[CP2K-user] [CP2K:18558] Re: Cutoff convergence for SCAN functional

Rajorshi Chattopadhyay rajorshichat at gmail.com
Sat Mar 18 14:53:29 UTC 2023
Previous message (by thread): [CP2K-user] [CP2K:18556] Re: Cutoff convergence for SCAN functional
Next message (by thread): [CP2K-user] [CP2K:18561] Request for assistance with MD simulation
Messages sorted by: [ date ] [ thread ] [ subject ] [ author ]
Hi,
I confused EPS_DEFAULT with EPS_SCF. The convergence data in the previous
email were done with wrong settings. The data with EPS_DEFAULT 1E-14 and
REL_CUTOFF 100 is as follows:

Cutoff

Rel cutoff

Energy

Charge

600

100

-1125.17821665946303

-0.0000000000

700

100

-1125.17827145715273

0.0000000000

800

100

-1125.17815095995729

-0.000000000

900

100

-1125.17818569659994

-0.000000000

1000

100

-1125.17816128538243

-0.000000000

1100

100

-1125.17813041874047

-0.000000000

1200

100

-1125.17820131280564

-0.000000000

1300

100

-1125.17820131281405

-0.000000000

1400

100

-1125.17817273531045

-0.000000000

1500

100

-1125.17817273531455

-0.000000000

I see an improvement in the total charge on the grids with these settings.
However, the energy convergence is still 10**-3 Ha.

I apologise for the confusion with the settings.

On Fri, 17 Mar 2023 at 20:20, Rajorshi Chattopadhyay <rajorshichat at gmail.com>
wrote:

