Hello users,
I have created a data file that contains atom_style full and bonds. But when trying use it in simulation, read_data gives me the following error.
“overflow input size in dendezvous_a2a”.
I was looking in mailing list but did not get any answer related to this error. My data contains 4554 atoms and 233550 bonds. Does this large number of bond creating this error?
Regards,
Tan H
Please always report the version of LAMMPS that you are using and the platform that you are running on.
That number of bonds is unusual. Can you explain why it is so large?
For these kind of problems it is usually best to report them as issue on GitHub with a simple way included to reproduce them.
Axel
This is due to create large particles and making them rigid by using bonds.
sorry, but this makes no sense. adding bonds will not result in generating rigid particles.
LAMMPS has rigid body integrators in the various fix rigid command variants for that purpose.
please explain why you are not using those.
more importantly if you do this kind of extreme crosslinking you can overflow the counters in the exclusion list generation.
by default, LAMMPS will exclude all 1-2, 1-3, and 1-4 non-bonded pairs from computation and the failing rendezvous communication is used in collecting this information (among other steps). If you have a large number of bonds, this search for excluded pairs can become 1) a performance bottleneck since this is an O(N**3) process with the number of bonds per atom (normally not a problem, since it is rare to find a configuration where individual atoms have more than ~12 bonds (and then the atoms they are bound to usually only have 1 or 2 bonds), and 2) it may also overflow the 32-bit counters used for communication buffers, which is likely the reason why you get the error message you are reporting.
so it seems the problem is less a problem of LAMMPS but rather a problem of an ill conceived or incorrectly realized model.
axel.
I have started with one particle consist of 506 atoms and 25950 bonds. This works fine but when replicate same particle 9 times LAMMPS create that error. So I also think that overflow 32 bit counter happens here.
Previously I have used fix rigid command instead of making bonds to perform DPD simulation. Fix rigid was used for the time integration of rigid body and fix nve for the nonrigid solvent. Which gives me unexpected pressure in the system. I have read several literature related to DPD and all of them used constant bonds to keep particles together.
Any suggestion from a person like you will be always helpful.
Regards,
Tan H
I have started with one particle consist of 506 atoms and 25950 bonds. This works fine but when replicate same particle 9 times LAMMPS create that error. So I also think that overflow 32 bit counter happens here.
so please check the output of that run. you should see the extreme growth when you look at the lines following:
Finding 1-2 1-3 1-4 neighbors …
the number of 1-4 neighbors should be very large.
Previously I have used fix rigid command instead of making bonds to perform DPD simulation. Fix rigid was used for the time integration of rigid body and fix nve for the nonrigid solvent. Which gives me unexpected pressure in the system. I have read several literature related to DPD and all of them used constant bonds to keep particles together.
“keeping particles together” is different from making them “rigid”.
question is: how many bonds need to be created for that purpose?
i.e. how is determined which particles need to be bound to which?
and are these bonds in addition to the non-bonded interactions or to replace the non-bonded interactions?
the default setting in LAMMPS is the second option, but my hunch is that this is not the model you describe.
so my assertion that this is an ill-conceived or incorrectly realized model remains.
axel.
I have created the particle using LAMMPS and created bond using Ovito. Each beads are connected to each other by bonds where bonds types are harmonic. Following this approach because in almost all the literature, they have used such types of bond to create large particle. Non bonded interactions are controlled by pair_style dpd. I am not sure if there are another approach to create coarse grained particles using bonds and use them in DPD simulation. Since using fix_rigid is giving abnormal pressure, I have used this method. If you suggest some other approach to create particles with bonds will be always helpful as my method of using bonds are not suitable for LAMMPS.
Regards,
Tanvir Hossain
Sorry, but a) this does not provide any explanations or clarifications to the concerns that I have voiced, b) the picture does not explain anything, c) I don’t believe that the model you describes requires bond from all particles of the nano particle to all other particles in the same nanoparticle (at least that won’t be what I’d do for something like this, but without checking the literature I cannot say for certain if that is consistent with the model you are describing), and most importantly d) I don’t have the time to figure out your research for you. I cannot say why fix rigid is producing unexpected pressure (most likely you didn’t set up your calculation correctly) or what is the most suitable approach to realize the model you want to do. Sufficient explanation should be in the published literature, or failing that you need to discuss with colleagues or collaborators that have experience in modeling the kind of system that you want to model.
Bottom line, from what you have shown and explained (or not), there is little evidence that what you are doing is done in a way that can result in a meaningful model. The fact that you are not addressing crucial concerns just reaffirms that there is something seriously wrong in your approach and that you need competent tutoring.
Axel.
Dear Tanvir
You mentioned that you are having trouble with running simulations with rigid molecules at constant pressure.
The “fix rigid” documentation offers many different ways to run these kinds of simulations.
If you are trying to mix rigid and flexible molecules together, then it can be a little bit more tricky. In that case, have you taken a look at the TIP5P example? (That example uses moltemplate, which might not be relevant to you.) However there is a LAMMPS input script for that example located here (which does not depend on moltemplate). The relevant lines from that example are here: