[Wien] Bug in lapw2 -almd ?

[Gavin Abo]

All the documentation I have seen only show or talk about the serial 
commands when using -alm and -almd:

https://github.com/jeandiehl/wannier-proj/blob/master/doc/wannier-proj_manual.pdf

https://github.com/TRIQS/triqs_0.x/blob/master/doc/user_manual/wien2k/TutorialDmftproj.pdf

The mpi for -qdmft seems to be for the DMFT python script case.py (not 
lapw2):

https://triqs.github.io/dft_tools/master/guide/dftdmft_selfcons.html?highlight=parallel

It sounds like you have confirmed what was previously thought [ 
https://www.mail-archive.com/wien@zeus.theochem.tuwien.ac.at/msg18447.html 
] that -almd is not valid with the parallel switch (-p).

On 4/25/2019 3:03 AM, Gary Amini wrote:
> Dear all,
>
> We are interested in obtaining the ALM, BLM and CLM coefficients.
> However, we have noticed that lapw2 -almd gives very different results 
> if we run the serial or parallel versions of Wien2k. The difference is 
> up to a factor of 10^10.
>
> Is this a bug, or is it not yet possible to run lapw2 -almd with the 
> parallel version?
>
> Best whishes,
> Gary