Speaker
Description
Almost every cosmological simulation of the past four decades has been run in a periodic cube. The choice is so universal that it is rarely stated as an assumption, yet it is one: it gives the simulated universe the topology of a $3$-torus. But what that assumption costs us and how to deal with the problems it introduces? A cube is not rotationally symmetric, so angular momentum is not conserved. The sum over infinitely many periodic images makes the gravitational force deviate from Newton's law in a direction-dependent way. The discrete set of allowed wavelengths cuts off large-scale power differently along different directions. None of this is motivated by observations. Why did the torus nevertheless became the standard? In what scenarios can we expect its artefacts to matter? What should a better, alternative simulation topology provide? As an example we present the first $\Lambda$CDM simulations in a spherical, cylindrical, and dodecahedral topologies, in which the cosmic web forms as expected from observations.