ADIOS:一种使用数据重构解决存储和I/O问题的系统框架
讲演者:美国橡树岭国家实验室数据科学中心 Scott A. Klasky博士
时 间:2016年4月7日(星期四),下午3:00
地 点:武汉光电国家实验室 A101
邀请人:谢长生 教授
摘要:
随着百亿亿次级(Exascale)计算的来临,美国将建设用于最新科学与工程前沿的系统。我们必须扪心自问是:面对如此高精度的仿真以及在实验和观察中验证数据时所带来的新挑战,我们准备好了吗?我们将探讨我的团队已做的工作:通过ADIOS来处理I/O挑战。但是我们看到了一个存在于未来的问题,这个问题需要通过新的合作和研究来解决。目前,数据的增长速度快于计算资源的增长速度,因此必须超越以往的数据存储方法,来探索以及寻找新的技术,以理解数据中所携带的信息,并将它放入存储中来帮助我们获取知识。
简介:
Scott A. Klasky博士是美国橡树岭国家实验室的杰出科学家,同时也是佐治亚理工大学、田纳西大学诺克斯维尔分校和北卡罗莱纳州立大学的兼职教授。Klasky博士在德州大学奥斯汀分校获得物理学博士学位,并且在2013年获得R&D 100奖。
TITLE: ADIOS to Storage and I/O Problems the SIRIUS way.
Abstract: Exascale computing is coming and our nations will build systems to new frontiers and science and Engineering. The question we must ask ourselves is: Are we ready for such high fidelity simulations and all of the new challenges it poses when faced in validating the data with experiments and observations? We will explore the work my team has done with ADIOS to cope with the I/O challenges, but we see a problem in the future which we need new collaborations and research to help solve. Data is growing faster than computational resources and we most look beyond to old methods of save data and then explore, and look for new techniques to understand the information in the data as it is being generated and then place this into storage to help us into the Knowledge away.
Bio: Dr. Scott A. Klasky is a distinguished scientist at the Oak Ridge National Laboratory, and an adjunct professor at Georgia Tech University, the University of Tennessee at Knoxville, and the North Carolina State University. He holds a Ph.D. in physics from the University of Texas at Austin, and the recipient of a R&D 100 award in 2013.