MRI-R2: Acquisition of a Compute Cluster for High-Fidelity Simulations of Gravitational Wave Sources -- Facilitating LIGO and Enabling Multi-Messenger Astronomy

MRI-R2：获取用于引力波源高保真模拟的计算集群——促进 LIGO 并实现多信使天文学

• 批准号: 0960291
• 负责人: Christian Ott
• 金额: $ 105万
• 依托单位: California Institute of Technology
• 依托单位国家: 美国
• 项目类别: Standard Grant
• 财政年份: 2010
• 资助国家: 美国
• 起止时间: 2010-04-01 至 2013-03-31
• 项目状态: 已结题
• 来源: https://www.nsf.gov/awardsearch/showAward?AWD_ID=0960291&HistoricalAwards=false
• 关键词: MRI; R2; Acquisition; Compute; Cluster

This award is funded under the American Recovery and Reinvestment Act of 2009 (Public Law 111-5). The award enables the acquisition of a computer cluster to allow the simulation of groundbreaking models of astronomical systems to yield high-fidelity, robust, quantitative predictions for multi-messenger astronomy in the form of gravitational waveforms, neutrino signatures, and electromagnetic signals. Gravitational waves (GWs) are ripples in the structure of spacetime that propagate across the universe at the speed of light. The NSF-funded Laser Interferometer Gravitational Wave Observatory (LIGO) is searching for these waves, which encode otherwise inaccessible details of their sources. This new information will be enriched and complemented by electromagnetic and neutrino observations of those sources, making multi-messenger astronomy a reality. Codes have been developed to take advantage of the growth in available computing power to allow improved, accurate modeling of several potential sources of gravitational waves, including double star systems containing black holes and/or neutron stars and the collapse of massive stars to yield supernovae. The cluster, to be located at Caltech's Center for Advanced Computing Research, will consist of approximately 2000 cores and approximately 100 TeraBytes (TB) of disk to start, with upgrades to approximately 325 TB by the end of the project. The machine be will be used by the Caltech/Cornell numerical relativity collaboration and by the Caltech Computational Astrophysics Group and collaborators at Princeton University, the University of Maryland and Louisiana State University. The cluster will concentrate on simulations where rapid turnaround time is required, on the development/testing of code to be run on (slow-turnaround, but much larger) TeraGrid/Lab resources, and on analysis/visualization of results obtained on TeraGrid/Lab machines as well as on the cluster itself. Graduate and undergraduate students and postdoctoral scholars will participate in the research and be trained to use the cluster. The cluster will foster extensive research collaboration across seven institutions and will significantly enhance the future scientific output of LIGO. The results of simulations performed on this cluster will be disseminated to scientists in all fields and to the general public, via web presentations, public lectures, articles, and movies.

该奖项是根据2009年美国复苏和再投资法案（公法111-5）资助的。该合同使计算机集群能够模拟天文系统的突破性模型，从而以重力波形、中微子特征和电磁信号的形式为多信使天文学提供高保真度、鲁棒性和定量预测。引力波（GWs）是时空结构中的波纹，以光速在宇宙中传播。美国国家科学基金会资助的激光干涉仪引力波天文台（LIGO）正在寻找这些引力波，这些引力波编码了它们来源的其他难以获取的细节。这些新信息将通过对这些来源的电磁和中微子观测得到丰富和补充，使多信使天文学成为现实。开发代码是为了利用可用计算能力的增长，对几种潜在的引力波源进行改进和精确的建模，包括包含黑洞和/或中子星的双星系统，以及产生超新星的大质量恒星的坍缩。该集群位于加州理工学院高级计算研究中心，将由大约2000个核心和大约100 TB的磁盘组成，到项目结束时将升级到大约325 TB。这台机器将由加州理工学院/康奈尔数值相对论合作项目、加州理工学院计算天体物理小组以及普林斯顿大学、马里兰大学和路易斯安那州立大学的合作者使用。集群将专注于需要快速周转时间的模拟，开发/测试在TeraGrid/Lab资源上运行的代码（周转缓慢，但规模更大），以及分析/可视化在TeraGrid/Lab机器上获得的结果以及集群本身。研究生、本科生和博士后学者将参与研究，并接受使用集群的培训。该集群将促进七个机构之间的广泛研究合作，并将显著提高LIGO未来的科学产出。在这个集群上进行的模拟结果将通过网络演示、公开讲座、文章和电影传播给所有领域的科学家和公众。