There are 3 points I am not quite clear in using NEB calculation.
I would like to make sure whether partition number I used in parallel computing means the image number in NEB method?
As the manual said, we can run on more virtual processors than the physical processors. How to achieve it? For example, the physical processors in my computer is 4, but I need 8 partition/images for NEB calculation. The executive command might be something like this: " mpirun -np 8 lmp_g++ -partition 8x1 -in inputfile ". I am not sure about it.
In neb command, we should have a filename to specify the final atom coordinates. It said that only the atoms with the diffident coordinates with the initial configuration should be included in this file. I wonder it will still work if I write all the atom coordinations in this file with distinguishing which atom coordination is change in comparison with the initial configuration?
There are 3 points I am not quite clear in using NEB calculation.
1) I would like to make sure whether partition number I used in parallel
computing means the image number in NEB method?
yes
2) As the manual said, we can run on more virtual processors than
the physical processors. How to achieve it? For example, the physical
processors in my computer is 4, but I need 8 partition/images for NEB
calculation. The executive command might be something like this: " mpirun
-np 8 lmp_g++ -partition 8x1 -in inputfile ". I am not sure about it.
yes
3) In neb command, we should have a filename to specify the final atom
coordinates. It said that only the atoms with the diffident coordinates with
the initial configuration should be included in this file. I wonder it will
still work if I write all the atom coordinations in this file with
distinguishing which atom coordination is change in comparison with the
initial configuration?
You can include atom coords that don't change if you like. It won't
make any difference.