[Pw_forum] different lattice representation

[Jess Kondor]

 Hello, everyone

 I have a question regarding how to set up lattice in pwscf.  Usually I like
present lattice in the following format ( aka vasp format)

lattice:
 a0  (in angstrom)

 a11 a12  a13
 a21 a22  a23
 a31 a32  a33

and atomic coordinates in the fractional (crystal) format. So, I guessed
that in pwscf it should be something like that:

ibrav       =  0
celldm(1) = 1.88973 (corresponds 1 Ang)

CELL_PARAMETERS cubic
 a11 a12  a13
 a21 a22  a23
 a31 a32  a33
ATOMIC_POSITIONS {crystal}
.....
Using this representation for my system (slab) I didn't get convergence at
all, and after a long careful examination, I found that this is due to wrong
geometry representation. To confirm this fact, I took the simple system from
example25 in espresso distribution and changed lattice like:

from
    ibrav=  0, celldm(1)=8.19
  CELL_PARAMETERS
  0.50 0.50 1.00
  0.50 1.00 0.50
 1.00 0.50 0.50
ATOMIC_POSITIONS {crystal}

to
   ibrav=  0, celldm(1)=1.00 ,
 CELL_PARAMETERS
  4.95 4.95 8.19
  4.95 8.19 4.95
  8.19 4.95 4.95


I thought these are equivalent  representation, but the code give me
different cell volumes and I obtained the different total energy, completely
different magnetic moments and long convergence in the second case.   So, I
suggest that this representation (case 2) is not correct. Am I right?

 Best regards,
  JK
-------------- next part --------------
An HTML attachment was scrubbed...
URL: /pipermail/attachments/20061212/2a9cc612/attachment.htm