Dear ADF users:
I am running some geometry optimizations with ADF (large molecules, both
with high and low symmetry). I am using the default thresholds in energy
(tolE = 1e-3) and gradient (tolG = 1e-2). Surprisingly, very often there
is no clear convergence but an oscillating behavior. In such cases,
(assuming the oscillations are small) does it make sense to stop at a
convenient point, for example at a low energy value as probed by the
"E-test" ?
Thanks for any help,

--
Alessandro Bagno
Centro CNR Meccanismi Reazioni Organiche, Dip. Chimica Organica
Universita' di Padova, via Marzolo 1 - 35131 Padova (Italy)
tel. +39 0498275660; fax 0498275239
http://www.chor.unipd.it/~alex