On Tue, Mar 06, 2007, Ilyas Yildirim wrote:
>
> This problem happens when multisander is used in a minimization
> (when Thermodynamic Integration Approach is followed).
This is certainly a possible limitation. The way sander handles
parallelization in minimization is completely different than what is used for
molecular dynamics, and it could well be that minimization is limited to
2 threads (i.e. -np 2) for TI.
I'm not quite sure why you are doing minimization with TI anyway, but I
suggest sticking with what works and using just 2 threads.
good luck...dac
Received on Wed Mar 07 2007 - 06:07:47 PST
