[CP2K-user] [CP2K:21165] Re: metadynamics energy convergence

Quentin Pessemesse q.pessemesse at gmail.com
Mon Feb 17 09:22:43 UTC 2025
Previous message (by thread): [CP2K-user] [CP2K:21162] Re: metadynamics energy convergence
Next message (by thread): [CP2K-user] [CP2K:21150] Grimme D3 Correction to CO2 System @ 1200 K
Messages sorted by: [ date ] [ thread ] [ subject ] [ author ]
 In my opinion with small hills deposited every 100 fs, you’re unlikely to 
converge the Metadynamics bias. 
You should have a look at J. Phys. Chem. B 2005, 109, 14, 6714–6721 
https://doi.org/10.1021/jp045424k, on choosing the optimal parameters for 
ab initio metadynamics. However, even with increased bias deposition rate, 
it is possible that meta dynamics bias convergence could take more than 
hundreds or thousands of ps depending on the system, although it’s hard to 
say. Maybe you can have a qualitative intepretation if you are close to 
convergence and do not require accurate barriers, and rather need orders of 
magnitude.
Take care :) 
Q. 


Le dimanche 16 février 2025 à 10:19:16 UTC+1, SAURABH SINGH a écrit :

> Thank you Quentin
>
> My system consists of simple molecules Cl CO2 and H2O
>
> &GLOBAL
>   PROJECT 1a
>   RUN_TYPE MD
>   PRINT_LEVEL MEDIUM
> &END GLOBAL
>
> #&EXT_RESTART
>  # RESTART_FILE_NAME 1a-1.restart
> #&END EXT_RESTART
>
> &MOTION
>   &MD
>     ENSEMBLE NVT # ENSEMBLE NVE
>     STEPS 50000
>     TIMESTEP 1
>     TEMPERATURE 300.0
>     TEMP_TOL   100
>     &THERMOSTAT
>       &NOSE
>         LENGTH 3 # 3 is the default
>         MTS 2 # 2 is the default
>         TIMECON 300.0
>         YOSHIDA 3 # 3 is the default
>       &END NOSE
>     &END THERMOSTAT
>   &END MD
>
>   &FREE_ENERGY
>     &METADYN
>       DO_HILLS
>       NT_HILLS 100
>       WW [kcalmol] 0.5
>       &METAVAR
>         SCALE 0.2
>         COLVAR 1
>         &WALL
>           TYPE QUADRATIC
>           POSITION [angstrom] 3.0
>           &QUADRATIC
>              DIRECTION WALL_PLUS
>              K [kcalmol] 100.0
>           &END
>         &END
>       &END METAVAR
>       &METAVAR
>         SCALE 0.2
>         COLVAR 2
>         &WALL
>           TYPE QUADRATIC
>           POSITION [angstrom] 3
>           &QUADRATIC
>              DIRECTION WALL_PLUS
>              K [kcalmol] 100.0
>           &END
>         &END
>       &END METAVAR
>
>       &PRINT
>         &COLVAR
>            COMMON_ITERATION_LEVELS 3
>            &EACH
>              MD 1
>            &END
>         &END
>         &HILLS
>            COMMON_ITERATION_LEVELS 3
>            &EACH
>              MD 1
>            &END
>         &END
>       &END
>
>     &END METADYN
>   &END
>
>   &PRINT
>     &RESTART_HISTORY
>       &EACH
>         MD 100
>       &END
>     &END RESTART_HISTORY
>     &VELOCITIES
>       &EACH
>         MD 1
>       &END
>     &END
>   # Normal restart file
>   &RESTART
>     &EACH
>       MD 1
>     &END
>   &END RESTART
>   &END PRINT
> &END MOTION
>
> &FORCE_EVAL
>   METHOD Quickstep              ! Electronic structure method (DFT,...)
>   &DFT
>     CHARGE = 0
>     MULTIPLICITY = 2
>     LSD
>     BASIS_SET_FILE_NAME  BASIS_MOLOPT.dat
>     POTENTIAL_FILE_NAME  GTH_POTENTIALS.dat
>     # if WFN restart is required
>     WFN_RESTART_FILE_NAME 1a.wfn
>     &MGRID
>      #CUTOFF 1000
>       CUTOFF 360
>       NGRIDS 4
>       REL_CUTOFF 60
>     &END MGRID
>     &QS
>       METHOD GPW
>       EPS_DEFAULT 1.0E-10 # 1.0E-10 is the default
>     &END QS
>    &SCF
>      SCF_GUESS RESTART
>      EPS_SCF 1.0E-5 # 1.0E-5 is the default
>      MAX_SCF 600 # 50 is the default
>      &OT
>      MINIMIZER CG
>      PRECONDITIONER FULL_ALL
>      LINESEARCH 3PNT
>      &END
>      &OUTER_SCF
>        EPS_SCF 1.0E-5 # 1.0E-5 is the default
>        MAX_SCF 20 # Keep > 100 when OT is on
>      &END
>    &END SCF  
