Our Vision

The REPACSS project envisions reducing the energy costs and improving the efficiency of large-scale scientific computing by developing innovative methods and tools. Over five years, the project aims to demonstrate the use of variable energy in powering advanced computing systems through a scalable high-performance computing data center and AI infrastructure.

By deploying a modern computing facility and addressing R&D challenges, REPACSS will explore variable energy sources for large-scale computing. The project will study barriers, tradeoffs, and the ability to predict and schedule scientific workloads based on varied energy production, enabling significant contributions to the research community.

Learn more about the project on the NSF award page.

Our Mission

• Deploy a modern computing facility to tackle R&D challenges in varaible energy market.
• Develop tools for remote data center control, automation, and scientific workflow management.
• Study barriers and tradeoffs in varaible energy and predict scientific workloads based on energy production.
• Support small- to mid-scale jobs across various scientific and engineering domains, including AI, machine learning, and deep learning.
• Analyze user workflows and behavior in response to power availability and cost versus quality-of-service tradeoffs.
• Promote adoption of variable energy sources and innovations by other facilities and the data center industry.
• Achieve significant reductions in energy costs and improve the efficiency for large-scale computing.