Problems with VBPT2 energetics

[etai....@mail.huji.ac.il]

Fuming hello,
We have recently encountered the following problem: VBPT2 results (energies) using two different versions of XMVB give very different results (70(!!) kcal/mol difference).

When using the old version of XMVB (2.0 from Jan 2015) we get different energy than the energy we get from the new version (2.1 from Sep 2015). For example, the difference for the system OH- + CH3F in the reactant state is 70(!) kcal/mol. Although, xmi, gus and integrals files are identical.

example files are attached.

We will be happy to know which of the versions is correct.

Thank,
Itay Karach.