I've been in the ATLAS Experiment for about 10 years and it still boggles my mind that it takes about 10 hours to simulate the detector's response to particles from 100 proton-proton collisions. (not bunch collisions - p-p collisions)