DC-DFT; ADMM with Non-Hybrid DFA

[Will Borrelli]

Hi all, 

I am trying to implement DC-DFT using the ENERGY_CORRECTION feature. I want to do this using both a hybrid functional and a non-hybrid functional, but my system is too large to do this without ADMM. By doing ADMM T in the ENERGY_CORRECTION section with the hybrid functional everything works, but when I try this with the non-hybrid it does not. I get an error that ADMM can't work if there is no HF section within XC, which of course I do not have since the XC functional is not a hybrid. I do have a HF section under my ENERGY_CORRECTION section (see attached input file). 

Thanks in advance for any guidance. 

[Jürg Hutter]

Hi

let's use the following nomenclature for DC-DFT : SCAN@HF means use SCAN on top of HF density.

Your input suggests you want to do LDA@HF(ADMM).
For this to work you have to set ADMM to false in the ENERGY_CORRECTION section, and to avoid later problems
change the ADMM section to

&AUXILIARY_DENSITY_MATRIX_METHOD
ADMM_PURIFICATION_METHOD NONE
EXCH_CORRECTION_FUNC PBEX
EXCH_SCALING_MODEL NONE
METHOD BASIS_PROJECTION
&END AUXILIARY_DENSITY_MATRIX_METHOD

BTW the following combinations are currently possible with CP2K

GGA@GGA
GGA@hybrid(ADMM)
hybrid(ADMM)@hybrid(ADMM)

what is not working (due to a bug) is hybrid(ADMM)@GGA

regards
JH

________________________________________
From: cp...@googlegroups.com <cp...@googlegroups.com> on behalf of Will Borrelli <will....@gmail.com>
Sent: Thursday, December 5, 2024 3:08 AM
To: cp2k
Subject: [CP2K:20946] DC-DFT; ADMM with Non-Hybrid DFA

Hi all,
I am trying to implement DC-DFT using the ENERGY_CORRECTION feature. I want to do this using both a hybrid functional and a non-hybrid functional, but my system is too large to do this without ADMM. By doing ADMM T in the ENERGY_CORRECTION section with the hybrid functional everything works, but when I try this with the non-hybrid it does not. I get an error that ADMM can't work if there is no HF section within XC, which of course I do not have since the XC functional is not a hybrid. I do have a HF section under my ENERGY_CORRECTION section (see attached input file).

Thanks in advance for any guidance.

--
You received this message because you are subscribed to the Google Groups "cp2k" group.
To unsubscribe from this group and stop receiving emails from it, send an email to cp2k+uns...@googlegroups.com<mailto:cp2k+uns...@googlegroups.com>.
To view this discussion visit https://groups.google.com/d/msgid/cp2k/606e9679-5c12-416a-a607-f5beec955b87n%40googlegroups.com<https://groups.google.com/d/msgid/cp2k/606e9679-5c12-416a-a607-f5beec955b87n%40googlegroups.com?utm_medium=email&utm_source=footer>.

[Will Borrelli]

JH, 

Thank you for the reply, I will give that a go. I would also like to verify that my setup (see newly attached file - efp_pb0.inp) is correct for doing PBE0(admm)@HF(ADMM). The calculation seems to be working as expected but I would like to verify. Also, just to make sure I understand, your possibilities mean: 

GGA@GGA: GGA functional corrected with GGA density
GGA@hybrid(ADMM): GGA functional corrected with hybrid density with ADMM used

hybrid(ADMM)@hybrid(ADMM): hybrid functional with ADMM corrected with hybrid density with ADMM

Best,

Will

[Jürg Hutter]

Hi

yes, looks good to me.
I just submitted a patch, so now all possible combinations of functionals with ADMM approximations
will be possible (including hybrid(ADMM)@GGA)

fun(A)@ fun(B) means: use functional B to calculate a Kohn-Sham density and then evaluate the energy
using functional A.

regards
JH

________________________________________
From: cp...@googlegroups.com <cp...@googlegroups.com> on behalf of Will Borrelli <will....@gmail.com>

Sent: Thursday, December 5, 2024 4:52 PM
To: cp2k
Subject: Re: [CP2K:20952] DC-DFT; ADMM with Non-Hybrid DFA

To view this discussion visit https://groups.google.com/d/msgid/cp2k/ff356d97-d2c8-4ec3-8c77-e4cc3742fad7n%40googlegroups.com<https://groups.google.com/d/msgid/cp2k/ff356d97-d2c8-4ec3-8c77-e4cc3742fad7n%40googlegroups.com?utm_medium=email&utm_source=footer>.

[Will Borrelli]

Thank you Jurg! I really appreciate your help.

Will