Category I: AMA27: Sustainable Cyber-infrastructure for Expanding Participation

The State University of New York (Stony Brook University, in collaboration with the University at Buffalo) will procure and operate for at least five years a high-performance, highly energy-efficient, large-memory computer system responsive to the interests and priorities of multiple science and engineering (S&E) and artificial Intelligence (AI)-focused research communities and research applications in national cyberinfrastructure. The project deploys a system utilizing the low-cost and low-energy AmpereOne ARM processors that excel at AI inference, as well as imperfectly-vectorized or pure scalar workloads that characterize much academic research computing but are inefficiently (cost, energy, and performance) supported by current NSF cyberinfrastructure. The large memory and excellent AI inference and generative performance of this processor and Qualcomm AI-200 accelerators will directly advance the mission of the National Artificial Intelligence Research Resource to broaden access to AI by being easily and tightly coupled into diverse applications, yet scalable to tens of thousands of simultaneous jobs or users nationwide. The project employs a comprehensive, multilayered strategy, with regional and national elements to ensure the widest possible benefit of AMA27. The team collaborates with multiple initiatives and NSF projects, reaching an audience that spans high-school students to faculty. Nationally and regionally, AMA27 will support a variety of projects and will provide tailored courses and training (both in-person and remote), host students and faculty, and encourage early registration access and flexible and novel access models. 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.