There is a strong need for high-accuracy and efficient modeling of extreme-mass-ratio binary black hole systems because these are strong sources of gravitational waves that would be detected by future observatories. In this article, we present sample results from our Teukolsky EMRI code: a time-domain Teukolsky equation solver (a linear, hyperbolic, partial differential equation solver using finite-differencing), that takes advantage of several mathematical and computational enhancements to efficiently generate long-duration and high-accuracy EMRI waveforms.
We emphasize here the computational advances made in the context of this code. Currently there is considerable interest in making use of many-core processor architectures, such as Nvidia and AMD graphics processing units (GPUs) for scientific computing. Our code uses the Open Computing Language (OpenCL) for taking advantage of the massive parallelism offered by modern GPU architectures. We present the performance of our Teukolsky EMRI code on multiple modern processor architectures and demonstrate the high level of accuracy and performance it is able to achieve. We also present the code’s scaling performance on a large supercomputer i.e. NSF’s XSEDE resource: Keeneland (a 201 TeraFLOP/s, 120-node HP SL390 system with 240 Intel Xeon 5660 CPUs and 360 NVIDIA Fermi M2070 graphics processors, with the nodes connected by an InfiniBand QDR network).
In this article we demonstrate that recent mathematical and computational advances made to our time-domain Teukolsky EMRI code have enabled it to achieve a high-level of accuracy and efficiency. We emphasize the computational advancements made, that make use of the OpenCL framework to take advantage of the massive parallelism offered by modern many-core GPU architectures. The order-of-magnitude gain in computational performance we obtain in this manner plays a critical role in our code achieving the desired level of accuracy and efficiency.
The ability to perform high-accuracy and long-duration EMRI computations has enabled various interesting advances in gravitational physics. Using data generated by this code we have been able to make significant contributions to the development of effective-one-body models and gravitational waveform generation, that will ultimately positively impact the data analysis of current and future detectors (such as NSF’s LIGO and future space-borne missions). In addition, results from our code have brought forth a better understanding of the “anti-kick” which is an intriguing aspect of the phenomenon of gravitational recoil in decaying binary systems due to gravitational wave emission. And finally, our code has also helped test Cosmic Censorship in the context of the capture of a small test particle by a near extremal Kerr black hole.
Justin McKennon, Gary Forrester and Gaurav Khanna. High Accuracy Gravitational Waveforms from Black Hole Binary Inspirals Using OpenCL. Proceedings of the 1st Conference of the Extreme Science and Engineering Discovery Environment: Bridging from the eXtreme to the campus and beyond. Article No. 14. 2012. arXiv:1206.0270v1 [gr-qc] [doi: 10.1145/2335755.2335808] [Free PDF]