[CP2K-user] [CP2K:18071] SCF run NOT converged
Krack Matthias (PSI)
matthias.krack at psi.ch
Tue Nov 22 13:17:38 UTC 2022
That’s your decision depending on the system and the properties you are interested in, e.g. you will need k points, if you want to plot the band structure.
On the other hand, a 3x3x3 supercell is roughly equivalent to using 3x3x3 k points with the 1x1x1 cell.
From: cp2k at googlegroups.com <cp2k at googlegroups.com> on behalf of eitan eidelstein <eitan.eidelstein at gmail.com>
Date: Tuesday, 22 November 2022 at 13:55
To: cp2k <cp2k at googlegroups.com>
Subject: Re: [CP2K:18071] SCF run NOT converged
So, your suggestion is not to use k points at all (or only Gamma point)?
On Tuesday, November 22, 2022 at 1:03:31 PM UTC+2 Matthias Krack wrote:
A supercell with at least ~12A in each direction is most likely more efficient than a primitive cell with a large number of k points. CP2K/Quickstep employs localized Gaussian functions as basis set and their interaction range is not impacted by the cell size. Therefore many computational steps do not benefit from a small cell. On the other hand, several features are not yet available with k points and thus, if k points are dispensable, more feature are available.
From: cp... at googlegroups.com <cp... at googlegroups.com> on behalf of eitan eidelstein <eitan.ei... at gmail.com>
Date: Tuesday, 22 November 2022 at 11:44
To: cp2k <cp... at googlegroups.com>
Subject: Re: [CP2K:18069] SCF run NOT converged
Thanks, I will try those.
Can you explain why MULTIPLE_UNIT_CELL 3 3 3 is helping for PERIODIC XYZ? and why 3 instead of 2?
On Monday, November 21, 2022 at 6:26:18 PM UTC+2 Matthias Krack wrote:
Hi
• try a supercell, e.g. 3x3x3 (MULTIPLE_UNIT_CELL keyword in &CELL and &TOPOLOGY section), instead of k points
• Increase the cutoff to 400 Ry
• Increase the smearing temperature if needed
HTH
M.
From: cp... at googlegroups.com <cp... at googlegroups.com> on behalf of eitan eidelstein <eitan.ei... at gmail.com>
Date: Monday, 21 November 2022 at 14:15
To: cp2k <cp... at googlegroups.com>
Subject: [CP2K:18066] SCF run NOT converged
Hi,
This calculation is not converging.
I will be delighted for any sugestion.
Thanks,
Eitan.
&FORCE_EVAL
METHOD Quickstep
&DFT
BASIS_SET_FILE_NAME BASIS_MOLOPT_UZH
#BASIS_SET_FILE_NAME BASIS_SET
#POTENTIAL_FILE_NAME POTENTIAL
POTENTIAL_FILE_NAME POTENTIAL_UZH
CHARGE 0
SPIN_POLARIZED .TRUE. # SPIN_POLIRIZED
MULTIPLICITY 1
RELAX_MULTIPLICITY 0.1
&MGRID
CUTOFF 200
&END MGRID
&POISSON
PERIODIC XYZ
poisson_solver periodic
&END POISSON
&SCF
SCF_GUESS ATOMIC
EPS_SCF 1.0E-10
MAX_SCF 1000
ADDED_MOS 100
&DIAGONALIZATION
ALGORITHM STANDARD
EPS_ADAPT 0.001
&END DIAGONALIZATION
&SMEAR ON
METHOD FERMI_DIRAC
ELECTRONIC_TEMPERATURE [K] 300
&END SMEAR
&MIXING
METHOD BROYDEN_MIXING
ALPHA 0.2
BETA 1.5
NBROYDEN 8
&END MIXING
&END SCF
&XC
&XC_FUNCTIONAL PBE
&END XC_FUNCTIONAL
&END XC
&KPOINTS
SCHEME MONKHORST-PACK 30 30 30
&END KPOINTS
&PRINT
&DOS
FILENAME dos
NDIGITS 14
&END DOS
&PDOS
FILENAME pdos
# print all projected DOS available:
NLUMO -1
# split the density by quantum number:
COMPONENTS .TRUE.
&END PDOS
&END PRINT
&END DFT
&SUBSYS
&PRINT
&SYMMETRY
ALL
&END
&END PRINT
&CELL # 8.104532945 4.0522664725
A [bohr] 8.104532945 4.0522664725 4.0522664725
B [bohr] 4.0522664725 8.104532945 4.0522664725
C [bohr] 4.0522664725 4.0522664725 8.104532945
PERIODIC XYZ
&END CELL
&KIND Co1
ELEMENT Co
BASIS_SET DZVP-MOLOPT-PBE-GTH-q17
#BASIS_SET DZV-GTH-PADE
#POTENTIAL GTH-PBE
POTENTIAL GTH-PBE-q17
#&BS
# &ALPHA
# &END
# &BETA
# &END
#&END BS
#MAGNETIZATION 1.0
&END KIND
&KIND Co2
ELEMENT Co
BASIS_SET DZVP-MOLOPT-PBE-GTH-q17
#BASIS_SET DZV-GTH-PADE
#POTENTIAL GTH-PBE
POTENTIAL GTH-PBE-q17
#&BS
# &ALPHA
# &END
# &BETA
# &END
#&END BS
#MAGNETIZATION -1.0
&END KIND
&KIND O
ELEMENT O
BASIS_SET DZVP-MOLOPT-PBE-GTH-q6
#BASIS_SET DZVP-GTH-PADE
#POTENTIAL GTH-PBE
POTENTIAL GTH-PBE-q6
&END KIND
&COORD
SCALED
Co1 0.0 0.0 0.0
Co2 0.5 0.5 0.5
O 0.25 0.25 0.25
O 0.75 0.75 0.75
&END COORD
&END SUBSYS
&END FORCE_EVAL
&GLOBAL
PROJECT CoO
RUN_TYPE ENERGY
PRINT_LEVEL MEDIUM
&END GLOBAL
--
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... at googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/cp2k/061d18c5-ce60-480d-851c-9715141d40d3n%40googlegroups.com<https://groups.google.com/d/msgid/cp2k/061d18c5-ce60-480d-851c-9715141d40d3n%40googlegroups.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+uns... at googlegroups.com
.
To view this discussion on the web visit https://groups.google.com/d/msgid/cp2k/0e614388-4c59-4540-b3cf-7b54182551f8n%40googlegroups.com<https://groups.google.com/d/msgid/cp2k/0e614388-4c59-4540-b3cf-7b54182551f8n%40googlegroups.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<mailto:cp2k+unsubscribe at googlegroups.com>.
To view this discussion on the web visit https://groups.google.com/d/msgid/cp2k/2e6f2ae5-ad82-4a32-bda2-3dbe4593e7f6n%40googlegroups.com<https://groups.google.com/d/msgid/cp2k/2e6f2ae5-ad82-4a32-bda2-3dbe4593e7f6n%40googlegroups.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/ZRAP278MB0827E3440CE462553DD0F738F40D9%40ZRAP278MB0827.CHEP278.PROD.OUTLOOK.COM.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://lists.cp2k.org/archives/cp2k-user/attachments/20221122/a8ebecdb/attachment-0001.htm>
More information about the CP2K-user
mailing list