Cosmological neutrino simulations at extreme scale

Abstract

Constraining neutrino mass remains an elusive challenge in modern physics. Precision measurements are expected from several upcoming cosmological probes of large-scale structure. Achieving this goal relies on an equal level of precision from theoretical predictions of neutrino clustering. Numerical simulations of the non-linear evolution of cold dark matter and neutrinos play a pivotal role in this process. We incorporate neutrinos into the cosmological N-body code CUBEP 3 M and discuss the challenges associated with pushing to the extreme scales demanded by the neutrino problem. We highlight code optimizations made to exploit modern high performance computing architectures and present a novel method of data compression that reduces the phase-space particle footprint from 24 bytes in single precision to roughly 9 bytes. We scale the neutrino problem to the Tianhe-2 supercomputer and provide details of our production run, named TianNu, which uses 86% of the machine (13,824 compute nodes). With a total of 2.97 trillion particles, TianNu is currently the world's largest cosmological N-body simulation and improves upon previous neutrino simulations by two orders of magnitude in arXiv:1611.01545v2 [astro-ph.CO] 28 Jul 2017 2 Emberson et al.scale. We finish with a discussion of the unanticipated computational challenges that were encountered during the TianNu runtime.

On the road to per cent accuracy – III. Non-linear reaction of the matter power spectrum to massive neutrinos

“…We compute our fiducial non-linear power spectra and halo properties with the publicly available N-body code cubep 3 m (Harnois-Déraps et al 2013), which has been modified to include neutrinos as a separate set of particles (Inman et al 2015;Emberson et al 2017). We run a suite of simulations both with and without neutrino particles.…”

Status of Neutrino Properties and Future Prospects—Cosmological and Astrophysical Constraints

“…The level of complexity of Nbody simulations has been increasing over the years, so that the physical processes included in the simulations and the final results are much closer to the observations than they used to be at the beginning. Recent examples are given by the MassiveNuS [69] suite, based on the Gadget-2 code [70] modified to include the effects of massive neutrinos, the DEMNUni suite [71][72][73], the TianNu simulation [74][75][76], the BAHAMAS project [77], the gevolution simulations [78], and the nuCONCEPT simulations [79] (see also [80] for a method combining the particle and fluid descriptions) 7 . Nevertheless, the uncertainties related to the non-linear evolution of cosmological structures are still higher than those affecting the linear theory, therefore reducing the constraining power coming from the inclusion of those scales in cosmological analysis.…”

Parity-odd neutrino torque detection

“…An accurate study of the neutrino torque requires mass and force resolutions to cover the wide range of halo mass, a large box size (> 600 Mpc/h) to account the neutrino tides at distance, and neutrino particles/fluids, further studies of reconstruction of I R (Appendix A) to calculate more precise nonlinear neutrino effects on an evolving halo. These require future simulations with computing power comparable to that of TianNu [28].…”