MRI: Acquisition of a Computing Cluster for Gravitational-Wave and Multimessenger Astrophysics in the Era of LIGO Detections

点击翻译按钮获取中文摘要

This MRI award supports a team of researchers at the Rochester Institute of Technology's (RIT) Center for Computational Relativity and Gravitation (CCRG) to acquire, deploy, and maintain a new state-of-the art computational cluster, GreenPrairie, to be used in research at the frontiers of gravitational physics, relativistic astrophysics, advanced high-performance computation, and scientific visualization. GreenPrairie will be the main computational workhorse of the CCRG, and it will directly enable the research of undergraduate and graduate students, postdoctoral researchers, and faculty across four departments and three colleges at RIT. It will foster the growth of three graduate programs in astrophysics, mathematical modeling, and computer science, and support research on behalf of several large community-wide collaborations in numerical relativity, gravitational waveform source modeling, and relativistic astrophysics. Scientific visualizations from the cluster will be used to support a funded REU program on multi-messenger astrophysics, and programs at RIT's National Technical Institute for the Deaf to promote science to the deaf and hard-of-hearing communities. The cluster will also be a vehicle for public outreach events on science, mathematics, and computing through site visits and annual community-wide public exhibits.GreenPrairie consists of a 1296-core, high-speed, large-memory computer cluster, with 768 TB of attached storage, and will be housed within the "Black Hole Lab" computer facility at the CCRG. Research with GreenPrairie will focus on some of the most extreme phenomena in the universe, where the strongest gravitational and magnetic fields interact with ultra-relativistic matter and high-energy radiation, that can only be studied through advanced, large-scale computation and visualization. The cluster will be used for timely simulations of key astrophysical sources for advanced LIGO, which uses information from theoretical waveforms to interpret the observed signals and determine the nature of the astrophysical sources. Specifically, the cluster will be used to simulate black-hole binaries, and their merger waveforms, across the entire space of parameters relevant to advanced LIGO and next generation detectors, as well as fast response simulations triggered by gravitational wave detections. The cluster will also be used to study accretion problems involving close and/or merging binary black hole systems which are expected to be observable in the electromagnetic spectrum by current and future astronomical surveys, such as LSST. Finally, the cluster will be a testbed for developing novel computational techniques that will permit efficient computation of heterogeneous systems involving multiple kinds of physics with very different length scales, and multiple evolution schemes.