The shrinking domain framework I: a new, faster, more efficient approach to cosmological simulations

The advent of the new generation of wide field galaxy surveys makes large N-body cosmological simulations a necessary evil. While the cosmological simulation codes have evolved a lot since the first calculations in the 80s, the computational requirements for generating data that is relevant for large surveys remain challenging. We propose an alternative approach that can speed up these simulations. The framework is based on the idea of reducing the size of the integration region following the lightcone of an observer at redshift zero and thus simulating only the parts of the Universe that can be observed. A possible implementation of this framework is presented, as well as tests of its accuracy and performance. These simple tests, based on conservative assumptions, show that the new framework gives a factor of three speed up with respect to the usual approach.