Dear Prof. Greame,
I am trying to use a force based conjugate gradient optimizer to find the ground state of a molecule (CH3OH) in a box. However, I see that the ions are not at all displacing. The positions of ions in CONTCAR are same as the positions in POSCAR. The same thing is observed after two consecutive runs of NSW=500.
I would be grateful if you could please take a the attached job folder and please let me know what I might be doing wrong.
Thank you so much.
Ions are not moving between POSCAR and CONTCAR
Moderator: moderators
Ions are not moving between POSCAR and CONTCAR
- Attachments
-
- CH3OH_molecule.zip
- (1.74 MiB) Downloaded 980 times
Re: Ions are not moving between POSCAR and CONTCAR
As I usually say, when you have problems with a second order optimizer such as IOPT/IBRION = {1,2} go with IBRION/IOPT = 3 and check for steady convergence. In the attached, you can see that the molecule optimization converged in two steps.
The problem was not that the molecule geometry wasn't moving with IOPT=2, it was that it was oscillating back and forth across the minimum. Reducing MAXMOVE and/or FDSTEP should fix the issue, but to everyone with similar problems, please, use a first-order conservative optimizer when you have convergence problems. Then, if things are slow, you can try to do better with a fancier optimizer.
The problem was not that the molecule geometry wasn't moving with IOPT=2, it was that it was oscillating back and forth across the minimum. Reducing MAXMOVE and/or FDSTEP should fix the issue, but to everyone with similar problems, please, use a first-order conservative optimizer when you have convergence problems. Then, if things are slow, you can try to do better with a fancier optimizer.
- Attachments
-
- CH3OH_molecule.tar.gz
- (338.57 KiB) Downloaded 972 times
Re: Ions are not moving between POSCAR and CONTCAR
Thank you Prof. Henkelman for your answer.