NEB

Please post to the list, not to me.\

But NEB is not getting initiated

What does that mean?

Steve

Steve,
NEB not getting initiated means, the simulation just hangs up and doesn’t proceed from the fix neb command. I have a minimize command before the fix neb; so the simulation just performs the minimization and hangs, without initiating the neb.

Thanks
Sankar

You'll have to post a (small) input script, data file
that has the problem.

Steve

Sankar and steve,

If you recall, I had the same situation before back to last December. I have not got this issue solved yet. I think there are some parameter settings that do not allow the program proceed further once the simulation system is relatively large (say thousands atoms). I guess because the NEB command generate multiple replicas within themselves and that may consume a lot of memory and could prohibit the program work. Also right now NEB is not parallel, if more replicas are produced in one processor that could be the problem. I may wait for new release of lammps that have parallel version of NEB. Hope this draws steve’s attention.

Best,

Ajing

Again, I need to see a script that reproduces the problem.
There should be no memory issues. There is one replica
per processor, not many replicas per processor.

Steve

Hi Steve,
Please find the initial and final coordinates for my simulation which was not working; and the input file I use. The system size here is 5400 atoms. What I mean by “simulation not working” is that the simulation just hangs up prior to the “fix neb” command and the neb is not initiated.
The same simulation works for a slightly smaller system of 2880 atoms.

Thank you !
Sankar

perfect_new_screw.Fe (300 KB)

perfect_new_screw_disp.in (427 KB)

in.neb.vac (643 Bytes)

Fixed a communication bug that was causing the problem
with Sankar's input. See the 4Apr patch. Hopefully
this fixes your problem as well Ajing.

Steve

Steve,

Thanks, now it is working great.

Ajing

Hi Steve, Ajing,
Yes, It is now working for me too (even for larger systems) !
Thank you very much, Steve.

Regards
Sankar