[Pw_forum] subroutine: star_q

[Katalin Gaal-Nagy]

Dear Paolo,

thank you for your reply!

>> I try to calculate for a given system the q points which are equivalent
>> to a given one. For this purpose I try to use the routine star_q given
>> in the ph directory of the quantum espresso package [...]
> the star_q routine is actually quite complicated. There are
> several other places where this is done in the code; one
> should be in PW/kpoint_grid.f90, another pwtools/kpoints.f

I had the hope to get the equivalent q points in tha same ordering as in, 
e.q., the files with the dynamical matrices. That was the reason to take 
the routine star_q, since the ordering and, sometimes, also the sign of 
the componets is different for q sets calculated with pwtools/kpoints.f 
and PW/kpoint_grid.f90. Anyway.

>> The parameter noinv sets the inversion symmetry of the system
> well, not exactly. "noinv" when .true. removes the effects of the
> symmetry sending z into -z. This was an input variable used for
> special purposes in cases (i.e. surfaces) but it is never read in
> the recent versions. You can safely set it to .false.

Bene!
I will do it.

>> The parameter modenum is the number of modes to be done
> no, it is the index of a specific phonon mode, used for
> calculations of a single mode. Just set it to 0.

I did it. Doing a small test it turned out that I get the same q points 
like in some dynmat-file, but with a different ordering. I am a little bit 
wondering, since the procedure is the same as in the ph.x program. But as 
long as it works for my (bct) system, I am fine. As soon as I have time 
I will do some checks also for systems with other lattices, and as soon as 
I have everything in a clean, documented way, I will give it to the pwscf 
(or better quantum espresso) community.

Thank you for your help!

Best wishes,
Katalin