No, if I remember correctly this happens at the point where siesta "moves" the atoms. A few things can occur, a) the user wants to extrapolate the DM from different geometries b) the DM is reused, but the overlap matrix changes
In both cases will the calculated charge be different from the initial geometry. Yet, siesta will normalize this subsequently and it should not be an issue. 2015-04-16 17:46 GMT+02:00 Marcos Veríssimo Alves < [email protected]>: > Hi all, > > I have been performing some calculations for graphene-like materials using > the trunk-462 version, and I have been obtaining some warnings at the > beginning of some SCF cycles: > > Note: For starting DM, Qtot, Tr[D*S] = 10.00000000 > 10.09624902 > > These warnings disappear as the structure is closer to being relaxed. Is > it cause to worry about some parameter in my calculation? > > Best regards, > > Marcos > -- Kind regards Nick
