[Pw_forum] Restarting calculations

stewart at cnf.cornell.edu stewart at cnf.cornell.edu
Fri Oct 28 18:30:52 CEST 2005


Hi all, 

   I would like to restart pw calculations that haven't converged.  
Unfortunately, I run into efermit errors when I try to restart the 
calculations.  I have encountered this before when I am taking a converged 
calculations and doing some processing on it to get band, dos, etc with 
non-scf runs.  In this case, I believe I removed the information on smearing 
and tetrahedra integration (found in a suggestion in a previous post) and 
everything ran.  Does anyone have suggestions for getting restart 
calculations to work when you are still seeking convergence.  I am looking 
at spin polarized calculations.  Could the spin polarization lead to 
problems?  In past posts, there seem to be some suggestions that information 
is lost that is required to restart spin calculations. 

Thanks, 

Derek 

#####################################
Derek Stewart, Ph. D.
Scientific Computation Associate
250 Duffield Hall
Cornell Nanoscale Facility (CNF)
Ithaca, NY 14853
stewart (at) cnf.cornell.edu
(607) 255-2856 




More information about the users mailing list