[CP2K-user] [CP2K:14674] SCAN_error: KS energy is an abnormal value (NaN/Inf).

[Amitava Banerjee]

Hi Fabian,

Right now I am using installed version 8.1 in our cluster. I do not have 
access the trunk version. Btw I have tried TPSS, but it shows "TPSS 
functional not implemented with LSD". I think probably I have to try LIBXC 
version.
On Wednesday, February 10, 2021 at 12:32:23 AM UTC-7 fa... at gmail.com 
wrote:

> Hi,
>
> All meta-GGA functionals that are available in libxc should work for you (
> https://www.tddft.org/programs/libxc/functionals/libxc-5.1.0/).
> If you have access to the trunk version of cp2k you can try R2SCAN which 
> is a revised SCAN functional designed for improved SCF convergence, see 
> doi:  10.1021/acs.jpclett.0c02405 and  doi: 10.1021/acs.jpclett.0c03077
> If you are stuck with an older version of cp2k (8.1 or older) you could 
> give TPSS a try.
>
> Cheers,
> Fabian
> On Tuesday, 9 February 2021 at 18:26:40 UTC+1 am... at gmail.com wrote:
>
>> Dear Prof. Hutter
>>
>> Thanks for your response. I have used cut off 800, 1000. Now, I am using 
>> 1200. With 800 and 1000, I did not have the previous error, but SCF did not 
>> converge. So I have increased the cutoff to 1200.
>> Is there any other meta-GGA in cp2k, which can calculate stress tensor 
>> with UKS/LSD?
>>
>>
>>
>> On Thursday, February 4, 2021 at 5:24:46 AM UTC-7 jgh wrote:
>>
>>> Hi 
>>>
>>> the SCAN functional is notorious for bad convergence/stability. 
>>> I would first try to set 
>>> CUTOFF 800 
>>> REL_CUTOFF 80 
>>> If this doesn't work increase the cutoff further. 
>>>
>>> I also don't have any experience with 
>>> &XC_GRID 
>>> XC_DERIV NN50_SMOOTH 
>>> XC_SMOOTH_RHO NONE 
>>> &END XC_GRID 
>>> I would suggest not to use it without any thorough testing for 
>>> the SCAN functional. The problem is that meta-functionals need 
>>> consistent rho - grad.rho - tau, smoothing destroys this. 
>>>
>>> regards 
>>>
>>> Juerg Hutter 
>>> -------------------------------------------------------------- 
>>> Juerg Hutter Phone : ++41 44 635 4491 <+41%2044%20635%2044%2091> 
>>> Institut für Chemie C FAX : ++41 44 635 6838 <+41%2044%20635%2068%2038> 
>>> Universität Zürich E-mail: h... at chem.uzh.ch 
>>> Winterthurerstrasse 190 
>>> CH-8057 Zürich, Switzerland 
>>> --------------------------------------------------------------- 
>>>
>>> -----c... at googlegroups.com wrote: ----- 
>>> To: "cp2k" <c... at googlegroups.com> 
>>> From: "Amitava Banerjee" 
>>> Sent by: c... at googlegroups.com 
>>> Date: 02/03/2021 04:33AM 
>>> Subject: [CP2K:14674] SCAN_error: KS energy is an abnormal value 
>>> (NaN/Inf). 
>>>
>>> hi 
>>> I am using SCAN to optimize bulk metal and oxide systems with SCAN. 
>>> After completing 2 steps in the SCF wavefunction optimization, it gives 
