I am not experienced with such systems, but I have some comments that may help you: - convergency values are quite harsh (1e-8) try reducing it. - why not to use gaussian smearing as a start - why didn't specify value for ecutrho? - maybe starting with smaller k-point mesh then restart upon the resultant relaxed structure can help
take care as 'O' can be problematic sometimes Karim KTH SE
_______________________________________________ Pw_forum mailing list [email protected] http://pwscf.org/mailman/listinfo/pw_forum
