Late Breaking Results: Leveraging Approximate Computing for Carbon-Aware DNN Accelerators

TL;DR

This paper presents a novel approach to designing sustainable DNN accelerators by applying approximate computing techniques to minimize environmental impact while maintaining performance.

Contribution

It introduces a method combining gate-level pruning, precision scaling, and genetic algorithms to optimize accelerator design for lower embodied carbon.

Findings

Reduced embodied carbon in DNN accelerators

Maintained performance and accuracy with approximate computing

Optimized accelerator design using genetic algorithms

Abstract

The rapid growth of Machine Learning (ML) has increased demand for DNN hardware accelerators, but their embodied carbon footprint poses significant environmental challenges. This paper leverages approximate computing to design sustainable accelerators by minimizing the Carbon Delay Product (CDP). Using gate-level pruning and precision scaling, we generate area-aware approximate multipliers and optimize the accelerator design with a genetic algorithm. Results demonstrate reduced embodied carbon while meeting performance and accuracy requirements.