The ionic coordinates in XDATCAR is wrapped or unwrapped?

[suen]

Dear all,
I find the ionic coordinates in XDATCAR file of VASP.5.4.4 may be unwrapped, but i am not very sure.
There is no more information about it in the VASP Manual or in this forum.
it is very important for me because i need to calculate the Mean Square Displacement.
Please help me if anyone know it.
Many thanks!

[suen]

This is my INCAR File:

SYSTEM = VIS
NWRITE = 0
LCHARG = .FALSE.
PREC = Normal
EDIFF = 1E-5
ENCUT = 500
LREAL= Auto
ALGO = Normal
ISMEAR = 0
SIGMA = 0.05
NELMIN = 4
ISIF = 2
ISYM = 0 
IBRION = 0
POTIM = 1.0 
NSW = 20000
NBLOCK = 1
MDALGO = 2 
SMASS = 2
TEBEG = 6000
TEEND = 6000

[chrstphr]

Hi,

my take would be to just compute the difference vector dv of all ionic displacements in a way such that they are consistent with periodic boundary conditions, i.e. something like

Code: Select all

dv = mod(dv+100.5,1.0)-0.5

in FORTRAN.

[suen]

Hi,

my take would be to just compute the difference vector dv of all ionic displacements in a way such that they are consistent with periodic boundary conditions, i.e. something like

Code: Select all

dv = mod(dv+100.5,1.0)-0.5

in FORTRAN.

Thank you for your reply.
I am confused about the "MDALGO".

My system is calculated at high temperature (6000 K) in vasp.5.4.4:
When I set MDALGO = 0 (default), all coordinates <1, it shows that all coordinates are wrapped in a cell using the PBC.
However, when I set MDALGO = 2, some coordinates > 1 (or < 0) in XDATCAR file (in direct, fractional coordinates), it shows that coordinates are unwrapped.

But when i use the same files for calculation in vasp.5.4.1, and set MDALGO = 2, then all coordinates <1 in XDATCAR.
So, i guess there are some changes in the new verson of vasp.5.4.4.
But i am not very sure.
I want to verify whether the coordinates in XDATCAR are transformed via PBC or not when i set MDALGO = 2 in vasp.5.4.4.