On Fri, 2012-11-02 at 17:07 -0500, Kyeong-hyun Park wrote:

> I am looking for the way of finding a proper tr2_ph value. Is there any 
> convergence test for that? My simulation would run more than 50 hrs, so I 
> don't want to just give it a shot with randomly picked value.

unfortunately there is no clear criterion right now. 1.d-14
should be fine, but sometimes calculations using ultrasoft 
pseudpootential may need tighter thresholds

P.
-- 
Paolo Giannozzi, IOM-Democritos and University of Udine, Italy


Reply via email to