CELL vs POISSON periodicities do not match. (I do not have an active POISSON section)

Michela Benazzi

unread,

Jan 14, 2025, 9:29:12 AM1/14/25

to cp2k

Hello everyone,

I am getting this error:
*** WARNING in pw_env_methods.F:742 :: The selected periodicities in the *** *** sections &CELL and &POISSON do not match

However, if you view my input file below, you will be able to see that I do not have an active POISSON section. SCF steps are converging as normal and my job is going through, so I am not quite sure if I should be worried or disregard the error. I would appreciate any help with interpreting it! Thank you kindly,

Michela

&GLOBAL
PROJECT_NAME 24h
RUN_TYPE MD
PRINT_LEVEL LOW
SAVE_MEM T
PREFERRED_DIAG_LIBRARY ScaLAPACK
&END GLOBAL

&MOTION
&MD
ENSEMBLE NPT_I
STEPS 25000
TIMESTEP 0.5
TEMPERATURE 1.2+03
&BAROSTAT
PRESSURE 1.0100000000000000E+00
TIMECON 50
&THERMOSTAT
TYPE NOSE
&NOSE
TIMECON 50
&END NOSE
&END THERMOSTAT
&END BAROSTAT
&THERMOSTAT
TYPE NOSE
&NOSE
TIMECON 50
&END NOSE
REGION GLOBAL
&END THERMOSTAT
&END MD

&PRINT
&TRAJECTORY SILENT
&EACH
MD 1
&END EACH
&END TRAJECTORY
&CELL SILENT
&EACH
MD 1
&END EACH
&END CELL
&VELOCITIES SILENT
&EACH
MD 1
&END EACH
&END VELOCITIES
&FORCES SILENT
&EACH
MD 1
&END EACH
&END FORCES
&RESTART SILENT
&EACH
MD 1
&END EACH
&END RESTART
&END PRINT
&END MOTION

&FORCE_EVAL
METHOD QS
STRESS_TENSOR ANALYTICAL
&DFT
BASIS_SET_FILE_NAME /opt/cp2k/data/BASIS_MOLOPT
POTENTIAL_FILE_NAME /opt/cp2k/data/POTENTIAL
&SCF
MAX_SCF 300
EPS_SCF 1.0000000000000002E-06
SCF_GUESS ATOMIC
ADDED_MOS 60
&DIAGONALIZATION T
ALGORITHM STANDARD
&END DIAGONALIZATION
&SMEAR T
METHOD FERMI_DIRAC
ELECTRONIC_TEMPERATURE 1.2000E+03
&END SMEAR
&MIXING T
METHOD BROYDEN_MIXING
ALPHA 5.0000000000000002E-02
&END MIXING
&PRINT
&RESTART OFF
&END RESTART
&END PRINT
&END SCF
&QS
EPS_DEFAULT 1.0000000000000000E-10
EXTRAPOLATION PS
METHOD GPW
&END QS
&MGRID
NGRIDS 5
CUTOFF 6.0000000000000000E+02
&END MGRID
&XC #should I add cutoffs?
&XC_FUNCTIONAL PBE
&END XC_FUNCTIONAL
&END XC
&END DFT
&SUBSYS
&CELL
A 11.452 0 0
B 0 11.452 0
C 0 0 34.356
PERIODIC XY
&END CELL
&KIND Al
BASIS_SET SZV-MOLOPT-SR-GTH
ELEMENT Al
POTENTIAL GTH-PBE-q3
&END KIND
&TOPOLOGY
COORD_FILE_NAME 64Al.xyz
COORD_FILE_FORMAT XYZ
NUMBER_OF_ATOMS 64
&END TOPOLOGY
&END SUBSYS
&END FORCE_EVAL

Frederick Stein

unread,

Jan 14, 2025, 9:30:52 AM1/14/25

to cp2k

Dear Michela,

By default, the Poisson solver assumes 3D-periodic systems (compare the manual https://manual.cp2k.org/trunk/CP2K_INPUT/FORCE_EVAL/DFT/POISSON.html).

Best,

Frederick

Michela Benazzi

unread,

Jan 15, 2025, 12:50:36 PM1/15/25

to cp2k

Hi Frederick, thank you for your response! However, I do not have an active POISSON keyword/section. Are you saying that the POISSON solver is being used anyways?

Michela

Frederick Stein

unread,

Jan 15, 2025, 1:03:59 PM1/15/25

to cp2k

Hi Michela,

Correct. CP2K defines default values for most input keywords. The default value for CUTOFF is 280 Ry by the way. In the manual, the defaults are provided in the format "<keyword>: <type of keyword> = <default value>".

Thus, you should always set the PERIODIC keyword.

Best,

Frederick

Michela Benazzi

unread,

Jan 21, 2025, 1:10:12 PM1/21/25

to cp2k

Hello Dr. Stein,

Thank you for the reminder. Just to update everyone on how I fixed it, my process was:

1. Trying IMPLICIT POISSON_SOLVER. That gave me this error: "No stress tensor is implemented for the implicit Poisson solver".
> My first MD step converged, then the simulation crashed and I got this error. I unfortunately do need a stress tensor (I was using ANALYTICAL) because I will eventually increase the pressure of my unit cell after converging at 1 bar. So, I tried:

2. ANALYTIC POISSON_SOLVER (kept ANALYTICAL STRESS_TENSOR). The error was: " Stress Tensor not tested for 2D systems".
> Even after checking CP2K source code, I was a bit confused. I am not sure if it is a CP2K limitation; I checked the papers linked under POISSON_SOLVER and I could not understand how to check for compatibility with different stress tensor. If anyone has a few minutes, I would love a tip!
3. ANALYTIC POISSON_SOLVER, NUMERICAL STRESS_TENSOR. Running smoothly with no errors.

Thank you kindly for your time!

Michela

Frederick Stein

unread,

Jan 21, 2025, 2:48:26 PM1/21/25

to cp2k

Hi Michela,

Yes, unfortunately, some combinations are either not implemented or well tested with 2D-periodic systems. Either employ numerical stress tensors or treat your system as 3D-periodic. In the latter case, you should set the CONSTRAINT keyword in the MOTION/CELL_OPT section.