Hi All,
I am trying to do a frequency analysis on a dft-converged geometry of 20-atom Ag cluster . The lkain routine ran out of iterations without achieving convergence. The natural thinking will be to increase the number of iterations but the residuals ain't consitently decreasing (below is a snippet) . Is there a way to improve on the convergence or an alternative linear equation solver that could be invoked? I am running NWchem version 6.1.1 Thank you for your comments
Jacob
iter nsub residual time
---- ------ -------- ---------
34 600 2.57D-02 17073.6
35 600 5.54D-02 17467.1
36 600 1.18D-01 17860.3
37 600 9.39D-02 18252.4
38 600 5.15D-02 18648.6
39 600 3.80D-02 19040.9
40 600 2.66D-02 19433.5
41 600 3.12D-02 19826.9
42 600 3.50D-02 20220.5
43 600 3.77D-02 20612.9
44 600 2.76D-02 21006.5
45 600 5.39D-02 21399.9
46 600 4.49D-02 21795.3
47 600 5.22D-02 22190.1
48 600 1.38D-01 22585.7
49 600 4.62D-02 22979.6
50 600 4.52D-02 23374.0
|