Computer models are used for all sorts of applications, from designing cars to predicting the weather. They are even used to simulate the whole universe, to work out how it was born, and what might happen next. Until now, the universe has been modelled using numerical simulations, which are quick and simple, but limited by the assumptions they make and their use of Newtonian gravity instead of Einstein’s general relativity.

A typical numerical simulation assumes that the universe is isotropic and homogeneous, meaning that all of its matter is distributed evenly throughout. This is true on a large scale, but on smaller scales the matter is gathered into clusters of galaxies and dark matter, and the rest of space is empty. This means that expansion of the universe occurs at different rates in different places – spots dense with matter will be pulled closer together by their gravity and expand slower, and empty spaces will expand unhindered (28% faster than the average rate of expansion!).

Two sets of code have now been written independently using…