Workshop on Petascale Computing and the Geosciences

这是一项规划活动，旨在加深对所需活动的理解，使地球科学研究界做好准备，利用NSF的战略计划，为科学和工程开发千万亿级的计算环境。它将把来自计算机科学、空间物理、海洋科学、大气科学和地球科学的参与者聚集在两个小研讨会上，旨在启动开发在千万亿级系统上表现良好的下一代地球科学研究代码的进程。长期目标是更好地定位地球科学家，以便在部署这些系统时能够使用千万亿级系统进行尖端研究。第一期研讨会将评估当前具有代表性的地球科学高端计算模型样本的可行性，分析现有算法的性能和结构，并描述地球科学家希望解决的千万亿级资源问题。相关专家将介绍有关性能测量和代码表征、性能建模和程序开发的新方法和工具的信息。预期的成果将是为早期合作或试点研究确定一些想法，目的是利用这些现代工具中的一些来研究和改进重大的地球科学研究应用。这些合作将在第一期讲习班之后进行大约六个月。结果和吸取的经验教训将在第二次讲习班上报告。将编写一份这一过程的公开报告，预计这将帮助更广泛的社区规划必要的步骤，使计算地球科学为使用千万亿级计算系统做好准备。预计这项活动将推动地球科学模拟能力的未来进展，使地球科学研究以外的应用扩展到使用地球科学计算代码的其他领域，如天气和季节性气候预报、沿海洪水模拟、地震灾害和空间天气。

This is a planning activity designed to develop understanding of the activities needed to prepare the geoscience research community to exploit NSF's strategic plan to develop a petascale computing environment for science and engineering. It will bring together participants from computer science, space physics, ocean science, atmospheric science, and Earth science in a pair of small workshops intended to jump-start the process of developing next-generation geoscience research codes that perform well on petascale systems. The long-term intent is to better position geoscientists to be able to conduct cutting-edge research with petascale systems when these are deployed. The first workshop will assess the viability of a representative sample of current, geoscience, high-end, computational models, analyze the performance and structure of existing algorithms, and characterize the problems that geoscientists wish to tackle with petascale resources. Appropriate experts will present information about new approaches and tools for performance measurement and code characterization, performance modeling, and program development. An intended outcome will be the identification of some ideas for early collaborations or pilot studies aimed at using some of these modern tools to study and improve significant geoscience research applications. These collaborations will be pursued for roughly six months after the first workshop. Results and lessons learned will be reported at the second workshop. A publicly available report of the process will be written and it is anticipated that this will help the broader community plan the steps needed to prepare computational geoscience for the use of petascale computing systems. It is anticipated that this activity will seed future advances in the capabilities of geoscience modeling, with applications beyond geoscience research to other areas in which geoscience computational codes are used, such as weather and seasonal climate forecasting, coastal inundation simulation, earthquake hazards, and space weather.