Restart files for MD with NPT_F ensemble not conserving all of the properties
Amandine Chouanard
achou... at gmail.com
Thu Aug 11 18:44:44 UTC 2016
Hello,
I am a beginner to cp2k and I have run into a problem. I am running a
molecular dynamics simulation with an NPT_F (flexible cell) ensemble, where
each a restart file is generated every 10 MD steps and resubmitted as the
input every 40 or so steps. However, I have noticed that the conserved
quantity increases after each restart resulting in a stepwise pattern (see
attached). The MD simulation runs without warning, though. I was wondering
if any one else has tried to use restart files with a flexible cell and if
so, what information is missing information in my restart files generated
to preserve the same conserved quantity..
Thanks a lot,
Amandine Chouanard
Additional information:
cp2k version: 2.7
Here is my input file: (The restart file is a bit long, so I can post it
separately if requested.)
&GLOBAL
PROJECT metaD-MgH2-
RUN_TYPE MD
PRINT_LEVEL LOW
&END GLOBAL
&MOTION
&MD
ENSEMBLE NPT_F
STEPS 100000
TIMESTEP 4.9999999999999994E-01
TEMPERATURE 2.9800000000000000E+02
&BAROSTAT
PRESSURE [bar] 1.0
&END
&THERMOSTAT
TYPE CSVR
REGION MASSIVE
&CSVR
TIMECON 4.9999999999999993E+01
&END CSVR
&END THERMOSTAT
&END MD
&PRINT
&RESTART SILENT
&EACH
MD 10
&END EACH
&END RESTART
&END PRINT
&FREE_ENERGY
&METADYN
DO_HILLS FALSE
&METAVAR
SCALE 1.0
COLVAR 1
&END METAVAR
&METAVAR
SCALE 1.0
COLVAR 2
&END METAVAR
&METAVAR
SCALE 1.0
COLVAR 3
&END METAVAR
&METAVAR
SCALE 1.0
COLVAR 4
&END METAVAR
&PRINT
&COLVAR
&EACH
MD 1
&END EACH
COMMON_ITERATION_LEVELS 100000
&END COLVAR
&END PRINT
&END METADYN
&END FREE_ENERGY
&END MOTION
&FORCE_EVAL
STRESS_TENSOR ANALYTICAL
METHOD QS # use Quickstep - electronic structure module of CP2K
&DFT
BASIS_SET_FILE_NAME ${DATA_PATH}/GTH_BASIS_SETS
POTENTIAL_FILE_NAME ${DATA_PATH}/GTH_POTENTIALS
&QS
ALMO_SCF F
EPS_DEFAULT 1.0E-13 # overall accuracy of the Kohn-Sham matrix build
EXTRAPOLATION USE_GUESS
&END QS
&MGRID
CUTOFF 400
NGRIDS 4
&END MGRID
&XC
&XC_FUNCTIONAL BLYP
&END XC_FUNCTIONAL
&END XC
&SCF
ADDED_MOS 200
EPS_SCF 1.0E-7
SCF_GUESS ATOMIC
MAX_SCF 500
&MIXING
METHOD PULAY_MIXING #DIRECT_P_MIXING
ALPHA 0.30
&END
&PRINT
&RESTART
BACKUP_COPIES 1
&EACH
QS_SCF 1
JUST_ENERGY 1
&END EACH
&END RESTART
&END PRINT
&END SCF
&KPOINTS
SCHEME MONKHORST-PACK 4 6 4
SYMMETRY FALSE
EPS_GEO 1.e-8
FULL_GRID TRUE
VERBOSE TRUE
PARALLEL_GROUP_SIZE 4
&END KPOINTS
&END DFT
&SUBSYS
&CELL
ABC 9.002 4.501 6.04
MULTIPLE_UNIT_CELL 1 1 1
&CELL_REF
ABC 10.8024 5.4012 7.248
&END CELL_REF
&END CELL
&TOPOLOGY
MULTIPLE_UNIT_CELL 1 1 1
&END
&COORD
SCALED T
Mg 0.75 0.0 0.75
H 0.902 0.304 0.75
H 0.598 0.696 0.75
Mg 0.0 0.5 0.0
H 0.152 0.196 0.0
H 0.348 0.804 0.0
Mg 0.25 0.0 0.75
H 0.402 0.304 0.75
H 0.098 0.696 0.75
Mg 0.5 0.5 0.0
H 0.652 0.196 0.0
H 0.848 0.804 0.0
Mg 0.75 0.0 0.25
H 0.902 0.304 0.25
H 0.598 0.696 0.25
Mg 0.0 0.5 0.5
H 0.152 0.196 0.5
H 0.348 0.804 0.5
Mg 0.25 0.0 0.25
H 0.402 0.304 0.25
H 0.098 0.696 0.25
Mg 0.5 0.5 0.5
H 0.652 0.196 0.5
H 0.848 0.804 0.5
&END COORD
&KIND Mg
BASIS_SET TZV2P-GTH
POTENTIAL GTH-BLYP-q2
&END KIND
&KIND H
BASIS_SET TZV2P-GTH
POTENTIAL GTH-BLYP-q1
&END KIND
&COLVAR
&QPARM
ATOMS_FROM 22
ATOMS_TO 2 3 5 6 8 9 11 12 14 15 17 18 20 21 23 24
L 4
RCUT 2.7
ALPHA 6.0
&END QPARM
&END COLVAR
&COLVAR
&QPARM
ATOMS_FROM 10
ATOMS_TO 2 3 5 6 8 9 11 12 14 15 17 18 20 21 23 24
L 6
RCUT 2.7
ALPHA 6.0
&END QPARM
&END COLVAR
&COLVAR
&QPARM
ATOMS_FROM 22
ATOMS_TO 1 4 7 10 13 16 19
L 4
RCUT 4.1
ALPHA 7.0
&END QPARM
&END COLVAR
&COLVAR
&QPARM
ATOMS_FROM 22
ATOMS_TO 1 4 7 10 13 16 19
L 6
RCUT 4.1
ALPHA 7.0
&END QPARM
&END COLVAR
&END SUBSYS
&END FORCE_EVAL
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://lists.cp2k.org/archives/cp2k-user/attachments/20160811/8f1ad219/attachment.htm>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: MgH2-consQ.png
Type: image/png
Size: 3499 bytes
Desc: not available
URL: <https://lists.cp2k.org/archives/cp2k-user/attachments/20160811/8f1ad219/attachment.png>
More information about the CP2K-user
mailing list