Equipment: CC Campus Compute: Expansion of GPU Compute Capacity for NC State University HPC to Support Research and Education

设备：CC Campus Compute：扩展北卡罗来纳州立大学 HPC 的 GPU 计算能力以支持研究和教育

• 批准号: 2321565
• 负责人: Andrew Petersen
• 金额: $ 46.76万
• 依托单位: North Carolina State University
• 依托单位国家: 美国
• 项目类别: Standard Grant
• 财政年份: 2023
• 资助国家: 美国
• 起止时间: 2023-08-01 至 2025-07-31
• 项目状态: 未结题
• 来源: https://www.nsf.gov/awardsearch/showAward?AWD_ID=2321565&HistoricalAwards=false
• 关键词: Equipment; CC; Campus; Compute; Expansion

This project expands advanced computing capacity at North Carolina State University (NC State). The resources enable larger models, faster simulations, larger datasets for more accurate Artificial Intelligence models, and faster processing of genome assemblies. The resource is deployed into a supportive environment - the High Performance Computing (HPC) facility - which provides outreach, training, documentation, tutorials, support, and tools for access, visualization, job execution, and an onboarding process for beginning users.The project installs six Lenovo SR670 nodes with the latest NVIDA Hopper architecture (with 12x H100 and 12x L40 NVIDIA GPUs) so that teaching, classroom use, and research (including research from the science drivers) can expand, relieving the compute congestion on the current GPU cards. More than twelve science applications will be supported on the system. Applications include ab-initio methods such as Density Functional Theory and Molecular Dynamics, neural network training, cryo-EM, genome assembly, and fluid simulation. More than 200 projects are supported by the HPC facility. Access to this system will be provided for all students, including undergraduates. Educational impacts include access to the system for teaching, classroom use, and research capability for HPC users. A significant portion of the resource will be made available to Open Science Grid, providing increased computing capacity for scientific users nationwide.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.

该项目扩展了北卡罗来纳州州立大学（NC State）的高级计算能力。这些资源可以实现更大的模型，更快的模拟，更大的数据集，以获得更准确的人工智能模型，以及更快的基因组组装处理。该资源被部署到一个支持性环境-高性能计算（HPC）设施-提供外展，培训，文档，教程，支持和工具，用于访问，可视化，作业执行，该项目安装了六个联想SR 670节点，采用最新的NVIDA Hopper架构（配备12个H100和12个L40 NVIDIA GPU），以便教学、课堂使用和研究（包括科学驱动程序的研究）可以扩展，从而缓解当前GPU卡上的计算拥塞。该系统将支持12个以上的科学应用。应用包括从头算方法，如密度泛函理论和分子动力学，神经网络训练，冷冻EM，基因组组装和流体模拟。HPC设施支持200多个项目。所有学生，包括本科生，都可以使用这个系统。教育影响包括访问系统用于教学，课堂使用和HPC用户的研究能力。很大一部分资源将用于开放科学网格，为全国的科学用户提供更高的计算能力。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。