IMR-MIP DANSE - Distributed Data Analysis for Neutron Scattering Experiments

位于田纳西州橡树岭的散裂中子源(SNS)正在建设中，预算为1.411美元，是世界上最大的科学建设项目。2006年，它将开始生产用作材料探测器的强流中子束。控制这些束流并探测从样品中散射的中子的仪器是最先进的。在SNS进行的中子散射实验将产生关于原子和自旋在材料、分子和凝聚态中的位置和运动的前所未有的详细数据。在国家科学基金会的IMR-MIP计划下，正在支持一项概念性工程设计工作，以建立软件，用于在一个名为DANSE分布式数据分析的系统中分析来自SNS和其他中子设施的数据，用于中子散射实验。DANSE项目包括一个中心资源活动，以及美国各地不同机构的中子散射科学不同子领域的子项目。牵头机构是加州理工学院。科学子项目由密歇根州立大学的教员领导。(衍射)，在爱荷华州立大学。(工程衍射)、马里兰大学(反射法)、田纳西大学(小角散射)、洛斯阿拉莫斯大学和加州理工学院。DANSE使用基于数据流范例的新软件架构。分析是通过可使用标准化数据流跨网络连接的可重用软件组件执行的。组件使用开放源码语言Python集成到一个连贯的解释框架中，因此可以在运行时轻松构建自定义分析过程。该体系结构支持在分布式资源上进行高性能计算，并打开了对基于网格计算的未来网络基础设施的访问。Danse提供了一个前所未有的机会，将数据分析、理论和模拟融合到一个统一的计算环境中。DANSE项目的目标是建立一个软件系统，1)使新的、更复杂的科学能够通过中子散射实验进行，2)使所有科学家更容易分析数据，3)提供一个强大的软件基础设施，可以在未来维护。散裂中子源正在田纳西州橡树岭建设，预算为14.11亿美元，是世界上最大的科学建设项目。2006年，它将开始生产用作材料探测器的强流中子束。控制这些束流并探测从研究中的样品中散射的中子的仪器是最先进的。在SNS进行的中子散射实验将产生有关原子在材料中的位置和运动的前所未有的详细数据。用这些仪器获得的原始实验数据并不容易解释，需要新的软件将数据转换为有用的形式。除了今天可用的这种数据缩减之外，还有机会利用计算材料科学在过去十年中取得的几项重大进展来解释数据。在NSF的IMR-MIP计划下，正在支持一项概念性工程设计工作，以建立一个名为DANSE-分布式数据分析的软件系统，用于中子散射实验。丹斯项目包括两个部分。第一个是软件工程方面的工作，目的是构建一个允许模块化软件组件互操作性的框架。第二个是美国各地不同机构的科学家努力为中子散射研究的不同子领域开发数据分析所需的软件组件。牵头机构是加州理工学院。科学子项目由密歇根州立大学、爱荷华州立大学、马里兰大学和田纳西大学的科学家领导。DANSE项目的目标是建立一个软件系统，该系统1)使新的、更复杂的科学能够通过中子散射实验进行，2)使所有科学家更容易分析数据，3)提供一个可在未来维护的强大的软件基础设施。

The Spallation Neutron Source (the "SNS"), under construction in Oak Ridge, Tennessee with a budget of $B1.411, is the world's largest science construction project. In 2006 it will start to produce intense beams of neutrons to be used as probes of materials. The instruments that control these beams, and detect the neutrons scattered from specimens, are state-of-the-art. Neutron scattering experiments performed at the SNS will produce data of unprecedented detail on the positions and motions of atoms and spins in materials, molecules, and condensed matter. Under the IMR-MIP program at the NSF, a conceptual engineering design effort is being supported to build software for the analysis of data from the SNS and other neutron facilities in a system called DANSE - distributed data analysis for neutron scattering experiments. The DANSE project includes a central resources activity, and subprojects in the different subfields of neutron scattering science at different institutions around the U.S. The lead institution is the California Institute of Technology. The scientific subprojects are led by faculty at Michigan State Univ. (diffraction), at Iowa State Univ. (engineering diffraction), the University of Maryland (reflectometry), the University of Tennessee (small-angle scattering), and Los Alamos, and Caltech.DANSE uses a new software architecture based on the data flow paradigm. Analysis is performed with reusable software components that can be connected across a network using standardized data streams. Components are integrated into a coherent interpretive framework using the open source language Python so that custom analysis procedures can be constructed easily at runtime. The architecture enables high performance computing on distributed resources and opens access to the future cyber infrastructure of grid-based computing. DANSE provides an unprecedented opportunity to merge data analysis, theory, and simulation into a uniform computing environment. The goals of the DANSE project are to build a software system that 1) enables new and more sophisticated science to be performed with neutron scattering experiments, 2) makes the analysis of data easier for all scientists, and 3) provides a robust software infrastructure that can be maintained in the future.The Spallation Neutron Source (the "SNS"), under construction in Oak Ridge, Tennessee with a budget of $ 1,411,000,000, is the world's largest science construction project. In 2006 it will start to produce intense beams of neutrons to be used as probes of materials. The instruments that control these beams, and detect the neutrons scattered from the specimens under study, are state-of-the-art. Neutron scattering experiments performed at the SNS will produce data of unprecedented detail on the positions and motions of atoms in materials. The raw experimental data acquired with these instruments are not simple to interpret, and new software is required to transform the data into useful forms. Beyond such data reductions that are available today, there is an opportunity to interpret data using several major advances in computational materials science that have occurred over the past decade. Under the IMR-MIP program at the NSF, a conceptual engineering design effort is being supported to build a software system called DANSE - distributed data analysis for neutron scattering experiments. The DANSE project includes two parts. The first is a software engineering effort to build a framework that permits the interoperability of modular software components. The second is an effort by scientistsat different institutions around the U.S. to develop the software components needed for data analysis for the different subfields of neutron scattering research. The lead institution is the California Institute of Technology. The scientific subprojects are led by scientists at Michigan State University, Iowa State University, the University of Maryland, and University of Tennessee. The goals of the DANSE project are to build a software system that 1) enables new and more sophisticated science to be performed with neutron scattering experiments, 2) makes the analysis of data easier for all scientists, and 3) provides a robust software infrastructure that can be maintained in the future.