DFCOEF not generated with .NOCC and --get=DFCOEF

[Carles]

Hello,

I tried to run a simple SCF calculation on H2O with B3LYP as in the official DIRAC tutorial, using --get=DFCOEF and also adding .NOCC in the input to ensure that DFCOEF would be generated.

However, DIRAC/25 only generates a CHECKPOINT.h5 file — never DFCOEF — even though the manual says it should.

Also, the .h5 file claims to be valid for restart, but when I try to use it later with .INIWFC DHFSCF in a LUCITA calculation, it fails with:

    ERROR: trying to read non-existent data set /result/wavefunctions/scf/mobasis/nz from CHECKPOINT.h5

Is this expected behavior in DIRAC/25?  
Should DFCOEF still be written when requested?  
Or is there a specific compile flag or keyword needed?

Many thanks in advance for any help,  

Carles

[Hans Jørgen Aagaard Jensen]

Dear Carles

 

Unfortunately the manual is not fully updated to the removal of DFCOEF, we plan to do that next week when we have our annual Dirac workshop. We will also look into the LUCITA issue, you mention.

 

Hans Jørgen.

 

Fra: <dirac...@googlegroups.com> på vegne af Carles <carles...@gmail.com>
Svar til: "dirac...@googlegroups.com" <dirac...@googlegroups.com>
Dato: torsdag den 29. maj 2025 kl. 17.51
Til: dirac-users <dirac...@googlegroups.com>
Emne: [dirac-users] DFCOEF not generated with .NOCC and --get=DFCOEF

 

Du får ikke ofte mails fra carles...@gmail.com. Få mere at vide om, hvorfor dette er vigtigt

--
You received this message because you are subscribed to the Google Groups "dirac-users" group.
To unsubscribe from this group and stop receiving emails from it, send an email to dirac-users...@googlegroups.com.
To view this discussion visit https://groups.google.com/d/msgid/dirac-users/ccf81bd6-1005-4677-9728-7acbc3e14852n%40googlegroups.com.

[Carles]

Dear Hans,

Thank you so much, this is a great help. For your reference, I am attaching the input and molecular files of the ultimate calculation that is failing.

Best regards,

Carles

El dia dijous, 29 de maig del 2025 a les 17:55:54 UTC+2, h...@sdu.dk va escriure:

[Carles]

Dear Hans,

I hope the workshop went well! Just a quick follow-up on the DFCOEF removal and the LUCITA restart issue, any updates would be very appreciated.

Cheers,

Carles

El dia dijous, 29 de maig del 2025 a les 18:10:00 UTC+2, Carles va escriure:

[Visscher, L. (Luuk)]

Dear Carles,

To make sure: can you give the exact link to “the official DIRAC tutorial” that you are using ? 

Also if you modified the input, please attach the output of both the first and the second run to make sure we can reproduce your issue. Restarting SCF with the checkpoint file is tested in various tests and no problems are known at present.

best regards,

Luuk

To view this discussion visit https://groups.google.com/d/msgid/dirac-users/1a146425-043b-455b-a749-d84ca53ce739n%40googlegroups.com.

[Carles]

Dear Luuk,

Thank you for your reply. The link to the manual I am using is:

https://www.diracprogram.org/doc/release-25/

I attach here the input files of the calculation that is giving problems and I would appreciate that you could test in your system.

Best regards,

Carles

El dia divendres, 13 de juny del 2025 a les 15:14:30 UTC+2, Lucas Visscher va escriure:

[Visscher, L. (Luuk)]

Dear Carles,

The code works normally. You forgot to supply 

.WAVE FUNCTION

 in your input. See https://www.diracprogram.org/doc/release-25/tutorials/getting_started.html

best regards,

Luuk

To view this discussion visit https://groups.google.com/d/msgid/dirac-users/7f4b689d-fa9d-4699-b9dc-7bc221b15587n%40googlegroups.com.
<co.mol><input.inp>

[Carles]

Dear Luuk,

I appreciate your reply. The point is that testing different commands, I arrived to the conclusion the installation in our cluster was not correct. For instance, if I just add the .WAVE FUNCTION I get the error here attached. Can you please send the input file that runs and finishes correctly in your system?

Best regards,

Carles

El dia divendres, 13 de juny del 2025 a les 17:33:05 UTC+2, Lucas Visscher va escriure:

[Carles]

here are the files

El dia divendres, 13 de juny del 2025 a les 18:23:04 UTC+2, Carles va escriure:

[Carles]

Dear Luuk,

I see I put .WAVE FUNCTION the wrong place..  it is now working, thanks a lot!

Carles

El dia divendres, 13 de juny del 2025 a les 18:23:45 UTC+2, Carles va escriure:

[Carles]

Dear Luuk,
Do you know if *RESPONSE can be used after a LUCITA calculation?
Cheers,
Carles

El dia dissabte, 14 de juny del 2025 a les 15:30:10 UTC+2, Carles va escriure:

[Andre Gomes]

hi carles,

the response module activated by the *RESPONSE keyword is completely independent from LUCITA -you can see it as depending only on the outcome of the scf.

if i understand what you want/mean: even if LUCITA runs before *RESPONSE, there will be no response done based on the MRCI wavefunction.

apart from mean-field wavefunctions, it is possible to carry our response calculations at the coupled cluster level with the EXACORR module, but please note that the input for that is separate from the one for *RESPONSE.

best,

andre

To view this discussion visit https://groups.google.com/d/msgid/dirac-users/611e1bda-9a41-47bf-9f08-369753256aban%40googlegroups.com.

[Carles]

Dear Andre,

Thank you for the clarification. What I would really need is the ability to compute quadratic response functions starting from a correlated wavefunction with a core hole, such as GASCI from LUCITA, in order to simulate x-ray difference-frequency generation signals. From what you say, this would require extending DIRAC’s response capabilities beyond SCF, to support quadratic response with MRCI-type wavefunctions. Do you think this could be feasible in the future?

Best regards,
Carles

El dia divendres, 20 de juny del 2025 a les 13:16:55 UTC+2, Andre Gomes va escriure:

[Andre Gomes]

hi carles, 

i am not aware of planned developments for mrci for core states. 

i'm not sure it would help specifically in your case, but we do have plans in exacorr to support core spectra with the response code. 

currently you can  do difference-frequency generation in the optical domain for closed-shell molecules with exacorr, using eom-cc quadratic response - standard cc quadratic response is likely to become publicly available by the next release.

best regards,

andre

To view this discussion visit https://groups.google.com/d/msgid/dirac-users/0afa1e63-a79c-48e0-8223-d820f63916a7n%40googlegroups.com.

[Andre Gomes]

On Fri, Jun 20, 2025 at 1:45 PM Andre Gomes <asp...@gmail.com> wrote:

hi carles, 

i am not aware of planned developments for mrci for core states. 

 

or on mrci response theory for that matter

[Carles]

Dear Andre,

Thanks a lot for the follow-up. I will definitely keep in touch, coupled-cluster with quadratic response in the x-ray regime could be very relevant for us at some point.

All the best,
Carles

El dia divendres, 20 de juny del 2025 a les 13:49:12 UTC+2, Andre Gomes va escriure: