To some extend we are already simulating the universe, from the smallest scales like lattice QCD simulations because we can not analytically solve the problems to the largest scales to test or hypothesizes about the history of the universe and everything in between. But how would this help to shed any light on the problem of a [non-]simulated universe?

And as mentioned before, making predictions seems even more problematic to me. Where would you even start predicting things? Nobody has ever seen a universe and a universe simulator next to each other and could therefore say we should watch out for this or that to tell the two apart. You probably don't even have good reasons to assume that a simulation would be anything like the outside universe, our universe could be the equivalent of a Tetris level running on a phone somewhere in the real universe.