Re: [Wien] Confusion about DOS

2021-11-27 Thread shamik chakrabarti
The material is a cubic spinel... On Sat, Nov 27, 2021, 22:56 shamik chakrabarti wrote: > Is it 2x2x1 > > On Sat, Nov 27, 2021, 21:45 shamik chakrabarti > wrote: > >> Dear Dr. Tran, >> >>How to set the number of k points to 4 as there are >> three variables nx, ny, nz? >> >>

Re: [Wien] Confusion about DOS

2021-11-27 Thread shamik chakrabarti
Is it 2x2x1 On Sat, Nov 27, 2021, 21:45 shamik chakrabarti wrote: > Dear Dr. Tran, > >How to set the number of k points to 4 as there are > three variables nx, ny, nz? > > On Sat, 27 Nov 2021 at 21:34, Tran, Fabien > wrote: > >> It would be a wrong procedure. There is no opt

Re: [Wien] Confusion about DOS

2021-11-27 Thread shamik chakrabarti
Dear Dr. Tran, How to set the number of k points to 4 as there are three variables nx, ny, nz? On Sat, 27 Nov 2021 at 21:34, Tran, Fabien wrote: > It would be a wrong procedure. There is no option -hf with lapw1. > "x hf -up" and "x hf -dn" after the two "x lapw1" are missing

Re: [Wien] Confusion about DOS

2021-11-27 Thread Tran, Fabien
It would be a wrong procedure. There is no option -hf with lapw1. "x hf -up" and "x hf -dn" after the two "x lapw1" are missing. Also, for a hybrid or Hartree-Fock calculation the scripts run(sp)_lapw do a few more things (like additional "x lapw2" and some "mv" and "cp") that are necessary. The

Re: [Wien] Confusion about DOS

2021-11-27 Thread shamik chakrabarti
Dear Prof. Blaha, I have 56 atoms supercell & it is converged with one k-point. In this regard should I do the following; (1) save_lapw case_1kpoint (2) run_kgenhf_lapw for 4 kpoints (3) x lapw1 -hf -up (4) x lapw1 -hf -dn (5) x lapw2 -hf -qtl -up (6) x lapw2 -hf -qtl -dn (7) x

Re: [Wien] Confusion about DOS

2021-11-27 Thread Peter Blaha
Is this a calculation with just one k-point ? Yes. It is a calculation with one k point Then everything is clear. Obviously, with one k-point in the BZ one cannot use the tetrahedra method. There are no bands (E as funktion of k), but just one eigenvalue "per band". In principle, your DO

Re: [Wien] Confusion about DOS

2021-11-27 Thread shamik chakrabarti
Dear Prof. Blaha, Answering your question as follows; On Sat, 27 Nov 2021 at 19:55, Peter Blaha wrote: > No. You have to "understand" your results. > Once you understand them, you can decide if you neglect something or not. > > Is this a calculation with just one k-point ? > Y

Re: [Wien] Confusion about DOS

2021-11-27 Thread Peter Blaha
No. You have to "understand" your results. Once you understand them, you can decide if you neglect something or not. Is this a calculation with just one k-point ? look into case.outputt These are data without broadening and you should see how many electrons (integral of DOS, 3rd column) are up

Re: [Wien] Confusion about DOS

2021-11-27 Thread shamik chakrabarti
Yes, the gap 2.875 eV is the value found from SCF. I have tried with broadening 0.001 & also with 0.000...however, it shows no improvement. Is it due to numerical noise & can be neglected? On Sat, 27 Nov 2021 at 18:30, Tran, Fabien wrote: > Is 2.875 eV the value of :GAP in case.scf? If yes, it l

Re: [Wien] Confusion about DOS

2021-11-27 Thread Tran, Fabien
Is 2.875 eV the value of :GAP in case.scf? If yes, it looks like if the Fermi energy is not placed correctly on the DOS, because the gap between the valence and conduction band seems to be close to 2.875 eV. Could it be due to the broadening in case.int? Try to execute tetra with a reduced (or no

Re: [Wien] compilation problem

2021-11-27 Thread Peter Blaha
I think you made everything ok. You are limited by the memory access (50GB/s with DDR4-3200 memory). The most recent Intel processors support DDR5 RAM, which boosts the memory access from 50 to 77 GB/s. They should be significantly faster under full load. Fast memory is more important than cl

Re: [Wien] qtl

2021-11-27 Thread Peter Blaha
And of course there must be a spin-orbit splitting into 5/2 and 7/2 (look at your core eigenvalues for l=1 or 2), but they should be completely occupied for one spin and thus the total L is zero. Am 27.11.2021 um 09:47 schrieb Fecher, Gerhard: when talking about the orbital momentum, you misse

Re: [Wien] qtl

2021-11-27 Thread Fecher, Gerhard
when talking about the orbital momentum, you missed the difference between capital L and lower case l If you would not find f-electrons (l=3) in Gd, then something would be completely wrong Ciao Gerhard DEEP THOUGHT in D. Adams; Hitchhikers Guide to the Galaxy: "I think the problem, to be quite