A cycle exact and semantically precise one, sure. One that is allowed to prune unobserved results, on the other hand could potentially just lazily evaluate or s

A cycle exact and semantically precise one, sure. One that is allowed to prune unobserved results, on the other hand could potentially just lazily evaluate or simulate approximate outcomes and skip whole chunks of the simulation. To make claims about the computational power needed to simulate something, we really need to first know at what level of precision, and whether there are e.g. any observable effects of taking shortcuts or not (e.g. if "outside time" are not observable "inside", then skipping anything that can be lazily evaluated only if actually referenced from something observable even if it makes computation slow might be fine)