>    &XC
>      &XC_FUNCTIONAL PBE
>      &END XC_FUNCTIONAL
>      &vdW_POTENTIAL
>        DISPERSION_FUNCTIONAL PAIR_POTENTIAL # POTENTIAL_TYPE is alias to 
> DISPERSION_FUNCTIONAL
>        &PAIR_POTENTIAL
>          TYPE DFTD3
>          PARAMETER_FILE_NAME dftd3.dat
>          REFERENCE_FUNCTIONAL PBE
>        &END PAIR_POTENTIAL
>     &END vdW_POTENTIAL
>    &END XC
>    &POISSON                    ! Solver requested for non periodic 
> calculations
>       PERIODIC NONE        ! Type of solver
>       POISSON_SOLVER MT
>    &END POISSON
>   &END DFT  
>   &SUBSYS
>     &CELL
>   A     20.0000000     0.00000000     0.00000000
>   B     0.00000000     20.0000000     0.00000000
>   C     0.00000000      0.00000000    20.00000000
>
>       PERIODIC NONE              ! Non periodic calculations. That's why 
> the POISSON section is needed
>     &END CELL
>     &TOPOLOGY                    ! Section used to center the atomic
>       COORD_FILE_FORMAT xyz
>       COORD_FILE_NAME  1a.xyz
>       &CENTER_COORDINATES
>       &END
>     &END
>     &KIND C
>       ELEMENT C
>       BASIS_SET DZVP-MOLOPT-SR-GTH-q4
>       POTENTIAL GTH-PBE-q4
>     &END KIND    
>     &KIND H
>       ELEMENT H
>       BASIS_SET DZVP-MOLOPT-SR-GTH-q1
>       POTENTIAL GTH-PBE-q1
>     &END KIND    
>     &KIND Cl
>       ELEMENT Cl
>       BASIS_SET DZVP-MOLOPT-SR-GTH-q7
>       POTENTIAL GTH-PBE-q7
>     &END KIND
>     &KIND O
>       ELEMENT O
>       BASIS_SET DZVP-MOLOPT-SR-GTH-q6
>       POTENTIAL GTH-PBE-q6
>     &END KIND
>
>     &COLVAR
>        &DISTANCE
>           ATOMS 5 3
>        &END DISTANCE
>     &END COLVAR
>     &COLVAR
>        &DISTANCE
>           ATOMS 7 3
>        &END DISTANCE
>     &END COLVAR
>   &END SUBSYS
> &END FORCE_EVAL
>
>
> I have attached my input file
>
>
> On Friday, February 14, 2025 at 3:34:42 PM UTC+5:30 Quentin Pessemesse 
> wrote:
>
>> Hello, 
>> I think you need to provide a bit more information such as the total time 
>> of your simulation, the type of system and maybe an input file. 
>> Metadynamics can be very long to converge, and it might be impossible to 
>> achieve convergence with ab initio depending on your CV for instance. If 
>> you observe crossing, it means that your CV is able to "push" the system 
>> over the barrier which is already a good sign. 
>> The plumed tutorials and google groups have information on how to 
>> understand how to see if a metadynamics simulation is converged, for 
>> instance:
>> https://groups.google.com/g/plumed-users/c/3R9CghiH8EE
>> https://www.plumed.org/doc-v2.9/user-doc/html/master-_i_s_d_d-2.html
>> There are other sampling methods that can help you achieve convergence of 
>> a free energy barrier if metadynamics does not work for your system.
>> Take care,
>> Quentin
>>
>> Le mercredi 12 février 2025 à 13:01:43 UTC+1, SAURABH SINGH a écrit :
>>
>>> Hii 
>>>
>>> I am facing a problem in metadynamics study
>>>
>>> I am using normal metadynamics.
>>> In my case I am getting the multiple recrossing, but my energy is not 
>>> converging 
>>> I am using wall height 1 kcal/mol and sigma as 0.3 
>>>
>>> Is it indicating issues with CV?
>>>
>>> I will be very grateful if you can give some valuable insights
>>>
>>> Thank you
>>
>>

-- 
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 visit https://groups.google.com/d/msgid/cp2k/11cfc1fd-76c4-41b3-85d5-f7113d2bfaf7n%40googlegroups.com.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://lists.cp2k.org/archives/cp2k-user/attachments/20250217/3613f753/attachment-0001.htm>
Previous message (by thread): [CP2K-user] [CP2K:21162] Re: metadynamics energy convergence
Next message (by thread): [CP2K-user] [CP2K:21150] Grimme D3 Correction to CO2 System @ 1200 K
Messages sorted by: [ date ] [ thread ] [ subject ] [ author ]
More information about the CP2K-user mailing list