Workshop on Clusters, Clouds and Data Analytics in Scientific Computing

科学计算中的集群、云和数据分析研讨会

• 批准号: 1606551
• 负责人: Jack Dongarra
• 金额: $ 2.41万
• 依托单位: University of Tennessee Knoxville
• 依托单位国家: 美国
• 项目类别: Standard Grant
• 财政年份: 2016
• 资助国家: 美国
• 起止时间: 2016-03-15 至 2017-02-28
• 项目状态: 已结题
• 来源: https://www.nsf.gov/awardsearch/showAward?AWD_ID=1606551&HistoricalAwards=false
• 关键词: Workshop; Clusters; Clouds; Data; Analytics

This workshop will be held in Lyon, France, and will bring together international experts from the US and France to discuss advancements in cyberinfrastructure (CI). The specific focus will be on commonalities that exist between cluster, cloud, and data analytics computing. CI has become of growing importance to nearly all NSF research, which makes increasing use of computational simulation and large-data analysis methods. The results will be published in an open report as well as in several journal papers. They will contribute to and leverage new ideas in cluster, cloud, and data analytics computing for the benefit of the research communityThe architectural similarities between the above issues, and the fact that high performance clusters typically make up the major compute nodes of computational grids and clouds, means that deployment, operational, and usage issues surrounding computational clouds form a superset of the issues that revolve around clusters. The workshop will discuss topics such as extreme scale performance and related software development issues, common protocols and APIs necessary to facilitate seamless movement, recombination and analysis of large data flows, creating dramatically easier programming paradigms for highly parallel systems, creating tools that can support application development in highly virtualized environments, and system management problems that arise in multicore-heterogeneous architectures and data intensive applications.

这次研讨会将在法国里昂举行，来自美国和法国的国际专家将齐聚一堂，讨论网络基础设施(CI)的进步。具体重点将放在集群、云和数据分析计算之间存在的共性上。CI对几乎所有的NSF研究都变得越来越重要，这些研究越来越多地使用计算模拟和大数据分析方法。研究结果将发表在一份公开报告以及几篇期刊论文上。他们将促进和利用群集、云和数据分析计算方面的新想法，以造福于研究界上述问题之间的体系结构相似性，以及高性能群集通常构成计算网格和云的主要计算节点的事实，意味着围绕计算云的部署、运营和使用问题构成了围绕群集的问题的超集。研讨会将讨论以下主题：极大规模的性能和相关的软件开发问题；促进大型数据流的无缝移动、重组和分析所需的通用协议和API；为高度并行的系统创建极其简单的编程范例；创建可支持高度虚拟化环境中的应用程序开发的工具；以及多核-异类体系结构和数据密集型应用程序中出现的系统管理问题。