The ABINIT Users Mailing List ( CLOSED )

["Matteo Giantomassi" <Matteo.Giantomassi@uclouvain.be>]

> Dear All,
> I try to use closure trick for empty states in GW calculations, but in
> several test calculations, i always came to WARNING below, and it seems
> that
> i cannot just tune "gwencomp" to avoid this WARNING.
> Would you please do me a favor for this problem ?

 The warning is not due to the value used for gwencomp.
 It's related to the fact that the closure trick requires
 the Fourier components of the product u^*(r) u(r) at G1-G2,
 where G1 and G2 are two plane waves used to describe the
 irreducible polarizability.

 Now, it might happen that G1-G2 falls outside the FFT box, in this
 case the contribution given by this term is neglected.
 This should not affect the accuracy of the results
 provided that the number of G1-G2 pairs is relatively small (as in your
case).

 To get rid of the warning there are several possibilities.
 Either you increase the value of ecutwfn  or you decrease ecuteps.
 The former is the preferred choice, usually I use all the G-vectors
 stored in the KSS file to avoid possible inaccuracies due to
 the truncation of the plane wave representation (the CPU effort
 scales almost linearly wrt npwwfn)


 Hope it helps
 Regards
 Matteo Giantomassi



>
>  completechi0_deltapart : WARNING  Number of G1-G2 pairs outside the
> G-sphere for Wfns =    72
>
>  completechi0_deltapart : WARNING  Number of G1-G2 pairs outside the
> G-sphere for Wfns =    72