MRI: Acquisition of an HPC System for Data-Driven Discovery in Computational Astrophysics, Biology, Chemistry, and Materials Science

MRI：获取 HPC 系统，用于计算天体物理学、生物学、化学和材料科学中的数据驱动发现

• 批准号: 1828187
• 负责人: Srinivas Aluru
• 金额: $ 369.93万
• 依托单位: Georgia Tech Research Corporation
• 依托单位国家: 美国
• 项目类别: Standard Grant
• 财政年份: 2018
• 资助国家: 美国
• 起止时间: 2018-10-01 至 2022-09-30
• 项目状态: 已结题
• 来源: https://www.nsf.gov/awardsearch/showAward?AWD_ID=1828187&HistoricalAwards=false
• 关键词: MRI; Acquisition; HPC; System; Data

The project funds the purchase of a high-performance computing and storage system at the Georgia Institute of Technology. This computing instrument will support data-driven research in astrophysics, biosciences, computational chemistry, materials and manufacturing, and computational science. These projects contribute to national initiatives in big data, strategic computing, materials genome, and manufacturing partnership; and NSF supported observatories such as the gravitational wave observatory and the South Pole neutrino observatory. The system also serves as a springboard for developments of codes, software prototyping, and scalability studies prior to using national supercomputers. Advances made in computational methods and scientific software are disseminated in the form of open-source codes and data analysis portals. Over 33 faculty, 54 research scientists/postdocs, 195 graduate students, and 56 undergraduate students will immediately benefit from the instrument. In addition, the system provides training opportunity at all levels from undergraduate students to early career researchers, in important interdisciplinary areas of national need. A fifth of the system capacity is utilized to enable research activities of regional partners, researchers from minority serving institutions, and other users nationally through XSEDE participation. The project involves undergraduate student participation from historically black colleges from Atlanta metropolitan area. Public outreach efforts are planned through videos of public interest and local events such as the Atlanta Science Festival.The cluster will combine regular compute nodes with others configured to emphasize one of the following: big memory, big local storage, solid state storage, Graphics Processing Units (GPU), and ARM processors. In doing so, the system can be employed by a diversity of projects. In astrophysics, the instrument bolsters data-driven research including detection of gravitational waves, astrophysical neutrinos, and gamma rays. It does it by leveraging data from leading astroparticle observatories and contributing to their mission. It also leads to improved insights into formation of supermassive black holes and large-scale structure of the universe. The computing system also aids the development of parallel software in computational genomics, systems biology, and health analytics. Important applications in assembly and network analysis of plant genomes, and environmental metagenomics are pursued. The instrument also enables next generation algorithms and software for computational chemistry and expands the boundaries of molecular simulation. The system enables advances in density function theory, enhances studies of crystal defects and nanostructures, and injects novel use of machine learning techniques in computational chemistry. It also fosters the development of data science methodologies to identify building blocks of materials at multiple scales, thus significantly reducing the development and deployments cycles for new materials.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.

该项目为格鲁吉亚理工学院购买高性能计算和存储系统提供资金。该计算仪器将支持天体物理学、生物科学、计算化学、材料和制造以及计算科学中的数据驱动研究。这些项目有助于大数据、战略计算、材料基因组和制造伙伴关系方面的国家倡议; NSF支持的观测站，如引力波观测站和南极中微子观测站。该系统也是在使用国家超级计算机之前开发代码、软件原型和可扩展性研究的跳板。在计算方法和科学软件方面取得的进展以开放源代码和数据分析门户的形式传播。超过33名教师，54名研究科学家/博士后，195名研究生和56名本科生将立即受益于该仪器。此外，该系统还在国家需要的重要跨学科领域提供从本科生到早期职业研究人员的各级培训机会。该系统五分之一的能力用于使区域伙伴、少数群体服务机构的研究人员和全国其他用户通过XSEDE的参与开展研究活动。该项目涉及来自亚特兰大大都会地区历史上黑人大学的本科生参与。计划通过公众感兴趣的视频和亚特兰大科学节等当地活动开展公众宣传工作。集群将联合收割机常规计算节点与其他配置的节点结合起来，以强调以下内容之一：大内存、大本地存储、固态存储、图形处理单元（GPU）和ARM处理器。在这样做时，该系统可用于各种项目。在天体物理学中，该仪器支持数据驱动的研究，包括引力波，天体物理中微子和伽马射线的探测。它通过利用来自领先的天体粒子观测站的数据并为其使命做出贡献来做到这一点。它还导致了对超大质量黑洞形成和宇宙大尺度结构的更好理解。该计算系统还有助于在计算基因组学、系统生物学和健康分析中开发并行软件。在植物基因组的组装和网络分析，以及环境宏基因组学的重要应用。该仪器还为计算化学提供了下一代算法和软件，并扩展了分子模拟的边界。该系统使密度函数理论的进步，增强了晶体缺陷和纳米结构的研究，并在计算化学中注入了机器学习技术的新用途。该奖项还促进了数据科学方法的发展，以在多个尺度上识别材料的构建块，从而大大缩短了新材料的开发和部署周期。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。

项目成果

Cartesian message passing neural networks for directional properties: Fast and transferable atomic multipoles

• DOI: 10.1063/5.0050444
• 发表时间: 2021-06-14
• 期刊: JOURNAL OF CHEMICAL PHYSICS
• 影响因子: 4.4
• 作者: Glick, Zachary L.;Koutsoukas, Alexios;Sherrill, C. David
• 通讯作者: Sherrill, C. David

Halo Environment for Population III Star Formation

III族恒星形成的光环环境

• DOI: 10.3847/2515-5172/ab9e78
• 发表时间: 2020
• 期刊: Research Notes of the AAS
• 作者: Grace, Justin;O’Shea, Brian W.;Wise, John H.
• 通讯作者: Wise, John H.

Parallel construction of module networks

模块网络的并行构建

• DOI: 10.1145/3458817.3476207
• 发表时间: 2021
• 期刊: Storage and Analysis (SC
• 作者: Srivastava, Ankit;Chockalingam, Sriram P.;Aluru, Maneesha;Aluru, Srinivas
• 通讯作者: Aluru, Srinivas

Decoding defect statistics from diffractograms via machine learning

• DOI: 10.1038/s41524-021-00539-z
• 发表时间: 2021-05-17
• 期刊: NPJ COMPUTATIONAL MATERIALS
• 影响因子: 9.7
• 作者: Kunka, Cody;Shanker, Apaar;Dingreville, Remi
• 通讯作者: Dingreville, Remi

Seeding the second star – II. CEMP star formation enriched from faint supernovae

• DOI: 10.1093/mnras/staa2144
• 发表时间: 2020-07
• 期刊: Monthly Notices of the Royal Astronomical Society
• 影响因子: 4.8
• 作者: G. Chiaki;J. Wise;S. Marassi;R. Schneider;M. Limongi;A. Chieffi