On Sat, Jan 23, 2010, xueqin pang wrote:
>
> I am running Nmode on a bigger system. And the minimization takes
> extremely long (about 50 steps per cpu hour ). So is there any way
> to run it in parallel form?
No. But see the following paper for an alternative:
%A R.A. Brown
%A D.A. Case
%T Second derivatives in generalized Born theory
%J J. Comput. Chem.
%V 27
%P 1662-1675
%D 2006
See especially Table 10. Also, if you are just doing minimization, you can
use the xmin minimizer in NAB. You don't say which minimizer you are using
in nmode: if it is Newton Raphson, it should take far fewer than 50 steps to
converge. Maybe you don't have a good enough starting structure(?)
...dac
_______________________________________________
AMBER mailing list
AMBER.ambermd.org
http://lists.ambermd.org/mailman/listinfo/amber
Received on Sat Jan 23 2010 - 12:00:03 PST
