[dsign]

I've been wondering about a similar approach for biomolecular simulations, where exact computations are still a hard bottleneck. I wonder if something like this could give us a few orders of magnitude more speed.