🚀 go-pugleaf

The dynamics server idea sounds interesting. How much compute power is
needed to do this in a reasonable time? Is the "standard regime" easily
partitionable? If so, a Distributed Dynamics Server might be employed.
To soon to volunteer this service, of course, but I'd be willing to try
running a few tests with such a package.  By sending identical dynamics
code and a standard test molecule to all potential server providers in
this group, you could generate relative benchmarks that would be useful
if this idea catches on, and a number of servers are established.

Ed Strong