Errors occurred when run some of the examples

[Jack Huang]

Dear developers, 

Thanks for your efforts to develop MBX. 

I think I installed the MBX correctly. I can pass all the tests if disable MPI, and most of them when MPI is enabled. (TestLog1, TestLog2)  and run the following two examples successfully. (Output logs)

MBX-0.7/plugins/lammps/examples/ch4

MBX-0.7/plugins/lammps/examples/ch4_h20

However, when I tried to run these two, errors occurred. (Errors)

MBX-0.7/plugins/lammps/examples/na+_bulk

MBX-0.7/plugins/lammps/examples/na+_slab

The main errors are as follows:

ERROR on proc 1: Unsupported molecule type in MBX (../fix_mbx.cpp:2312)
Last command: fix 2 all mbx 2 1 na 277 h2o json mbx.json
ERROR on proc 0: Unsupported molecule type in MBX (../fix_mbx.cpp:2312)
Last command: fix 2 all mbx 2 1 na 277 h2o json mbx.json

I  can't find any solution to this problem on the internet. I hope you can give me some suggestions. 

Thank you!

Best regards, 

Jie

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

Jie HUANG
Research Assistant
N23-6002, University of Macau, Avenida da Universidade, Taipa, Macau, China
Institute of Applied Physics and Materials Engineering, University of Macau

[Marc Riera]

Hello Jie,

Thanks for letting us know. Indeed, it is a problem with the definition of the monomers in the LAMMPS input file. We changed the definition of Na ion from "na" to "na+", and I forgot to update those examples. I will patch this right away, but you can just fix this by replacing the "na" monomer type by "na+". this means that, in the line:

fix 2 all mbx 2 1 na 277 h2o json mbx.json

you must replace "na" by "na+":
fix 2 all mbx 2 1 na+ 277 h2o json mbx.json

After this, the issue should be solved.

Regarding the unittests, make check will only work without MPI. this reminds me to open an issue to deactivate "make check" when the code has been configured with MPI. The MPI compilation should only be used with LAMMPS, not for anything else.

All the best,

Marc

[Marc Riera]

Hello Jie,

I have fixed the issues in the main repository. You can either change it locally or do a git pull and everything should be good to go.

I really appreciate feedback and problem reports, so please let us know if there is anything else not working as expected.

All the best,

Marc

[Jack Huang]

Hi Marc, 

Thank you for your reply! 

The problem is solved.  

Best regards,

Jie

PS: 

Another fact I noticed in my case is the compilation process consumes over 30G of memory on Linux HPC. I'm not sure this is the normal case.  I believe this is the reason why I can't install MBX on my Linux/macOS laptop because my laptop only has 8G of memory. (Install Log)

[Marc Riera]

That sounds about right. The polynomial files are huge and they require a lot of memory if compiled with optimizations. We are working on this, but you should be able to compile the code in your laptop if you use only 1 thread to compile it. Howevevr, it will be a 30 or 40 min compilation, in that way. If you just wanna test things before you submit them on the supercomputer, my advice would be to compile the code with --disable-optimization (when configuring the makefile). This will add the -O0 flag and make the compilation fast and memory free. I do not recommend this if you want to run production runs.

Please let us know if there is anything else we can help with!

All the best,

Marc