[dekhn]

Sure, a DC full of computers can fully simulate a sucrose molecule at a detailed level of description. You wouldn't need a DC- it could be done on a single machine. The real question is, why would you need to model things at the molecular level of detail if that detail is not necessary to recapitulate the behavior of a cell?

One thng I've learned from over a decade of simulating proteins and nucleic acids is that those methods, while mathematically interesting, don't provide useful data given the amount of resources they require. Instead, reduced models (effectively embeddings) and careful statistical methods are much, much more productive.