rundos/kpoints

Message

phtobias · #1 Post by **phtobias** » Fri May 12, 2006 1:24 am

Dear all,

Does anyone has an experience using the rundos/kpoints utility?

I tried 'make kpoints' under a previously built vasp.4.6, but it turned out the makepoints.F is not there.

Also, is the description in the link of accurate dos calculation still valid? Or, it has been taken care of in vasp.4.6 by IBZKPT? Or, does there exist a better way to specify the (full star?) kpoints in a dos calculation?
http://cms.mpi.univie.ac.at/vasp/vasp/node211.html

Thanks in advance for the help.

--yi-ren

#2 Post by **admin** » Thu May 18, 2006 2:06 pm

1) rundos/kpoints : these features are obsolete for vasp.4.6, the respective code has been included in the program. The DOS is automatically calculated (and written to DOSCAR) for every vasp-run
-- to get an accurate DOS for the converged results,
a) set ICHARG=11 and do one more run from the self-consistent charge density with a denser k-grid spanning the full BZ (this is automatically done if ISYM = 0 in INCAR)
b) if necessary (narrow peaks), increase the energy grid for which the DOS is calculated (NEDOS)

phtobias · #3 Post by **phtobias** » Fri May 19, 2006 2:31 am

Thank a lot for the explanantion. That clears many questions about accurate DOS. --yi-ren

lahaye · #4 Post by **lahaye** » Thu May 25, 2006 10:35 am

This info appears to be also very useful for what I'm doing.
I tried it and got this in the info lines of vasp at the second run:

number of k-points has changed, file: 16 present: 901
trying to continue reading WAVECAR, but it might fail

I suppose I should then ignore that, since I have intentionally
increased the number of k-points for the second run.
Or have a done something wrong?

I'm trying to calculate graphene (single graphite layer) with
k-points only in X and Y direction. My first run KPOINTS
file was then:

k-points with Monckhorst-Pack
0
Monkhorst-Pack
11 11 1
0 0 0

For the second run, I changed the one but last line into
101 101 1
to increase the number of k-points.

Is that kind of right?

Thanks,
Rob.

#5 Post by **admin** » Thu Jun 01, 2006 2:03 pm

If the k-grids are changed, you must not continue from WAVECAR, on this file the wavefunction coefficients are given explicitely for each k-point, so the grid must not be changed between writing and reading. Continue from CHGCAR in such a case (setting ICHARG appropriately)

ziulfer · #6 Post by **ziulfer** » Fri Feb 03, 2012 5:40 pm

regarding rundos
this section of the manual is therefore useless:
http://cms.mpi.univie.ac.at/vasp/vasp/A ... tions.html