MRI: Acquisition of a High-Performance Computing Cluster to Unveil the Sources of Gravitational Waves

The first direct detections of gravitational waves, this past year, from two merging black-hole binary systems has ushered in the era of gravitational-wave astronomy. These observations from the Advanced Laser Interferometer Gravitational-wave Observatory (Advanced LIGO), and expected future observations, bring two major questions to the foreground. First, what do the observations of gravitational waves tell us about the particular astrophysical sources that launch those waves? Second, and more globally important, what do the growing variety of gravitational-wave observations tell us about the populations of exotic objects, such as black holes and neutron stars (stellar remnants), which are nearly impossible to detect in any other way? Related to this last question, can we bring other observational methods to bear on those gravitational-wave sources, using methods like optical, X-ray, or even neutrino telescopes (for instance), to learn more about these sources? This last question moves us beyond gravitational-wave astronomy, and into multi-messenger astronomy, with astrophysicists utilizing information from beyond just the electromagnetic spectrum. Moving into multi-message astronomy is one of the NSF's six "Big Ideas", singled out for their particular importance for future research. PI Kalogera, her collaborators, and the rest of the proposing team are at the forefront of addressing all of these questions; central to their work is the use of a dedicated computer cluster for both data analysis & simulations of gravitational-wave data & sources from Advanced LIGO. The proposed cluster will consist of 53 general-purpose computational nodes, 9 nodes with dual General-Purpose Graphics Processing Units (GPGPUs), and three nodes with a large amount of memory (512 GB as opposed to 128 GB, as normal). We also request funds for a 65TB storage system, directly connected to the cluster, for storing simulations for large-scale analyses.