US-Africa High Performance Computing Workshop: TACC Ranger Lives On

该研讨会将通过帮助非洲科学家提高当地的超级计算能力，增强美国科学家与非洲同事在广泛的数据密集型项目上合作的潜力。研讨会将由德克萨斯大学和奥斯汀大学德克萨斯高级计算中心(TACC)高级计算系统主任汤米·明亚德组织和管理，并将建立在2013年向三个非洲国家捐赠近2000个计算节点的基础上，这些节点来自退役的？Ranger？超级计算机，它本身就是由美国国家科学基金会资助的。2013年向博茨瓦纳、坦桑尼亚和南非的机构捐赠的Ranger计算组件是由德克萨斯大学独立资助的。这次研讨会将有助于确保接受这种计算基础设施的非洲科学家拥有必要的技能和信息，以利用这一资源推动天文学、农业、气候学、地球科学、交通、电信和更多对美国和非洲感兴趣和有价值的研究领域的发现。TACC将组织一系列培训和信息交流活动，以支持非洲对超级计算的更广泛了解和使用，并为美非合作研究项目奠定基础。2014年11月，大约8名非洲研究人员将在出席在新奥尔良举行的超级计算会议后访问TACC。奥斯汀的访问将允许展示一系列广泛的应用程序和工具，这些应用程序和工具将增强非洲科学家？了解并有能力推动在南非举行的第二阶段研讨会。此外，访问奥斯汀将允许非洲科学家与几名使用超级计算的美国研究人员会面并讨论潜在的合作，以研究南部非洲或与南部非洲相关的主题。2014年12月，南非高性能计算中心将主办第二次规模更大的讲习班活动。2014年12月的活动将在CHPC主办的一年一度的南非超级计算会议之后举行，本次会议的与会者将被邀请参加随后的研讨会。TACC的六名工作人员将前往南非，就将服务器配置为单一集群以及为一系列研究活动安装和使用适当的开源应用程序进行为期三天的密集讨论和培训。TACC工作人员还将与他们的非洲同事讨论正在出现的多国非洲高性能计算(HPC)发展计划。加强非洲的高性能计算能力可能会在短期内惠及两项由国家科学基金会支持的具体全球活动：平方公里阵列(SKA)和iPLANT合作。国际资助的SKA将在南非境内定位仪器和数据处理设备，CHPC将支持该仪器的计算要求，该仪器将为美国和世界各地的研究人员提供数据。美国国家科学基金会资助的iPLANT项目与盖茨基金会国际育种平台(IBP)合作，共享非洲主要粮食作物基于计算的分子育种工具。TACC提供的基础设施和培训将帮助坦桑尼亚和南非的HPC中心在当地托管必要的工具和数据，以提高非洲主要作物的生产率。产生的数据集将在全球范围内共享，为美国植物科学研究人员提供丰富的比较数据集，他们的研究往往局限于美国用于食品生产的作物品种，进一步受到农业公司私人持有的数据的限制。这项活动的主要更广泛影响将是激发人们对高性能计算的兴趣和技能，这将鼓励学生学习构建和管理集群计算资源，并帮助科学家了解如何使用HPC来推动改变世界的发现。随着非洲数据处理能力的提高，并基于本研讨会支持的会议，美国和非洲的研究人员和他们的学生之间可能会开展新的合作。此外，HPC的改进有可能改善这三个非洲国家的以计算机为重点的教育和科学教育。该项目由主任办公室的国际科学和工程和计算机和信息科学与工程局的高级网络基础设施共同资助。

This workshop will enhance the potential for U.S. scientists to collaborate with African colleagues on a wide range of data-intensive projects by helping African scientists improve local supercomputing capacity. The workshop will be organized and managed by Tommy Minyard, Director of Advanced Computing Systems at the Texas Advanced Computing Center (TACC) of the University of Texas and Austin, and will build upon the 2013 donation to three African countries of almost 2000 compute nodes from the decommissioned ?Ranger? supercomputer, which was itself funded by NSF. The 2013 donation of Ranger computing components to institutions in Botswana, Tanzania, and South Africa was funded independently by the University of Texas. This workshop will help ensure that the African scientists who received this computing infrastructure have the skills and information necessary to use this resource to advance discoveries in astronomy, agriculture, climatology, earth sciences, transportation, telecommunications and many more fields of study that are of interest and value to the U.S. and Africa. TACC will organize a series of training and information exchange events to support broader understanding and use of supercomputing in Africa and to build a foundation for US-Africa collaborative research projects. In November 2014 approximately eight African researchers will visit TACC following their attendance at the SuperComputing Conference in New Orleans. The Austin visit will allow demonstration of a broad range of applications and tools that will enhance the African scientists? understanding and ability to facilitate the second phase of the workshop in South Africa. In addition visiting Austin will allow the African scientists to meet and discuss potential collaboration with several U.S. researchers using supercomputing to study topics in or relevant to southern Africa. In December 2014 the South African Centre for High Performance Computing (CHPC) will host a second, much larger workshop activity. The December 2014 event will follow the annual South African Supercomputing Conference hosted by CHPC and participants from this conference will be invited to attend the subsequent workshop. Six TACC staff members will travel to South Africa to conduct three-days of intensive discussions and trainings on configuring servers into single clusters and installing and using appropriate open source applications for a range of research activities. TACC staff will also discuss the emerging multi-national African high performance computing (HPC) development plan with their African colleagues. The enhancement of HPC capacity in Africa is likely to produce near term benefits to two specific NSF-supported global activities: the Square Kilometer Array (SKA) and the iPlant Collaborative. The internationally-funded SKA will be locating instruments and data processing equipment within South Africa and the CHPC will be supporting the computational requirements of this instrument, which will produce data for researchers in the United States and around the world. The NSF-funded iPlant project works with the Gates Foundation International Breeding Platform (IBP) to share computationally based molecular breeding tools for staple food crops within Africa. The TACC-provided infrastructure and training will help HPC Centers in Tanzania and South Africa to locally host the tools and data necessary to improve productivity in staple crops in Africa. The datasets produced will be shared around the globe, providing rich comparative datasets to U.S. plant science researchers, whose research is often limited to the strains of crops used for food production in the U.S. and further limited by data held privately by Agribusiness firms.The primary broader impact of this activity will be spurring interest and skills in high performance computing that will encourage students to learn to build and manage cluster computing resources and help scientists to understand how they can use HPC to power discoveries that change the world. New collaborations between researchers and their students in the U.S. and Africa are likely to develop as African capacity for data processing improves and based on the meetings that this workshop supports. In addition, improvements in HPC have the potential to improve both computer-focused education and science education in these three African countries.This project is co-funded by International Science and Engineering in the Office of the Director and Advanced Cyberinfrastructure in the Directorate for Computer and Information Science and Engineering.