English

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

Hardware Architecture 2025-02-04 v1

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.

Keywords

Cite

@article{arxiv.2502.00286,
  title  = {Late Breaking Results: Leveraging Approximate Computing for Carbon-Aware DNN Accelerators},
  author = {Aikaterini Maria Panteleaki and Konstantinos Balaskas and Georgios Zervakis and Hussam Amrouch and Iraklis Anagnostopoulos},
  journal= {arXiv preprint arXiv:2502.00286},
  year   = {2025}
}

Comments

Accepted in DATE 2025 - Late Breaking Results

R2 v1 2026-06-28T21:28:45.146Z