[Pw_forum] PWscf can not find out symmetry operations!!!

[Yingli Niu]

Respected Prof. Baroni
    Sorry for that I didn't browsed in the mailing list's archive carefully!
    I think you didn't understand me due to my english language problems.
1. I did a relax calculation for a slab with 10 layers, 20 C atoms. Than I
got 20 relaxed positions.
2. I did a scf calculation using the 20 positions. And pw.x found "No
symmetry".
3. I inputed the structure into Material Studio, and it find 8 symmetry
operations and 7 inequivalent atoms. The tolerance is 0.01A.
4. I used 8 symmetry operations to act on the 7 positions, then got 56
positions(some of them are superpositions).
5  I selected 20 positions in the primary unit cell(there were only 20
positions in one primary unit cell) and input them into the PWscf's run_scf
script. Then pw.x find 8 symmetry operations.
    This method is equivalent to modify the atom positions manually in order
to make the pw.x to find symmetry operations!
    The manual say that :

>the number of significant figures in the atomic positions is not large
enough. In file
>PW/eqvect.f90, the variable accep is used to decide whether a rotation is a
symmetry operation.
>Its current value (10-5 ) is quite strict: a rotated atom must coincide
with another atom to
>5 significant digits. You may change the value of accep and recompile.

    In the last email I said that I had changed the value of accep and
recompiled. But it made no use to find any symmetry operations. So I have no
method other than to modify the atom positions by symmetry manually.
  Now I think the problem has been solved!
  Many thanks again and forgive me to bring so many troubles!
Best wishes!
Yingli Niu


On Aug 31, 2006, at 04:01:40 +0200, Prof. Baroni wrote:

>> Dear Pro. Baroni

>>* Thanks a lot for your help!
*>>*   I have said that It was no use to set accep=3D1.0D-2
*>>* and recompiled pw.x. The program still could not find
*>>* any symmetry operation.
*>>*   Then I selected 7 inequivalent atoms, then used 8
*>>* symmetry operations:
*>>*   x,y,z
*>>*   -x,-y,z
*>>*   -x+1/2,y+1/2,-z
*>>*   x+1/2,-y+1/2,-z
*>>*   -x+1/2,-y+1/2,-z
*>>*   x+1/2,y+1/2,-z
*>>*   x,-y,z
*>>*   -x,y,z
*>>* to transform the 7 inequivalent atoms to 20 atoms in
*>>* the primary unit cell.
*>
>I can hardly understand what you did. If by applying some symmetry =20
>operations to a bunch of positions ("atoms") you obtain a larger =20
>number of atoms, doesn't this mean that those operations are not =20
>"symmetrry operations"???
>>*   Then I ran scf. I use Norm-Conserving PP and chose
*>>* ecutwfc=3D70,ecutrho=3D600.
*>
>there is ABSOLUTELY no point in using ecutrho>ecutwfc with norm =20
>conserving PP's
>
>>*   There appeared message "warning:symmetry operation #
*>>*  3/4/5/6 not allowed  fractional translation:" in the
*>>* out file as discussed before. Only 4 symmetry
*>>* operations without inversion could be found!
*>>* I set ecutrho =3D 500 and calculated it again. Then
*>>* all 8 symmetry operations was found!
*>>*   The stress tensor did not change too much!
*>>*   stress tensor(kbar) from original atom positions:
*>>*     49.20      0.01      0.07
*>>*      0.01    -16.72      0.16
*>>*      0.07      0.16      2.61
*>>*   stress tensor(kbar) from changed atom positions:
*>>*     49.98      0.00      0.00
*>>*      0.00    -16.13      0.00
*>>*      0.00      0.00      7.30
*>>*   All the nondiagonal elements changed to 0.00.
*>>*   I think changed accpe from 1.0D-5 to 1.0D-2 will be
*>>* harmful for the calculations. So I reset it to 1.0D-5
*>>* and recompiled.
*>>*   I have no other method to make the program find out
*>>* all 8 symmetry operations than this method.
*
>Have you browsed in the mailing list's archive, as I suggested? =20
>Please, let me NOT rediscuss this point that I and others discussed =20
>the last time a few days ago
>
>>*
*>>*   Another question:
*>>*   I use Yahoo mail box. How can I post the reply
*>>* message next to the message which is replied in the
*>>* mailing list. I alway started a new discussion line
*>>* with reply message! Sorry for that!
*>>
>Please, ask Terry Semel. He certainly knows better than me.
>
>SB
>
>---
>Stefano Baroni - SISSA  &  DEMOCRITOS National Simulation Center - =20
>Trieste
>[+39] 040 3787 406 (tel) -528 (fax) / stefanobaroni (skype)
-------------- next part --------------
An HTML attachment was scrubbed...
URL: /pipermail/attachments/20060901/6b1d6746/attachment.htm