[Wannier] pw2wannier crashes when calculating uHu

[Jonathan Yates]

Christoph,

 Can you take the latest version of pw2wannier90.f90 from GitHub, and see if that has the same problem. There was a issue with the use of ‘exclude_bands’ and uHu that we fixed recently.

 If that doesn’t fix the issue, can you get the input files to me (I will share with the other developers). I think that running this locally will be the only way to investigate.

 Jonathan

> On 30 Jan 2018, at 01:41, Christoph Wolf <wolf.christoph at qns.science> wrote:
> 
> Dear all, 
> 
> I recently tried computing orbital magnetization (and some other postw parameters) following example 19 of the wannier package; Works all great, however when I try to adapt this for other systems (e.g. SrVO3) pw2wannier90.x (from pwscf 6.0, 6.1 and 6.2, compiled with and without MKL, serial and parallel runs) crashes when attempting to calculate uHu.
> 
>   -----------------
>   *** Compute  uHu 
>   -----------------
> 
>  iknum =      512
>        1
> 
> Program received signal SIGSEGV: Segmentation fault - invalid memory reference.
> 
> Backtrace for this error:
> #0  0x7F15C242AE08
> #1  0x7F15C2429F90
> #2  0x7F15C1B5B4AF
> #3  0x7F15C24EC8F0
> #4  0x7F15C24EDFEF
> #5  0x7F15C24FCB87
> #6  0x40A2CB in compute_orb_ at pw2wannier90.f90:2658
> #7  0x42D3D0 in MAIN__ at pw2wannier90.f90:325
> Segmentation fault
> 
> 
> Since  example 19 works fine (with pseudos from pslibrary 1.0.0 in LDA/Normcons.) I suspect something else to be the problem (what is the secret ingredient?); I am a bit lost finding the problem so any hint is very much appreciated!
> 
> pseudos were generated as below:
> 
> <PP_INPUTFILE>
>  &input
>    title='Sr',
>    zed=38.,
>    rel=2,
>    config='[Kr] 5s2.0 5p0.0 4d0.0',
>    iswitch=3,
>    dft='LDA'
>  /
>  &inputp
>    pseudotype=2,
>    file_pseudopw='Sr-rel.LDA-n-nc.UPF',
>    author='TM',
>    lloc=0,
>    nlcc=.true.,
>    tm=.true.
>  /
> 3
> 5P  2  1  0.00  0.00  3.80  3.80  0.0
> 4D  3  2  0.00  0.00  3.60  3.60  0.0
> 5S  1  0  2.00  0.00  3.60  3.60  0.0
> 
> 
> Interpolated bands look great (compared to A. Scaramucci ArXiv:1405.3804v1) and wannierization has no problems...
> 
> TIA for any help!
> 
> Best,
> Chris 
> 
> PS: this also went out to the pwscf mailing list as I am not sure which one is more suitable for interfaces!
> 
> -- 
> Postdoctoral Researcher
> Center for Quantum Nanoscience, Institute for Basic Science
> Ewha Womans University, Seoul, South Korea
> <nscf.in><scf.in><pw2wan.in><VO.win>_______________________________________________
> Wannier mailing list
> Wannier at quantum-espresso.org
> http://mailman.qe-forge.org/cgi-bin/mailman/listinfo/wannier