symmetry warning in HSE jobs

Message

apple · #1 Post by **apple** » Wed Nov 17, 2010 12:13 am

hi all,
I was trying to run a single point HSE calculation with a Monkhorst-Pack 3x3x1 k-point grid. The job ran without any problems (i.e. no error messages).
Then, I was trying to repeat this calculation with the IBZKPT file copied to KPOINTS and no other changes made, which should give me the identical result one would think (right?), but got the following warning:
=====
Your generating k-point grid is not commensurate to the symmetry |
| of the lattice. This can cause slow convergence with respect |
| to k-points for HF type calculations |
| suggested SOLUTIONS: |
| ) if not already the case, use automatic k-point generation |
| ) shift your grid to Gamma (G) (e.g. required for hex or fcc lattice)
=======================
Can someone comment what could be the source of this warning.
I ran this calculation with ISYM=0 and then with ISYM=1, but the warning persists. Also, I tried a calc where I copied k-points from the OUTCAR file below the line: "Subroutine IBZKPT_HF returns following", but no success of getting rid of the warning yet...
I would appreciate if you could share your experience with me with running HSE jobs, in particular, how to avoid the symmetry warning mentioned above.

Thanks.

alex · #2 Post by **alex** » Wed Nov 17, 2010 8:07 am

Hi apple,

try to increase SYMPREC. You have to play a little with the value, because if it is too high, the program crashes, if it is too low, you end up like above.

Hth

alex

apple · #3 Post by **apple** » Wed Nov 17, 2010 3:48 pm

hi Alex,
thanks for replying to my post. I played with the SYMPREC flag changing it from 1E-3 to 1E-6, 1E-5, and 1E-4, but I still get the warning. It seems that coping the IBZKPT file to KPOINTS (which is equivalent to setting the k-points manually) does not work for some reason. If I use the automatically generated k-point file, I don't get the warning. Also, I just tried to grep SYMPREC from OUTCAR to check if vasp is reading it correctly, but it is not showing. Now I wonder if it is actually reading this flag. I tried with NWRITE=2. If you (or any other forum participant or the admin) have other suggestions on how to get rid of the symmetry warning, I would appreciate if you could share them with me.
Thanks.

alex · #4 Post by **alex** » Thu Nov 18, 2010 7:33 am

Hello apple,

have you also tried large values like 0.1, 0.01? There should be a change in message at least. You may need to adjust SYMPREC in even smaller steps once you have the range right (thinking of e.g. 0.02 0.04 0.06 or so). Are you sure about the Gamma centred k-point mesh?
Cheers,

alex