Non-Iterative Triples Corrections to EOM-CCSD in 5.1 version

Questions
#1

Dear Q-Chem Community,
I am trying to reproduce the examples 7.65-7.67 in the manual (vers. 5.1) with and without
the corrections fT/dT, the numerical values in the output are identical in both cases. I would expect some difference, or perhaps I am doing something wrong?

#2

In general they should not be identical, e.g, for CH+, DZP basis set (one of the sample jobs provided with Q-Chem):

 dE(fT)=-0.0013798, E_CCSD(fT)=-38.0190500

 dE(dT)=-0.0017826, E_CCSD(dT)=-38.0194527

I don’t have a version 5.1 manual so I can’t see the examples to which you refer.

#3

It is available here

pp. 385-386
I can’t attach the outputs, but all the values are identical for both eom-ccsd(ft) and eom-ccsd.

#4

I have checked the version 6.0 and it’s working. This the example 7.107 from manual for 6.0 (p. 572).
This was a trial 6.0 version and we have only 5.1 license. Perhaps this problem can be somehow circumvented in the 5.1 version? Some additional keywords or some patch?

#5

Hello @Leo137 ,

Could you try adding CCMAN2 = FASLE in $rem section and see if it works?

Thanks.

#6

It doesn’t. Different jobs fail with different errors, e.g.

CCMAN error!NSDavidsonRight<T>::NSDavidsonRight() : invalid Nroots

or

*Nls: 441
lsGoal: 343
 Q-Chem fatal error occurred in module libint/aoloop_2BSet.C, line 672:
 AOI batch size too small.
 Please submit a crash report at q-chem.com/reporter
#7

Could you send the output file to support@q-chem.com? I used 5.1.2 for the provided example and it finished with any issues. Note that if you are not using 5.1.2, you can upgrade to it without additional charge.

#8

Yes, I’m using the version 5.1.2.
I have sent the outputs. The original example finishes without any issues, but does not contain the desired triples, see my first post.