[QE-developers] Bug in Modules/read_pseudo.f90 (Quantum ESPRESSO v.6.8)

[Paolo Giannozzi]

You are very right. As a quick fix, just remove the protection against
zero-size allocations: I don't think it serves any purpose any longer

Paolo

On Fri, Sep 24, 2021 at 10:16 AM Lothar Brendel <lothar.brendel at uni-due.de>
wrote:

> Dear espressoers,
>
> I'm pretty sure to have found an inconsistency regarding the treatment
> of the case of zero wavefunctions in a PP file (nwfc = 0):
>
> In upflib/read_upf_v1.f90:subroutine read_pseudo_pswfc(), the allocation
> of the chi array is protected against the zero case:
>
> ALLOCATE( upf%chi( upf%mesh, MAX( upf%nwfc, 1 ) ) )
>
> But for the MPI slave ranks in Modules/read_pseudo.f90:SUBROUTINE
> upf_bcast(), it is *not*:
>
> ALLOCATE( upf%chi(upf%mesh,upf%nwfc) )
>
>
> Hence, when subsequently broadcasting the array, the slave ranks are
> receiving the data into a zero sized array.
>
>
> Kind regards
>          Lothar Brendel
>
> _______________________________________________
> developers mailing list
> developers at lists.quantum-espresso.org
> https://lists.quantum-espresso.org/mailman/listinfo/developers
>


-- 
Paolo Giannozzi, Dip. Scienze Matematiche Informatiche e Fisiche,
Univ. Udine, via delle Scienze 206, 33100 Udine, Italy
Phone +39-0432-558216, fax +39-0432-558222
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.quantum-espresso.org/pipermail/developers/attachments/20210924/6d589a10/attachment.html>