Equipment: MRI: Track 2 Acquisition of a High-Performance Computing Cluster for Boosting Artificial Intelligence Enabled Science, Engineering, and Education in South Carolina

该重大研究仪器（MRI）项目旨在在南卡罗来纳州大学（USC）建立一种新的高性能计算（HPC）仪器，具有平衡的中央处理单元（CPU）和图形处理单元（GPU）节点。该设施将显着改善南加州大学现有的HPC资源，这将促进材料科学，物理，化学，工程，计算机科学，生物信息学，健康科学和人文等不同领域的研究，所有这些都将通过HPC，大数据和人工智能（AI）工具得到增强。此外，该工具将使南卡罗来纳州（SC）10多所本科院校的学生和教师能够获得HPC资源和人工智能培训，包括历史上的黑人和少数民族服务学院，为STEM中代表性不足的群体创造尖端HPC/AI科学研究的机会。通过南加州大学系统和南加州大学州的许多合作机构之间的合作，并为能源，气候和国家安全方面的国家研究重点做出贡献。新的HPC设施将使研究人员能够为AI模型训练进行大规模模拟数据采集。新仪器将用于最先进的分析、人工智能算法和数据可视化。研究工作集中在从头算量子物理和化学，分子动力学，计算流体动力学，多宇宙随机模型和数据密集型深度学习和图形卷积神经网络，将推进我们对网络物理系统和材料的了解。这些知识将有助于开发创新应用，例如材料的生成设计，复杂疾病的有效干预和预防机制。该项目将大大拓宽SC的HPC-AI计算设施的使用范围，并将通过技术研讨会，本科生研究奖学金和AI for Science项目以及高中夏令营提供全面的AI和HPC培训。该工具还将加强和促进区域联盟研究人员之间的跨学科合作，包括为历史上代表性不足的群体服务的私立本科院校（PUI）的科学家和学生，从而促进SC研究社区的有机增长。南加州大学和10个PUI的90名高中生和30个部门将立即受益于这种HPC设备。公众访问的项目网站将是http://ai4science.sc.edu。它将作为整个项目产生的所有研究产品的存放处，如数据，代码，结果，仿真器和模拟器。该项目网站将在项目期结束后至少再维护5年。该项目由重大研究仪器计划和刺激竞争性研究的既定计划（EPSCoR）共同资助。该奖项反映了NSF的法定使命，并通过使用基金会的知识价值和更广泛的影响审查标准进行评估，被认为值得支持。

This Major Research Instrumentation (MRI) project aims to establish a new High-Performance Computing (HPC) instrument with balanced Central Processing Unit (CPU) and Graphics Processing Units (GPU) nodes at the University of South Carolina (USC). This facility will significantly improve the existing HPC resources at USC, which will boost research in diverse fields such as materials science, physics, chemistry, engineering, computer science, bioinformatics, health science and humanities, all enhanced by the HPC, big data and artificial intelligence (AI) tools. Moreover, this instrument will enable access to HPC resources and AI training for students and faculty at over 10 undergraduate institutions in South Carolina (SC), including historically black and minority-serving colleges, creating opportunities for cutting-edge HPC/AI-for-Science research for underrepresented groups in STEM.This project will advance fundamental science, engineering, health sciences and humanities, through collaboration among many partnering institutions across the USC system and SC state, and contribute to national research priorities in energy, climate and national security. The new HPC facility will enable researchers to perform large-scale simulated data acquisition for AI model training. The new instrument will be employed for state-of-the-art analytics, AI algorithms and data visualization. Research efforts centered on ab initio quantum physics and chemistry, molecular dynamics, computational fluid dynamics, multi-universe stochastic models and data-intensive deep learning and graph convolutional neural networks, will advance our knowledge of cyber-physical systems and materials. This knowledge will contribute to the development of innovative applications such as generative design of materials, effective intervention and prevention mechanisms for complex diseases.This project will greatly broaden access to the HPC-AI computing facility in SC and will provide holistic AI and HPC training through technical workshops, undergraduate research scholarships and AI-for-Science projects and high-school summer camps. This instrument will also enhance and promote interdisciplinary collaborations among regional consortium researchers, including scientists and students at Primarily Undergraduate Institutions (PUIs) serving historically underrepresented groups, leading to an organic growth of the research community in SC. Over 3,000 people including 315 faculty, 2,550 graduate and undergraduate students, 90 high school students and 30 departments across USC and 10 PUIs will immediately benefit from this HPC equipment.The project website for public access will be http://ai4science.sc.edu. It will serve as the depository for all research products generated throughout the project, such as data, codes, results, emulators and simulators. The project website will be maintained for at least 5 additional years beyond the project period.This project is jointly funded by the Major Research Instrumentation program and the Established Program to Stimulate Competitive Research (EPSCoR).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.