> Hi Prof. Hutter,
>
> I thank you for your reply.
>
> I tried out with the settings as suggested by you. The data is as follows:
>
> Cutoff
>
> Rel cutoff
>
> Energy
>
> Charge
>
> 600
>
> 100
>
> -1125.17821814826993
>
> -0.0000000138
>
> 700
>
> 100
>
> -1125.17827294369863
>
> -0.0000000147
>
> 800
>
> 100
>
> -1125.17815245523525
>
> -0.0000000120
>
> 900
>
> 100
>
> -1125.17818718982608
>
> -0.0000000115
>
> 1000
>
> 100
>
> -1125.17816278218288
>
> -0.0000000122
>
> 1100
>
> 100
>
> -1125.17813191141613
>
> -0.0000000121
>
> 1200
>
> 100
>
> -1125.17820280116666
>
> -0.0000000130
>
> 1300
>
> 100
>
> -1125.17820280127080
>
> -0.0000000125
>
> 1400
>
> 100
>
> -1125.17817423150291
>
> -0.0000000108
>
> 1500
>
> 100
>
> -1125.17817423131828
>
> -0.0000000112
>
> 1600
>
> 100
>
> -1125.17813474004834
>
> -0.0000000119
>
> I still find the highest convergence of 10**-3 Ha.
>
> On Fri, 17 Mar 2023 at 12:37, Jürg Hutter <hutter at chem.uzh.ch> wrote:
>
>> Hi
>>
>> Maybe:
>>
>> EPS_DEFAULT  1.E-14
>>
>> REL_CUTOFF   100
>>
>> helps.
>> regards
>> JH
>>
>> ________________________________________
>> From: cp2k at googlegroups.com <cp2k at googlegroups.com> on behalf of
>> Rajorshi Chattopadhyay <rajorshichat at gmail.com>
>> Sent: Friday, March 17, 2023 12:27 PM
>> To: cp2k at googlegroups.com
>> Subject: Re: [CP2K:18549] Re: Cutoff convergence for SCAN functional
>>
>> Hello Fabian,
>>
>> I thank you for your reply.
>>
>> When I do cutoff convergence test with r2SCAN, I get the following energy
>> values till 1500Ry (REL_CUTOFF:60):
>>
>> Cutoff
>> Rel cutoff
>> Energy
>> Charge
>> 600
>> 60
>> -1125.17821812053535
>> -0.0000000141
>> 700
>> 60
>> -1125.17827283731867
>> -0.0000000147
>> 800
>> 60
>> -1125.17815235733997
>> -0.0000000121
>> 900
>> 60
>> -1125.17818711771633
>> -0.0000000128
>> 1000
>> 60
>> -1125.17816274805796
>> -0.0000000131
>> 1100
>> 60
>> -1125.17813293059771
>> -0.0000000132
>> 1200
>> 60
>> -1125.17820271100982
>> -0.0000000140
>> 1300
>> 60
>> -1125.17820271126402
>> -0.0000000141
>> 1400
>> 60
>> -1125.17817412922477
>> -0.0000000128
>> 1500
>> 60
>> -1125.17817458499053
>> -0.0000000139
>>
>> I find that the convergence is only till 10**-3 Ha.
>>
>>
>> On Fri, 17 Mar 2023 at 10:18, Fabian <belleflamme.fabian at gmail.com
>> <mailto:belleflamme.fabian at gmail.com>> wrote:
>> Dear Rajorshi,
>>
>> The original SCAN functional is known to be very sensitive to the
>> integration grid used during calculation.
>> An easy fix would be to switch to r2SCAN (acs.jpclett.0c02405) instead:
>>
>>       &XC
>>         &XC_FUNCTIONAL
>>           &MGGA_X_R2SCAN
>>           &END
>>           &MGGA_C_R2SCAN
>>           &END
>>         &END
>>       &END XC
>>
>> Typically, the cutoff should be converged at 1200Ry with SCAN.
>>
>> Kind regards,
>> Fabian
>> Rajorshi Chattopadhyay schrieb am Freitag, 17. März 2023 um 00:11:52
>> UTC+1:
>> Dear all,
>>
>> I am trying to do cutoff convergence with the SCAN functional in a
>> aqueous H2O-H2SO4 system. I need an energy convergence of 10**-6 Ha or
>> lower. However, I find that even at a cutoff of 4000Ry (REL_CUTOFF:60),
>> energy convergence is only 10**-4 Ha. The input script is as follows:
>>
>>
>> &GLOBAL
>>
>>     PROJECT Sulphate_amb_dens
>>
>>     RUN_TYPE ENERGY_FORCE
>>
>>     PRINT_LEVEL MEDIUM
>>
>> &END GLOBAL
>>
>> &FORCE_EVAL
>>
>> &DFT
>>
>>       MULTIPLICITY 1
>>
>>       BASIS_SET_FILE_NAME ./BASIS_MOLOPT
>>
>>       POTENTIAL_FILE_NAME ./SCAN_POTENTIALS
>>
>>   &MGRID
>>
>>       CUTOFF 3800
>>
>>       REL_CUTOFF 60
>>
>>       NGRIDS 4
>>
>>   &END MGRID
>>
>>   &SCF
>>
>>      SCF_GUESS ATOMIC
>>
>>      MAX_SCF 3000
>>
>>      EPS_SCF 1.00000000E-007
>>
>>      &OT
>>
>>         MINIMIZER DIIS
>>
>>         PRECONDITIONER FULL_SINGLE_INVERSE
