurements from random input data, providing estimates unsuitable for safe WCEC analysis. We examine probabilistic energy distributions of instructions and propose a model for composing instruction sequences using distributions, enabling WCEC analysis on program basic blocks. The worst case is predicted with statistical analysis. Further, we verify that the energy of embedded benchmarks can be characterised as a distribution, and compare our proposed technique with other methods of estimating energy consumption.