James Sullivan, University of California, Berkeley

Photo of James Sullivan

The widely accepted Lambda-Cold Dark Matter (Lambda-CDM) cosmological model and associated simulations have been extremely successful in replicating observations at so-called linear scales in perturbation theory. Below these scales, non-linear effects become important for the growth of structure, complicating models. However, powerful high-performance computing machines are increasingly providing opportunities for modeling physics at non-linear scales, including massive neutrino physics. Detailed modeling of neutrinos in the context of cosmological structure formation is motivated at both the cosmological and particle scales. At cosmological scales, effects such as neutrino damping of the growth of cosmic structure can more precisely describe the observed large-scale matter distribution. On the particle scale, cosmological measurements have the power to constrain the total mass of the three observed neutrino flavors, whereas neutrino oscillation experiments have only measured the mean-square mass difference between flavors. Cosmological simulations with the Hardware/Hybrid Accelerated Cosmology Code (HACC) can achieve the resolution necessary to properly model massive neutrino physics by leveraging HACC's remarkable parallel performance. We extend the code through the implementation of an effectively higher-order technique for reducing noise in N-body simulations that include massive neutrinos.

Abstract Author(s): James Sullivan, Matthew Becker, Salman Habib