Page 1 of 1

LUCITA calculation

Posted: 13 Sep 2020, 07:47
by cymantren
Dear Users and developers,

I tried to do a GASCI-Lucita calc. on B5, but it fails.
-----------
Output from 'grep -n ERROR' :
928: SIRSET-ERROR, NCONF .le. 0 : NCONF = 0
932: SIRSET - FATAL ERRORS DETECTED, SEE ABOVE
-------------

Because I have not yet experience with this kind of calculation, it may be due to input errors.

Please corect me.

Thank you !

Kind regards

Alfred

Re: LUCITA calculation

Posted: 13 Sep 2020, 08:06
by taylor
I am quite unfamiliar with doing GAS calculations, but from what little I know I don't quite understand your proposed calculation in terms of electron numbers. Your INACTIVE specification is 8 MOs total, accounting for sixteen electrons, and your GAS specification accommodates between 3 and 5 electrons in a single GAS shell. But B5 has 25 electrons, as you specify for your open-shell SCF starting calculation, which certainly corresponds to the expected electron number. Your GAS calculation seems to have a total of 21 electrons maximum?

Possibly I simply don't understand what you are trying to do. In particular, a GAS calculation with a single GAS space --- is this not something that could be done with a CASSCF or RASSCF calculation?

By the way, for a CASSCF/RASSCF starting guess I think it would be simpler (although there is nothing wrong with what you are doing) to start with a closed-shell SCF calculation on B5+. This is not only faster and likely to converge better in the SCF step, it also usually provides a better starting guess at the final orbitals.

Best regards
Pete

Re: LUCITA calculation

Posted: 13 Sep 2020, 10:18
by hjaaj
Always use newest release! I tried with dalton release/2018 - and it worked on my computer.

PS. I saw you wanted +5 in charge, so that is OK. It does not make sense to specify "3 5" electrons when you only have one GAS space (which is the same as a CAS space, as Pete writes). "3 5" for last gascip space would mean +7, +6, and +5 charge, and LUCITA can only handle one value for charge in each calculation. It seems it has not been programmed to test for this type of inconsistency, and that it uses the last number and therefore gives the number you want.

Re: LUCITA calculation

Posted: 20 Sep 2020, 11:25
by cymantren
Dear Mr. Taylor and Mr. Jensen,

thank you very much foryour reply. I have no experience with that kind of calculation and the problems that arises are certanily due to this fact.
I did'nt want to calculate a cation, but wanted to do a calculation which is similar to CAS(5,6).

For clearness I want to switch to the test-suite „energy_lucita_gasci“ namly HBr
^1A_1-State.
There are 8 active electrons which are distributed among the GAS1-3-shells.
My problem is with the input:
e.g. in C2v-sym, GAS 1: 5 8/ 2 1 1 0
Why 5 electrons, because a 5 electr. Would not lead to the ^1A_1-state ??
This is the minimal occupation, but it should be 8 elect. Otherwise there are no 36 elec.
What is meant by „minimal occupation“ ?
Later in the ouput of e.g. ROOT 1 there is
„..Occupation of active sets : 5 1 2“
….
All occupation add up to 8, but why there are only 3 entries ? I expected something according to GAS 1, 2 or 3.

Please tell my why and I will be able to change and adjust this to my B5-case.

Thank you very much.

Kindly regards

Alfred