[Pw_forum] Conjugate Gradient minimization not converged after 50 iterations

Paolo Giannozzi giannozz at democritos.it
Mon May 4 10:02:38 CEST 2009


Jonas Baltrusaitis wrote:

> conjugate gradient minimization does not converge after 50 iterations in 
some modes.
 > By inspection, some modes converge in 46-48 iterations. Would 
increasing number of
>  iterations solve that problem?

likely, yes

> Is it a problem at all as residual after 50 iterations is quite small

maybe it isn't a problem at all. It simply means that the prefixed
threshold on the residual vector has not been reached. How this
translates into errors in the final results, it is hard to say.
One should go to the definition of the threshold and try to
figure out an estimate for the corresponding error on the
frequeencies; or else, one should try a test casse with
different thresholds

Paolo
-- 
Paolo Giannozzi, Democritos and University of Udine, Italy


More information about the Pw_forum mailing list