D2h symmetry discrepancy between MRCC and DIRAC

Dear MRCC developers,

I recently came across a puzzling issue. Using the DIRAC interface there seems to be large discrepancy between MRCC results when C2v and D2h symmetries are used. The resulting energies in DIRAC are consistent between C2v and D2h. Also the MRCC results agree in the C2v case. However, in the D2h case all MRCC energies starting from the reference are wrong. I feel like I am missing something obvious, but I am unable to identify the problem. The reference determinant vectors in the corresponding MINP files are consistent between C2v and D2h, fort.55 files are consistent as well.

As a very simple and fast example I attach inputs and outputs for Ne atom correlating just the 2p electrons. The resulting CCSD(T) energies:
DIRAC C2v -128.837980709791
MRCC C2v -128.837980709789
DIRAC D2h -128.837980709791
MRCC D2h -146.402199565869

Thank you for any insight.

Lukas

Dear Lukas,
You can only use the Abelian double groups C^*_2v and C^*_2h and their subgroups, see section 5.3 in the manual.

Dear Prof. Kallay,
thank you for a quick reply. Of course, I was missing something obvious after all. I have to be more careful in reading the manual
Best regards

Lukas