We present a framework for characterizing the large scale movement of baryons relative to dark matter in cosmological simulations, requiring only the initial conditions and final state of the simulation. This is performed using the spread metric which quantifies the distance in the final conditions between initially neighbouring particles, and by analysing the baryonic content of final haloes relative to that of the initial Lagrangian regions defined by their dark matter component. Applying this framework to the SIMBA cosmological simulations, we show that 40% (10%) of cosmological baryons have moved $> 1h^{-1}~ {\rm Mpc}^{-1}$ ($3h^{-1}~ {\rm Mpc}^{-1}$) by $z=0$, due primarily to entrainment of gas by jets powered by AGN, with baryons moving up to $12h^{-1}~ {\rm Mpc}^{-1}$ away in extreme cases. Baryons decouple from the dynamics of the dark matter component due to hydrodynamic forces, radiative cooling, and feedback processes. As a result, only 60% of the gas content in a given halo at $z=0$ originates from its Lagrangian region, roughly independent of halo mass. A typical halo in the mass range $M_{\rm vir} = 10^{12}$--$10^{13}{\rm M}_\odot$ only retains 20% of the gas originally contained in its Lagrangian region. We show that up to 20% of the gas content in a typical Milky Way mass halo may originate in the region defined by the dark matter of another halo. This inter-Lagrangian baryon transfer may have important implications for the origin of gas and metals in the circumgalactic medium of galaxies, as well as for semi-analytic models of galaxy formation and "zoom-in" simulations.

Cosmological simulations require the use of a multiple time-stepping scheme. Without such a scheme, cosmological simulations would be impossible due to their high level of dynamic range; over eleven orders of magnitude in density. Such a large dynamic range leads to a range of over four orders of magnitude in time-step, which presents a significant load-balancing challenge. In this work, the extreme adaptivity that cosmological simulations present is tackled in three main ways through the use of the code SWIFT. First, an adaptive mesh is used to ensure that only the relevant particles are interacted in a given time-step. Second, task-based parallelism is used to ensure efficient load-balancing within a single node, using pthreads and SIMD vectorisation. Finally, a domain decomposition strategy is presented, using the graph domain decomposition library METIS, that bisects the work that must be performed by the simulation between nodes using MPI. These three strategies are shown to give SWIFT near-perfect weak-scaling characteristics, only losing 25% performance when scaling from 1 to 4096 cores on a representative problem, whilst being more than 30x faster than the de-facto standard Gadget-2 code.

For the Royal Society Summer Science Exhibition 2016, the Institute of Computational Cosmology from Durham University created the Galaxy Makers exhibit to communicate our computational cosmology and astronomy research. In addition to the physical exhibit we created an online component to foster re-engagement, create a permanent home for our content and to allow us to collect the ever-more important information about participation and impact. Here we summarise the details of the exhibit and the successes of creating an online component. We also share suggestions of further uses and improvements that could be implemented for online components of other science exhibitions.