[Wannier] VASP doesn't write the *amn files

[Dita Puspita]

As long as my experience, I think you need to prepare the .win file and prepare the projections. There is an explanation about the projections in the user manual.

Regards,
Dita Puspita

> On Sep 6, 2016, at 2:43 AM, abhishek khetan <askhetan at gmail.com> wrote:
> 
> Dear all,
> 
> I compiled vasp with the Wannier90 v1.2. interface and use the following commands in my INCAR to write all the files:
> LWANNIER90 = .TRUE.
> LWRITE_UNK = .TRUE.
> LWRITE_MMN_AMN=.TRUE.
> 
> However, vasp writes only the following files:
> UNK00001.1
> UNK00001.2
> UNK00002.1
> UNK00002.2
> UNK00003.1
> UNK00003.2
> ..................
> ..................
> ..................
> wannier90.dn.eig
> wannier90.dn.mmn
> wannier90.up.eig
> wannier90.up.mmn
> wannier90.win
> wannier90.wout
> 
> As you can see, the files wannier90.up.amn and wannier90.dn.amn have not been written at all. When I try to run the wannier executable using the command (of course after renaming wannier90.win as wannier90.up.win):
> $ ./wannier90.x wannier90.up.win
> 
> I get the error:
> Reading overlaps from wannier90.up.mmn : File generated by VASP: unknown system
> Exiting.......
> Error: Problem opening input file wannier90.up.amn
> 
> Could someone please suggest what could be going wrong ? and why are the wannier90.up.amn and wannier90.dn.amn files are not being written ? This is related to num bands =! num_wann. When I use num wann = num_bands, then it doesnt ask for *amn files because there is no need for disentanglement. But what If I want to disentangle? I did not get any erros while the vasp code was running and writing the wannier files.
> 
> Is there a way to write the *amn files afterwards (i.e. not using the library mode). I am doing very heavy and intensive GW calculations which require a lot of memory and time, so its hard for me to repeat the calculations. Is there a way i can get the *amn files just from the final WAVECAR/CHG/CHGCAR files ?
> 
> My vasp is intel mkl compiled. COuld it have something to do witht he choice of FCL
> FCL        = mpif90 -mkl=parallel
> 
> is it supposed to be -mkl=sequential or -mkl=cluster instead of -mkl=parallel ?
> 
> Any help is highly appreciated.
> 
> 
> -- 
> MfG,
> abhishek
> _______________________________________________
> Wannier mailing list
> Wannier at quantum-espresso.org
> http://mailman.qe-forge.org/cgi-bin/mailman/listinfo/wannier