Reply to: Questions about i-Pi+Lammps+plumed by MLPs have a wrong Trajectory

[Mariana Rossi]

Hi Chris -- and sorry, I accidentally deleted your original thread in the forum, trying to answer you!

The higher temperature is due to the fact that you are using PIGLET, which "messes up" with the momenta (in the desired way, but that makes the temperature output meaningless).

I was wondering how your plumed.dat file looks like? It is not in the zip file you sent.

Additionally, have you checked that the example in i-pi/examples/features/metadynamics/pimd_metadynamics_zundel/plumed/plumed.dat works as expected with your installation?

Sorry for the accidental deletion, maybe you want to repost? 

Cheers,

Mariana

N.B. I also sent this to Chris Wu and this is his original message:
---

Hello

When I was running metadynamics using lammps+ipi+plumed, some problems occurred.

I used MLPs to run the self-decomposition of water molecules. I can run ipi+lammps without plumed.When i ran lammps+plumed without adding I-PI, I could decompose the water molecules into a hydroxide ion and a hydrated hydrogen ion very well. However, after adding I-PI with the same control input file, I found that the trajectory was very strange and the system seemed to have disintegrated. The water molecules in the box have been divided into oxygen atoms and hydrogen atoms. I don't know if it's a problem with the version or the parameters I input.

-------------------

When i delet these code,  I found that the trajectory was normal and the atoms wouldn't scatter, but plumed's COLVAR file couldn't be output.

         <bias>
           <force forcefield="plumed" nbeads="1">
             <interpolate_extras> [ doo, mtd.bias ] </interpolate_extras>
           </force>
         </bias>

-------------------

Furthermore, is the temperature (kelvin) in the fourth column of the .out file output by ipi the actual temperature for each step? Why do I set the temperature to 300K in input.xml, but the output file shows 500K or even higher

Besides, my i-Pi version is 3.0 and the plumed version is 2.9.0.

I have placed the relevant documents in the attachment for your reference.

Looking forward to your answer

[Michele Ceriotti]

Note also that as stated in the manual, fetching extras from plumed requires 2.10 (the github main version)

This might explain the symptoms.

M

[wu chris]

Thank you very much for your patient answer!

I now provide you with the necessary files, including plumed.dat, the initial structure and the input files. I hope you can help me find the errors in them. Thank you very much!!!!

Since the trajectory file is very large, I took the first 40 frames and made a gif for convenient viewing.

Explanation of the parameters in plumed.dat

VORONOID1 is a function concerning the division of distances between atoms, cc is a parameter that controls the number of ions. The cc value of pure water is 0, while that of completely decomposed water is 2. Therefore, in COLVAR, I hope the cc value is within the range of 0 to 2, and dd is the distance between two ions.

[Mariana Rossi]

Hi Chris,

Did you try with PLUMED 2.10 (main dev version) as Michele suggested? If you still have issues after trying that and after making sure the example in the examples folder I pointed to actually works, we can go deeper in your inputs.

Cheers,

Mariana

[wu chris]

Thank you very much for your patient reply. However, I tried that plumed v2.10 did not match the cv I defined, so I couldn't use plumed v2.10. But I changed the input.xml file. I cancelled the parameter <interpolate_extras> [doo, mtd.bias] </interpolate_extras>, and it worked. However, another problem occurred. After running for some time, the output file of lammps would report an error. The error: ERROR on proc 0: Non-numeric atom coords - simulation unstable (.. /domain_omp.cpp:58)
Last command: run 50000000
However, such an error does not occur when i-pi is not added. May I ask what the problem is?
Another issue is that the timestep I took in the input.xml file is 0.5 femtoseconds, and the timestep in the input file run.in of lammps is 0.0005ps. But why is the first column output in the output file COLVAR of plumed not time? Instead, it is the number of steps. When i don't add i-pi, the output is the normal time
-------
All the input files and some output files are attached in the new folder. Since the COLVAR file is too large, I only took the first few lines
Looking forward to your answer.