Post-PASI Workshop - Solidifying networks and staying current in parallel computing; Seattle, Washington; November 12-18, 2011

【摘要】[1143988]；波士顿大学董事会该奖项将支持来自美国的两名高级研究人员和15名初级科学家，在Lorena a . Barba博士的指导下，参加并主持在华盛顿州西雅图举行的研讨会，作为2011年11月12日至18日同时举行的超级计算会议的边会。该研讨会是2011年1月在智利瓦尔帕莱索举行的泛美高级研究所（PASI）的后续研讨会，题为“美洲的科学计算：大规模并行的挑战”。本次研讨会的主要目标是加强PASI期间建立的网络，将年轻科学家与资深研究人员（PASI讲师）及其同行重新联系起来，并最终促进新的合作。所选择的场地将能够与高性能计算领域的顶级活动之一密切互动。最初的PASI汇集了许多计算科学领域的高级研究人员和不同层次的学生，专注于图形处理单元（GPGPU）加速器上的通用计算对科学生产力的影响。GPU集群正在成为高性能计算领域的领军人物，目前全球排名前5位的超级计算机中有3台是基于这种硬件的。在对这些加速器进行编程时，需要学习一些重要的经验教训，以及与理解算法和编程模型相关的更一般的问题，以实现预期的结果。拟议的讲习班将为这些参与者提供所需的网络，以继续他们在PASI开始的工作并实现新的合作，这将是非常宝贵的。

Abstract1143988: Barba, Lorena A.; Trustees of Boston UniversityThis award will support two senior researchers and fifteen junior scientists from the United States, under the direction of Dr. Lorena A. Barba, to attend and conduct a workshop in Seattle, Washington, as a side meeting to the Supercomputing-11 Conference held concurrently during November 12-18, 2011. The workshop is a follow up to a Pan-American Advanced Studies Institute (PASI) held in Valparaiso, Chile in January 2011, titled "Scientific Computing in the Americas: the challenge of massive parallelism." The primary goal of this workshop is to reinforce the network created during the PASI, reconnecting young scientists with senior researchers (PASI lecturers) and their peers, and, ultimately, to stimulate new collaborations. The venue chosen will enable close interactions with one of the top events in the field of high performance computing.The original PASI brought together many high-level researchers in computational science and students at various levels to focus on the impact of general-purpose computation on Graphics Processing Unit (GPGPU) accelerators on scientific productivity. GPU clusters are becoming a leading player in the high-performance computing arena, with 3 of the top 5 supercomputers in the world now based on this hardware. There are important lessons to be learned in programming these accelerators as well as the more general issues associated with understanding the algorithms and programming models to achieve desired results. The proposed workshop will be invaluable in providing the networking needed for these participants to continue with the work they started at the PASI and enable new collaborations.