>>> error 
>>>
>>> ******************************************************************************* 
>>>
>>> * ___ * 
>>> * / \ * 
>>> * [ABORT] * 
>>> * \___/ KS energy is an abnormal value (NaN/Inf). * 
>>> * | * 
>>> * O/| * 
>>> * /| | * 
>>> * / \ qs_ks_methods.F:890 * 
>>> ******************************************************************************* 
>>>
>>>
>>> Input is here: &GLOBAL 
>>> PROJECT_NAME Ni 
>>> RUN_TYPE CELL_OPT 
>>> PRINT_LEVEL LOW 
>>> FLUSH_SHOULD_FLUSH .TRUE. 
>>> &END GLOBAL 
>>>
>>>
>>> &MOTION 
>>> &CELL_OPT 
>>> TYPE DIRECT_CELL_OPT 
>>> MAX_DR 4.5E-04 
>>> MAX_FORCE 1.0E-04 
>>> RMS_DR 4.5E-04 
>>> RMS_FORCE 4.5E-04 
>>> MAX_ITER 200 
>>> OPTIMIZER BFGS 
>>> &END CELL_OPT 
>>> &GEO_OPT 
>>> OPTIMIZER BFGS 
>>> &END GEO_OPT 
>>> &END MOTION 
>>>
>>>
>>> &FORCE_EVAL 
>>> METHOD QS 
>>> STRESS_TENSOR ANALYTICAL 
>>>
>>>
>>> &DFT 
>>> BASIS_SET_FILE_NAME BASIS_MOLOPT 
>>> BASIS_SET_FILE_NAME BASIS_MOLOPT_UCL 
>>> POTENTIAL_FILE_NAME GTH_POTENTIALS 
>>> UKS 
>>>
>>> &MGRID 
>>> NGRIDS 5 
>>> CUTOFF 400 
>>> REL_CUTOFF 60 
>>> &END MGRID 
>>>
>>>
>>> &QS 
>>> METHOD GPW ! to optimize the geometry the GPW method will be used 
>>> &END QS 
>>>
>>>
>>> &SCF 
>>> MAX_SCF 400 
>>> EPS_SCF 1.0E-6 
>>> SCF_GUESS ATOMIC 
>>> &DIAGONALIZATION ON 
>>> &END DIAGONALIZATION 
>>> &MIXING 
>>> ALPHA 0.1 
>>> &END MIXING 
>>> &SMEAR 
>>> METHOD FERMI_DIRAC 
>>> &END SMEAR 
>>> &END SCF 
>>>
>>>
>>> &XC 
>>> &XC_FUNCTIONAL 
>>> &LIBXC 
>>> FUNCTIONAL MGGA_X_SCAN 
>>> &END LIBXC 
>>> &LIBXC 
>>> FUNCTIONAL MGGA_C_SCAN 
>>> &END LIBXC 
>>> &END XC_FUNCTIONAL 
>>> &XC_GRID 
>>> XC_DERIV NN50_SMOOTH 
>>> XC_SMOOTH_RHO NONE 
>>> &END XC_GRID 
>>>
>>>
>>> &vdW_POTENTIAL 
>>> POTENTIAL_TYPE NON_LOCAL 
>>> &NON_LOCAL 
>>> TYPE RVV10 
>>> PARAMETERS 15.7 0.0093 
>>> KERNEL_FILE_NAME rVV10_kernel_table.dat 
>>> &END 
>>> &END vdW_POTENTIAL 
>>> &END XC 
>>>
>>>
>>> &END DFT 
>>>
>>>
>>> &SUBSYS 
>>> &TOPOLOGY 
>>> COORD_FILE_FORMAT xyz 
>>> COORD_FILE_NAME NiNiO-pos-1.xyz 
>>> &END TOPOLOGY 
>>> &CELL 
>>> PERIODIC XYZ ! we are considering the system periodic in the three 
>>> directions 
>>> ALPHA_BETA_GAMMA 90 90 90 
>>> ABC 3.505798 3.505798 3.505798 
>>> &END CELL 
>>>
>>>
>>> &KIND Ni 
>>> ELEMENT Ni 
>>> BASIS_SET DZVP-MOLOPT-SR-GTH 
>>> POTENTIAL GTH-PBE-q18 
>>> &END KIND 
>>>
>>>
>>> &END SUBSYS 
>>>
>>> &END FORCE_EVAL. 
>>> -- 
>>> 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 cp... at googlegroups.com. 
>>> To view this discussion on the web visit 
>>> https://groups.google.com/d/msgid/cp2k/71a1f8cb-7220-4b57-a1fe-23e8b6f9bb1bn%40googlegroups.com. 
>>>
>>>
>>>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://lists.cp2k.org/archives/cp2k-user/attachments/20210210/e5700ce6/attachment.htm>