Exploring Sustainable Computing
The work of 2024 PhD Joint Program participant Rajini Wijayawardan, graduate student in Computer Science working in the lab of Andrew Chien, addresses a question at the forefront of many minds, particularly amidst the rise of AI: how can the environmental impact of data centers be reduced?
In a recent paper supported in part by the CNRS-UChicago PhD Joint Program, Wijayawardana, Chien, and co-investigators Joachim Cendrier, Anne Benoit, Yves Robert, and Frédéric Vivien (CNRS/INRIA/ENS Lyon) examine how job scheduling and resource management techniques may reduce carbon emissions in data centers.
“Computing’s carbon footprint is a growing concern, and it’s not just happening in the cloud. The rapid growth in computing is also happening at the edge, where consumer AI services, intelligent city applications, and machine-learning models are increasingly deployed,” Wijayawardana shares. “This raises the question: when and where should we run our jobs, given that the carbon cost of power varies by location and time of day?”
In a promising set of findings, the authors show that carbon-aware schedulers can robustly reduce carbon emissions for workload, across a variety of load and communication carbon costs. According to Wijayawardana, the carbon-aware scheduling algorithms designed by the team “shows they reduce carbon costs by an average of 42% compared to standard approaches.”
Wijayawardana’s work is part of a project awarded through IRC Discovery’s PhD Joint Program, which provides mobility funding for PhD students working on a project co-directed by PIs at UChicago and the CNRS. “Efficient and environment-friendly scheduling and resource management algorithms” is co-led by Andrew Chien (UChicago) and Anne Benoit (CNRS/INRIA/ENS Lyon) and includes Wijayawardana’s PhD counterpart, Joachim Cendrier (ENS Lyon).
Paper: Joachim Cendrier, Rajini Wijayawardana, Anne Benoit, Yves Robert, Frédéric Vivien, and Andrew Chien. Green Scheduling on the Edge. Euro-Par 2025- 31st International European Conference on Parallel and Distributed Computing, Aug 2025, Dresden, Germany. pp.380-394.
This article was originally written by Salajaji and published on IRC Discovery News
