CC* Compute: The MSU Data Machine - a high-memory, GPU-enabled compute cluster for data-intensive and AI applications

CC* 计算：MSU 数据机 - 一个高内存、支持 GPU 的计算集群，适用于数据密集型和人工智能应用程序

• 批准号: 2200792
• 负责人: Brian O'Shea
• 金额: $ 39.99万
• 依托单位: Michigan State University
• 依托单位国家: 美国
• 项目类别: Standard Grant
• 财政年份: 2022
• 资助国家: 美国
• 起止时间: 2022-07-01 至 2024-06-30
• 项目状态: 已结题
• 来源: https://www.nsf.gov/awardsearch/showAward?AWD_ID=2200792&HistoricalAwards=false
• 关键词: CC; Compute; MSU; Data; Machine

The MSU Data Machine addresses the exponential growth of large and complex datasets in many fields of study, particularly those where computing has not been widely used or where the research and teaching approaches needed to work with “big data”, which present a different set of computing requirements than in traditional high performance computing. The Data Machine facilitates data-intensive research by having computing nodes with large amounts of memory, a high speed file system, graphics processing units that are optimized for machine learning and artificial intelligence-based analysis techniques, and a high speed file system. It also includes software, usage policies, and training that makes it easy for users to interactively analyze and visualize their data. This project focuses on four specific research areas - in the areas of microbial genomics, social system modeling, spatial and community ecology, and data-driven turbulence modeling - however, the Data Machine broadly enables MSU’s research community to pursue data-intensive research projects by lowering barriers to engaging with these types of resources. The project also provides a valuable computational resource to the nation via the Open Science Grid and MSU’s NSF-funded Science DMZ project, advancing research in a wide spectrum of areas. Furthermore, MSU undergraduate and graduate students are participating in the deployment and administration of the Data Machine as well as using it for research and educational activities, contributing to the development of a globally competitive STEM workforce and promoting the advancement of under-represented groups in computational and data science.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.

密歇根州立大学数据机解决了许多研究领域中大型和复杂数据集呈指数增长的问题，特别是那些计算尚未广泛使用或研究和教学方法需要使用“大数据”的领域，这些领域提供了一组不同的数据与传统高性能计算相比，数据机通过拥有大量内存的计算节点、高速文件系统、针对机器学习和人工智能分析技术进行优化的图形处理单元来促进数据密集型研究，它还包括软件、用法。政策和培训，使用户可以轻松地交互式分析和可视化他们的数据。该项目重点关注四个特定的研究领域——微生物基因组学、社会系统建模、空间和群落生态学以及数据驱动的湍流建模。然而，数据机广泛地使密歇根州立大学的研究社区能够通过降低使用此类资源的障碍来开展数据密集型研究项目。该项目还通过开放科学网格和密歇根州立大学的 NSF 资助为国家提供了宝贵的计算资源。 Science DMZ 项目推动了广泛领域的研究，此外，密歇根州立大学的本科生和研究生正在参与数据机的部署和管理，并将其用于研究和教育活动，为发展具有全球竞争力的项目做出贡献。 STEM 劳动力并促进计算和数据科学领域代表性不足群体的进步。该奖项反映了 NSF 的法定使命，并通过使用基金会的智力价值和更广泛的影响审查标准进行评估，被认为值得支持。