With reference to LMP2 "An Integral-Direct Linear-Scaling Second-Order Møller−Plesset Approach" 2016 (which was recently credited with enabling the largest correlated calculation ever performed), are analytic gradients available at this time? If not, are there plans in this regard? Geometry optimization and density calculation (results given in NBO format) would be most useful.
NBO input format would also be useful for investigating the revSCAN functionals when they become available. There is a particular error in density calculation when DFT methods are applied to proteins, and NBO is good at detecting this error. The existing SCAN functionals are not an improvement over many other DFT methods in this regard, but the principles of SCAN seem worth persisting with.
