Collaborative Research: EAGER: CET: GreenSW: Fostering Sustainable HPC and Cloud Software Systems through AI-Enabled Energy-aware Code Smell Refactoring

This EArly-concept Grants for Exploratory Research (EAGER) award is made in response to Dear Colleague Letter 23-109, as part of the NSF-wide Clean Energy Technology initiative. Software applications and workloads, especially within the domains of High-Performance Computing (HPC), Cloud computing, and large-scale Artificial Intelligence (AI) model training, exert considerable demand on computing resources, thus contributing significantly to the overall carbon footprint. Currently, the carbon emissions attributed to the software industry rival those of the aviation sector, and this trend is projected to escalate further by 2030. This project targets the creation of sustainable and environmentally friendly software by identifying and selectively restructuring energy-draining code segments (known as code smells) and addressing coding inefficiencies. This innovative approach has the potential to yield substantial savings in energy consumption, reduce carbon emissions, and make a significant contribution to the environment. The project encompasses a comprehensive education and outreach program, featuring science projects for K-12 students, the development of new undergraduate and graduate-level courses, mentoring of minority and underrepresented students, and the dissemination of the project outcomes to the wider society. This project facilitates sustainable and low-carbon software development through three significant contributions: (1) a comprehensive analysis of code smells and investigation of the impact of their refactoring on application energy consumption and carbon footprint; (2) development of novel machine-learning models tailored to intelligently and judiciously guide code smell refactoring while prioritizing energy efficiency; and (3) application of the developed models across a diverse spectrum of HPC, Cloud and AI workloads, enabling robust validation and comprehensive evaluation. The research outcomes of this project will revolutionize energy optimization in software systems by offering a holistic framework that addresses the complex challenges of code smell refactoring and energy consumption. By integrating sophisticated machine learning models and a rigorous validation process, the project will pave the way for sustainable and low-carbon software development practices, benefiting both the software industry and the environment on a broader scale. This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the Foundation's intellectual merit and broader impacts review criteria.