mpi parallel plumed/lammps

[jutta...@gmail.com]

Hello,

I'm having trouble to run lammps patched with plumed in mpi parallel, and I was wondering if someone has some experience in compatibility with certain mpi-versions, or possible other sources of the error.  Compilation of the programmes works fine, and it's all compiled on the same machine with the same settings, but I get errors when running my simulations.

The problem is that the calculation doesn't crash with a segmentation fault or anything, but I randomly get errors in lammps that only occur when using the 'fix plumed'.  Here are two of the error messages that appear in the lammps log file:

ERROR: Non-numeric pressure - simulation unstable (../fix_nh.cpp:1029)

ERROR: Lost atoms: original 4000 current 3998 (../thermo.cpp:433)

The first error appeared e.g. directly at the beginning of the run, the second only after a couple of hundred MD steps.  Also, I don't always get the same error, even if I run exactly the same simulation twice.  Or if I run on a different number of processors. Or if I change a keyword in the plumed input file, although that keyword is not even relevant to the MD run.  So, I think that there is something wrong with the parallel communication between lammps and plumed.

If I run the same lammps executable without any 'fix plumed' in the input file it runs fine.  Also if I run plumed by itself it's all fine.

My lammps version is 170330 with plumed v2.3.1

I'm using intel openmpi 2.0.2 and gcc 4.8.4

Within lammps I use the following command to use plumed:

fix 2 all plumed plumedfile plumed.dat outfile dAFED.log

And my plumed file contains the following lines:

UNITS LENGTH=A

q6: Q6 SPECIES=1-4000 SWITCH={TANH R_0=0.01 D_0=4.4 D_MAX=4.5} MEAN LOWMEM

q4: Q4 SPECIES=1-4000 SWITCH={TANH R_0=0.01 D_0=4.4 D_MAX=4.5} MEAN LOWMEM

aq6: LOCAL_AVERAGE SPECIESA=q6 SPECIESB=q6 SWITCH={TANH R_0=0.01 D_0=4.4 D_MAX=4.5} MEAN LOWMEM

aq4: LOCAL_AVERAGE SPECIESA=q4 SPECIESB=q4 SWITCH={TANH R_0=0.01 D_0=4.4 D_MAX=4.5} MEAN LOWMEM

ex: EXTENDED_LAGRANGIAN ARG=aq4.mean,aq6.mean  KAPPA=1000000,1000000 TAU=5,5  FRICTION=1,1 TEMP=100000

PRINT ARG=aq4.mean,ex.aq4.mean_fict,ex.aq4.mean_vfict  FILE=colvar_aq4 

PRINT ARG=aq6.mean,ex.aq6.mean_fict,ex.aq6.mean_vfict  FILE=colvar_aq6

But it also crashes if I don't use the extended lagrangian and just print, or even if I just calculate something like the volume and print that.  If anyone has any suggestions, or if there are some known incompatibility with certain mpi versions would be great.

Thanks!

Jutta

[Pablo Piaggi]

Dear Jutta,

I started to see this issue a few months ago. I am not sure if it is due to a change in lammps or plumed. However, specifying plumed as the first fix in lammps input has always solve the problem for me. 

Let me know if it works for you as well. 

Regards,

Pablo

--
You received this message because you are subscribed to the Google Groups "PLUMED users" group.
To unsubscribe from this group and stop receiving emails from it, send an email to plumed-users+unsubscribe@googlegroups.com.
To post to this group, send email to plumed...@googlegroups.com.
Visit this group at https://groups.google.com/group/plumed-users.
To view this discussion on the web visit https://groups.google.com/d/msgid/plumed-users/24fc19a9-313e-40be-b4e4-bd0de269c133%40googlegroups.com.
For more options, visit https://groups.google.com/d/optout.

[jutta...@gmail.com]

Dear Pablo,

yes, thanks, strangely enough switching the order of the fix in the lammps input seems to work for now, but only if I do not change any units in plumed.  If I have the line

UNITS LENGTH=A

in my plumed input file it still crashes.  If I remove this and specify everything in 'nm' it seems to work so far.

To me it seems as if there is some inconsistency when passing variables back and forth between lammps and plumed.  Maybe one of the contributors of the lammps-patch could have a look to figure out what's going on?

Thanks again!

Jutta

To unsubscribe from this group and stop receiving emails from it, send an email to plumed-users...@googlegroups.com.