Collaborative Research: Elements: A task-based code for multiphysics problems in astrophysics at exascale

California Institute Of Technology, Pasadena CA

Investigators

Abstract

Current and upcoming computers will run at exascale, over a hundred times more powerful than typical machines of today. Many algorithms used in current codes will not be able to take advantage of these new machines. The researchers will complete the development of an open-source community code for multi-scale, multi-physics problems in astrophysics and gravitational physics. The code uses transformative algorithms to reach the exascale. The techniques can be applied across discipline boundaries in fluid dynamics, geoscience, plasma physics and nuclear physics and engineering. The development of this new code has been driven by the current deployment of gravitational wave detectors such as LIGO. To fully understand and analyze the signals and waveforms measured with such detectors, it is essential that accurate, robust, and efficient computational tools be available for solving the dynamical Einstein equations over very long time scales. The recent detection of the merger of a neutron star-neutron star merger by LIGO and by a host of electromagnetic telescopes has ushered in the era of multi-messenger astronomy. The extreme energy densities of matter and radiation and the highly dynamic spacetimes of these events probe fundamental physics inaccessible to terrestrial experiments. The new code will be made available as open-source community cyberinfrastructure. The researchers will reach out to other communities within astrophysics (e.g., star formation, space plasma physics) and across discipline boundaries to fluid dynamics, geoscience, plasma physics, nuclear engineering etc. Young researchers trained in these techniques are in great demand, both in academia and as highly-skilled members of the industrial STEM workforce. Undergraduates will participate in the research by producing visualizations. The new code uses discontinuous Galerkin methods and task-based parallelism to accomplish its desired goals. This framework will allow the multi-physics applications to be treated both accurately and efficiently on the new architectures of petascale and exascale machines. The code is designed to scale to over a million cores for efficient exploration of the parameter space of potential sources and allowed physics, and for the high-fidelity predictions needed to realize the promise of multi-messenger astronomy. The code will allow astrophysicists to explore the mechanisms driving core-collapse supernovae and the properties of stellar remnants, to understand electromagnetic transients and gravitational-wave phenomena in compact objects, and to reveal the dense matter equation of state. The two key algorithmic innovations in the code, the discontinuous Galerkin method coupled with task-based parallelism, promise revolutionary impact in other fields relying on numerical solution of partial differential equations at the exascale. This project advances the objectives of "Windows on the Universe: the Era of Multi-Messenger Astrophysics", one of the 10 Big Ideas for Future NSF Investments. This project advances also the objectives of the National Strategic Computing Initiative (NSCI), an effort aimed at sustaining and enhancing the U.S. scientific, technological, and economic leadership position in High-Performance Computing (HPC) research, development, and deployment. This project is supported by the Office of Advanced Cyberinfrastructure in the Directorate for Computer & Information Science & Engineering and the Division of Physics and the Division of Astronomical Sciences in the Directorate of Mathematical and Physical Sciences. This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the Foundation's intellectual merit and broader impacts review criteria.