The ABINIT Users Mailing List ( CLOSED )

["Anglade Pierre-Matthieu" <anglade@gmail.com>]

Hi,

For a long time now  there has been one or two failing GW tests when
using gfortran. You may have find the reasons. Thanks.

PMA

On 8/31/07, Takeshi Nishimatsu <takeshi@physics.rutgers.edu> wrote:
> Hi,
>
> I encountered this bug in another program.
>    http://gcc.gnu.org/bugzilla/show_bug.cgi?id=32962
>                  (This bug is fixed on trunk of gcc.)
> So I did grep in ABINIT src.
>
> $ pwd ; grep -i 'conjg(transpose' */*.F90
> abinit-5.3.5/src
> 15gw/setup_hamilt.F90:   
> h(:,:,ik,is)=matmul(conjg(transpose(c(:,:,ik,is))),ctmp(:,:))
> 15gw/setup_hamilt.F90:  htotal(:,:,is)=0.5 * ( htotal(:,:,is) + 
> conjg(transpose(htotal(:,:,is))) )
>
> $ pwd ; grep -i 'conjg(transpose' */*.F90
> abinit-5.4.2/src
> 15gw/csigme.F90:       call 
> calc_sig_ppm(sp%npwc,nomega,rhotwgp(1:sp%npwc),conjg(transpose(botsq(:,:))),&
> 15gw/setup_hamilt.F90:   h(:,:,ik,is)= 
> matmul(conjg(transpose(c(:,:,ik,is))),ctmp(:,:))
> 15gw/setup_hamilt.F90:  htotal(:,:,is)= 0.5*( htotal(:,:,is) + 
> conjg(transpose(htotal(:,:,is))) )
> 15gw/symf12.F90:  ! herm(:,:,io) = 
> half*(eps(:,:,io)+conjg(transpose(eps(:,:,io))))
> 15gw/symf12.F90:  ! aherm(:,:,io) = 
> -(0.,1)*half*(eps(:,:,io)-conjg(transpose(eps(:,:,io))))
>
> What I did is only grep.
> So it might be false alarm for ABINIT. Sorry,
> --
> Takeshi Nishimatsu
>
>
>


-- 
Pierre-Matthieu Anglade