Scientists have run a complete simulation of a virus, down to the atomic level, for 50 ns. This seems like a pretty big deal. Admittedly, we already do really sophisticated modeling of the interplay between cell receptors and viruses. Without a complete model of a cell, a viral model doesn't seem likely to do you a whole lot of good, and cells are waaay more complicated than viruses. Still, it's encouraging just to know that this kind of thing can be done. It's going to be an amazing breakthrough when a complete viral lifecycle can be simulated on an atomic level.

And hey, maybe it wouldn't be that hard to build a complete model of a ribosome (here's where we overstep my knowledge of cell biology). That seems likely to do a lot of good in and of itself (although I have no idea if it'd let you sidestep the computation necessary to solve protein-folding problems or not — if not, you'd be headed from one supercomputing problem straight into another).