>>
>>      &END OT
>>
>>   &END SCF
>>
>>   &XC
>>
>>      &XC_FUNCTIONAL
>>
>>          &LIBXC
>>
>>           FUNCTIONAL MGGA_X_SCAN
>>
>>          &END LIBXC
>>
>>          &LIBXC
>>
>>            FUNCTIONAL MGGA_C_SCAN
>>
>>            &END LIBXC
>>
>>      &END XC_FUNCTIONAL
>>
>>     &END XC
>>
>>    &END DFT
>>
>>   &SUBSYS
>>
>>     &TOPOLOGY
>>
>>      &CENTER_COORDINATES TRUE
>>
>>      &END CENTER_COORDINATES
>>
>>     &END TOPOLOGY
>>
>>
>>     &CELL
>>
>>        ABC 12.32 12.32 12.32
>>
>>        PERIODIC XYZ
>>
>>     &END CELL
>>
>>     &COORD
>>
>>       @include ../coord_eql
>>
>>    &END COORD
>>
>>    &KIND H
>>
>>       BASIS_SET DZVP-MOLOPT-SR-GTH
>>
>>       POTENTIAL GTH-SCAN-q1
>>
>>    &END
>>
>>    &KIND O
>>
>>       BASIS_SET DZVP-MOLOPT-SR-GTH
>>
>>       POTENTIAL GTH-SCAN-q6
>>
>>    &END
>>
>>   &KIND S
>>
>>       BASIS_SET DZVP-MOLOPT-GTH
>>
>>       POTENTIAL GTH-SCAN-q6
>>
>>    &END
>>
>>
>>   &END SUBSYS
>>
>>  &END FORCE_EVAL
>>
>>
>> I use the pseudopotentials provided by Dr. Joerg Hutter in his github.
>>
>> Am I missing something in the input script ? Is it possible to achieve
>> such tight energy convergences using SCAN the functional?
>>
>> I thank you in advance for your help/suggestions.
>>
>> --
>> With Regards,
>> Rajorshi Chattopadhyay,
>> PhD Researcher in Mineralogy/Crystallography,
>> Institut für Geologie und Mineralogie,
>> Universität zu Köln, Germany
>>
>>
>> --
>> 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+unsubscribe at googlegroups.com<mailto:
>> cp2k+unsubscribe at googlegroups.com>.
>> To view this discussion on the web visit
>> https://groups.google.com/d/msgid/cp2k/a68b39f8-10cc-4a1f-a6a4-a3aa8c0893ffn%40googlegroups.com
>> <
>> https://groups.google.com/d/msgid/cp2k/a68b39f8-10cc-4a1f-a6a4-a3aa8c0893ffn%40googlegroups.com?utm_medium=email&utm_source=footer
>> >.
>>
>>
>> --
>> With Regards,
>> Rajorshi Chattopadhyay,
>> PhD Researcher in Mineralogy/Crystallography,
>> Institut für Geologie und Mineralogie,
>> Universität zu Köln, Germany
>>
>> 
>>
>> --
>> 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+unsubscribe at googlegroups.com<mailto:
>> cp2k+unsubscribe at googlegroups.com>.
>> To view this discussion on the web visit
>> https://groups.google.com/d/msgid/cp2k/CACTJSRfhgJHLbZRTFmKnor5O58G%3DNriRQVaHm-2KK_xbuK-gEg%40mail.gmail.com
>> <
>> https://groups.google.com/d/msgid/cp2k/CACTJSRfhgJHLbZRTFmKnor5O58G%3DNriRQVaHm-2KK_xbuK-gEg%40mail.gmail.com?utm_medium=email&utm_source=footer
>> >.
>>
>> --
>> 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+unsubscribe at googlegroups.com.
>> To view this discussion on the web visit
>> https://groups.google.com/d/msgid/cp2k/ZR0P278MB075931AD020B7A19C3CA23E99FBD9%40ZR0P278MB0759.CHEP278.PROD.OUTLOOK.COM
>> .
>>
>
>
> --
> With Regards,
> Rajorshi Chattopadhyay,
> PhD Researcher in Mineralogy/Crystallography,
> Institut für Geologie und Mineralogie,
> Universität zu Köln, Germany
>
> 
>


-- 
With Regards,
Rajorshi Chattopadhyay,
PhD Researcher in Mineralogy/Crystallography,
Institut für Geologie und Mineralogie,
Universität zu Köln, Germany



-- 
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+unsubscribe at googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/cp2k/CACTJSRfjpzO_d6AmEtS5GPj2p2JzY1%2B_tS%2BB73P4zkU-qGwxuA%40mail.gmail.com.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://lists.cp2k.org/archives/cp2k-user/attachments/20230318/2858f4fd/attachment-0001.htm>
Previous message (by thread): [CP2K-user] [CP2K:18556] Re: Cutoff convergence for SCAN functional
Next message (by thread): [CP2K-user] [CP2K:18561] Request for assistance with MD simulation
Messages sorted by: [ date ] [ thread ] [ subject ] [ author ]
More information about the CP2K-user mailing list