VASP with BoltzTraP

Message

kldas-05 · #16 Post by **kldas-05** » Sat Jan 11, 2014 5:19 am

Hello kzhuo and bioggg,

I already found the error cause. The problem was in the case.energy file and it is generated by the presence of some Kpoint repeated. So, this kpoints repeated and its energies must be deleted of the case.energy file.

kldas-05 · #17 Post by **kldas-05** » Sat Jan 11, 2014 5:19 am

Hello kzhuo and bioggg,

I already found the error cause. The problem was in the case.energy file and it is generated by the presence of some Kpoint repeated. So, this kpoints repeated and its energies must be deleted of the case.energy file.

bioggg · #18 Post by **bioggg** » Mon Jan 13, 2014 12:33 am

Hello kldas-05,

I deeply appreciate your effort and sharing. However, I did not find any repeated Kpoints in my case.energy(so) file. I found those kpoints are exactly the same as the data in IBZKPT file. Does BoltzTraP have any limitation on the number of kpoints? In my case, I have 1000 kpoints.

I am wondering whether you are able to describe how you debugged, and I may be able to work on it. For now, I have no idea how to solve this problem.

Thank you so much.

kldas-05 · #19 Post by **kldas-05** » Sat Jan 18, 2014 10:08 pm

Hello bioggg,
When you perform the bands calculation in VASP, the strings of k-points Have the form: Gamma-X, XW .....
In this case, X is the kpoint repeated and this kpoint is repeated in the case.energy too.

bioggg · #20 Post by **bioggg** » Tue Jan 21, 2014 8:11 am

Hello Kldas-05,

I am wondering why you used the data from band structure calculation rather than the ones from self-consistence. I had this question in my mind for a while, but I never got an answer.

However, even though I plugged in the band structure data and removed repeated kpoints, the error message, STOP error in factorization, was still there.

Thank you so much for your kind help.

bioggg · #21 Post by **bioggg** » Tue Jan 21, 2014 7:11 pm

The eigenvalue gotten from band structure calculation worked. However, I still don't understand why you did not use self-consistent result.

kzhuo3 · #22 Post by **kzhuo3** » Sat Feb 22, 2014 11:05 pm

I think you should be using a standard KPOINTS file so that you sample the entire Brillouin zone evenly.e.g.

comments
0
Monkhorst Pack
16 16 16
0 0 0

instead of the line input method for band structure plots along specific directions only.

That should naturally get rid of your repeated k-points problem too.

szpunarb · #23 Post by **szpunarb** » Tue Apr 08, 2014 8:36 pm

I also get this factorization error and tried to simplify bands by using supercell but still not working. Has your problem been resolved and how? Thank you.

szpunarb · #24 Post by **szpunarb** » Thu Apr 10, 2014 8:28 pm

In my case i finally fixed by neglecting small magnetic moment and using non-spin polarized calculations for supercell. I used Monkhorst Pack large number of k points as discussed above, nr bands was set to the number of electrons in scf calculations and lpfac eq. 10.