Page 1 of 1

EDIFF and EDIFFG

Posted: Tue Mar 08, 2011 12:17 pm
by hexiujie
Hi,everyone. The endpoint images are fully minimized using EDIFF=1.E-04 and EDIFFG=-0.02. But the Vasp TSTtools say “……these modifications to the force mean that the energy is no longer consistent with the force being optimized”. How can I set the two parameters for cNEB calculation?Many thanks.

Re: EDIFF and EDIFFG

Posted: Tue Mar 08, 2011 4:21 pm
by graeme
There is no problem using those settings for the cNEB. The issue is only with optimizers which rely on the force being consistent with the energy -- and specifically the built-in conjugate gradients optimizer in vasp. You can use any other optimizer, including these: http://theory.cm.utexas.edu/vtsttools/optimizers/