It is critical to consider and address the environmental implications and challenges of AI. Why?
The environmental impact of training Meena, a chatbot developed by Google, which is equivalent to 242,231 miles driven by an average passenger vehicle.[1]
The announcement of the TPU V3.0 by the Alphabet CEO highlighted the need for liquid cooling in data centers to accommodate the high power consumption of these chips.[2]
Projections indicate that data centers will account for 5% of global energy demand in 2030.[3]
It is imperative to address these challenges to ensure that AI is developed and deployed in an environmentally sustainable manner.
Project Overview
DESC: Type I: Towards Greener AI Computing: Designing and Managing Sustainable Heterogeneous Edge Data Centers
(09/2023 - 08/2026)
Funded by: NSF
This project is designed to significantly enhance the sustainability of edge data centers specialized in AI computing. Utilizing a comprehensive, carbon-first vertical approach, the project will address multiple levels of optimization. We will develop robust hardware accelerators that are energy-efficient, and also implement system-level improvements aimed at better power usage and strategic workload distribution. Additionally, renewable energy sources will be integrated into the management of these edge data centers. The ultimate goal is not only to reduce the overall carbon footprint through operational and embodied efficiencies but also to shift towards a more localized, renewable energy-based infrastructure. This holistic approach aims to pave the way for a new standard in energy-efficient, sustainable data center operations.
Project Outcomes
Our research will lead to the development of best practices and tools for creating sustainable EDCs that can be adopted by industry and academia.
Journals
- Balancing Throughput and Fair Execution of Multi-DNN Workloads on Heterogeneous Embedded Devices IEEE Transactions on Emerging Topics in Computing (IEEE TETC), 2024 [PDF]
- Pythia: An Edge First Agent for State Prediction in High-Dimensional Environments IEEE Embedded Systems Letters (IEEE ESL), 2024 [PDF]
- Hardware-Aware DNN Compression via Diverse Pruning and Mixed-Precision Quantization IEEE Transactions on Emerging Topics in Computing (IEEE TETC), 2023 [PDF]
Conferences
- RankMap: Priority-Aware Multi-DNN Manager for Heterogeneous Embedded Devices ACM/IEEE Conference on Design, Automation and Test in Europe (DATE), 2025 [PDF]
- Less is More: Optimizing Function Calling for LLM Execution on Edge Devices ACM/IEEE Conference on Design, Automation and Test in Europe (DATE), 2025 [PDF]
- MapFormer: Attention-based multi-DNN Manager for Throughout & Power Co-optimization on Embedded Devices ACM/IEEE International Conference on Computer-Aided Design (ICCAD), 2024 [PDF]
- Carbon-Aware Design of DNN Accelerators: Bridging Performance and Sustainability EEE Computer Society Annual Symposium on VLSI (ISVLSI), 2024 [PDF]
Contact Us
For more information about this project, please reach out to our research team at iraklis.anagno@siu.edu.