[QE-users] A question about SCF cycles and BFGS steps during relaxation calculation

[杜红波]

I do a relaxation about titanium oxide clusters with 22 atoms using QE. I found the number of SCF cycles in the .out file is changed all the time, but the number of BFGS steps is changed far slower than the number of SCF cycles during the relaxation process, in fact, the number of SCF cycles is unchanged after a while since it’s starting during the relaxation process. The energy of new cycle is larger than the old one. And it can’t converge for 5 hours. Could someone tell me something about this problem? Thank you for your help!

 

 

     Total force =     0.012964     Total SCF correction =     0.000239

 

     number of scf cycles    = 358

     number of bfgs steps    =  58

 

     energy   old            =   -1662.9726523325 Ry

     energy   new            =   -1662.9726263315 Ry

 

     CASE: energy  _new > energy  _old

 

     new trust radius        =       0.0011710614 bohr

     new conv_thr            =       0.0000000100 Ry

 

 
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.quantum-espresso.org/pipermail/users/attachments/20181010/b76be5f4/attachment.html>