Supercomputing
Advancing sustainability in high-performance computing
December 5, 2025
Share
Server infrastructure at the Centre for Advanced Computing enables data-intensive and computational research.
Supercomputing helps scientists design vaccines, predict extreme weather, and train the artificial intelligence tools that power everyday life. But this progress presents new challenges. The data centres that support these systems use vast amounts of electricity and water to stay cool. As global demand for computing grows, researchers are racing to make these systems more efficient.
At Queen’s, that work is happening inside the . PhD student Shaina Smith (Electrical and Computer Engineering) and master’s student Jordan Abt (Electrical and Computer Engineering) are part of a team led by (Electrical and Computer Engineering) that studies how to keep computing progress aligned with environmental responsibility.
Smith examines how supercomputing systems can limit waste through improved cooling, and energy reuse. Abt focuses on computational efficiency, looking at how information moves through processors to save both time and power. Together, they approach the challenge from different angles but share the same goal of reducing the environmental cost of digital innovation.
Their collaboration earned international recognition when they co-authored research that won Best Paper at the Energy Efficiency and Sustainable Practices Workshop at the 2025 International Supercomputing Conference (ISC) in Hamburg, Germany.
In the CAESAR Lab, graduate students Shaina Smith (left) and Jordan Abt (right) collaborate on research that reduces the environmental impact of high-performance computing.
Measuring sustainability
Their paper What a Waste tackles one of computing’s biggest inefficiencies, the enormous amount of heat released by data centres. In most facilities, this heat is treated as waste and vented through large cooling systems, but this is now changing with more facilities interested in finding new ways to capture and reuse this heat.
Benchmarks such as Power Usage Effectiveness have long been used to measure the efficiency of data centres. However, they focus only on how effectively a system uses electricity and do not account for how much waste heat is recovered or how those efforts reduce emissions.
To address this gap, the team introduced two new metrics, the Heat Reuse Factor and the Green Supercomputing Ratio. The first measures how much of a data centre’s total heat output is captured and put to use, while the second combines that information with carbon data to estimate how much greenhouse gas is avoided through heat reuse.
“These metrics give us a clearer picture of what sustainability looks like in practice,” says Smith. “They help operators see how their choices, like using renewable energy or investing in heat recovery, translate into real environmental benefits.”
The researchers tested their approach through three case studies. They explored using waste heat for district heating in Ottawa, greenhouse heating in Alaska, and desalination in California. Each example showed how geography and climate affect performance. For instance, district heating offered strong results in colder months, while desalination provided consistent, year-round benefits.
“Sustainability in computing isn’t one-size-fits-all,” says Abt. “The best strategy in one region might not work in another, so understanding those variables is key to making meaningful progress.”
Preparing for the next generation of computing
The CAESAR Lab currently has 22 graduate students gaining hands-on experience with high-performance computing while embedding environmental responsibility in its research approach. Smith and Abt’s achievements also highlight the growing role of women in a field where they remain underrepresented.
“Canada is raising a new generation of computing talent that understands the balance between performance and sustainability,” says Dr. Grant. “Seeing students like Shaina and Jordan take the lead shows how innovation and bright young researchers can move this field forward.”
Work in areas such as sustainable computing systems and high-performance computing is supported by a strong network of researchers and technical expertise across campus, including resources like the , a leading research data centre and analytics hub. This environment helps position Queen’s to contribute to the next generation of high-performance and energy-efficient computing.
“Canada’s supercomputing capacity is still growing,” says Smith. “By thinking about how data centres can support local energy systems, we can position Canada as a leader in green supercomputing.”
