MRI-R2: Acquisition of a Hybrid CPU/GPU and Visualization Cluster for Multidisciplinary Studies in Transport Physics with Uncertainty Quantification

MRI-R2：获取混合 CPU/GPU 和可视化集群，用于不确定性量化的运输物理学多学科研究

• 批准号: 0960306
• 负责人: Eric Stefan Shaqfeh
• 金额: $ 402.96万
• 依托单位: Stanford University
• 依托单位国家: 美国
• 项目类别: Standard Grant
• 财政年份: 2010
• 资助国家: 美国
• 起止时间: 2010-04-01 至 2013-09-30
• 项目状态: 已结题
• 来源: https://www.nsf.gov/awardsearch/showAward?AWD_ID=0960306&HistoricalAwards=false
• 关键词: MRI; R2; Acquisition; Hybrid; CPU

0960306Shaqfeh"This award is funded under the American Recovery and Reinvestment Act of 2009 (Public Law 111-5)."This proposal is for the acquisition of a high performance, Hybrid CPU/GPU and Visualization Cluster (HPVC) that is designed to serve the needs of 7 working groups where each is involved in HPC research to solve problems of large global and national impact. The unique aspects of the HPVC cluster lies, in the combination of state-of-the art computer visualization, enhanced performance with large datasets via GPU architecture, and the development of computational uncertainty quantification (UQ). These elements, , will be employed in the solution of interdisciplinary, multiphysics transport problems including (a) engineering Nutrient Transport in Oceanographic Coral Reef Flows (to protect our environment and save the reefs), (b) the Development of Trauma and Nanoparticle Cancer Therapies including understanding hemostasis in the microcirculation (impacting human health), (c) Modeling the Dispersion of Biological Agents in Urban Environments (for Homeland Security), (d) using Quantum Chemistry to Design Catalysts for Fuel Cells (toward sustainable energy development), (e) simulating the Molecular Dynamics of Proteins and the Ribosome (for understanding the basic processes of life), (f) developing a fundamental understanding of Jet Noise Creation (to develop effective noise mitigation strategies for jets, fans, etc.), and finally, (g) to create algorithms for Design Optimization for Sustainable Building Development (as a contribution toward the larger issue of sustainability). The HPVC will play a key role in education, via bringing CPU with GPU enhanced acceleration to the classroom in a spectrum of Stanford courses as well as new workshops on GPU computing. The concept of Uncertainty Quantification as applied to large scale HPC problems will be introduced via a new UQ pipeline as a novel software deliverable within the HPVC research program.

0960306 Shaqfeh“该奖项是根据2009年美国复苏和再投资法案（公法111-5）资助的。“该提案旨在收购高性能混合CPU/GPU和可视化集群（HPVC），旨在满足7个工作组的需求，每个工作组都参与HPC研究，以解决具有重大全球和国家影响的问题。HPVC集群的独特之处在于结合了最先进的计算机可视化、通过GPU架构增强大型数据集的性能以及计算不确定性量化（UQ）的发展。这些要素将用于解决跨学科、多物理场的输运问题，包括（a）海洋学珊瑚礁流中的营养物输运工程（保护我们的环境和拯救珊瑚礁），（B）创伤和纳米粒子癌症治疗的发展，包括理解微循环中的止血（c）模拟生物制剂在城市环境中的扩散（国土安全），（d）利用量子化学设计燃料电池催化剂（朝向可持续能源发展），（e）模拟蛋白质和核糖体的分子动力学（用于理解生命的基本过程），（f）发展对喷气噪音产生的基本理解（为喷气机，风扇等开发有效的噪音缓解策略），最后，（g）创建可持续建筑发展的设计优化算法（作为对可持续性更大问题的贡献）。HPVC将在教育中发挥关键作用，通过在斯坦福大学的一系列课程以及GPU计算的新研讨会中将具有GPU增强加速的CPU带入课堂。将通过新的UQ管道引入应用于大规模HPC问题的不确定性量化概念，作为HPVC研究计划中的一种新型软件交付。