LAMMPS WWW Site - LAMMPS Documentation - LAMMPS Mailing List Archives
Re: [lammps-users] Different potential energy calculation results
[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: [lammps-users] Different potential energy calculation results


From: Xiao Jia <xiaoj@...1508...>
Date: Tue, 19 Dec 2017 16:36:20 -0500

Dear Dr. Kohlmeyer,

Sorry for the missing information. 
I am using the latest stable version of lammps, 11Aug2017. 
I tried another case with Morse potential, which generates consistent potential value to the log file.
While for EAM potential, it does not. 

I am not sure how to quickly reproduce the simulation, because I do the modification in the cpp code.
This is how I did in the code.
I put this code segments within the end_of_step part in fix_ttm_mod.cpp and output the summation of all the potential energy as potential_all[0],
which is used for comparison with the log file.

int i, j, ii, jj, inum, jnum, itype, jtype;
  double xtmp, ytmp, ztmp, delx, dely, delz, evdwl, fpair;
  double rsq, eng, r, dr, dexp, factor_coul, factor_lj;
  int *ilist, *jlist, *numneigh, **firstneigh;
  double vir[3];

  double *special_coul = force->special_coul;
  double *special_lj = force->special_lj;
  int newton_pair = force->newton_pair;

  inum = force->pair->list->inum;
  ilist = force->pair->list->ilist;
  numneigh = force->pair->list->numneigh;
  firstneigh = force->pair->list->firstneigh; 

   potential_single[0] = 0.0;
   potential_all[0] = 0.0;

for (ii = 0; ii < inum; ii++)
  {
   i = ilist[ii];
   xtmp = x[i][0];
   ytmp = x[i][1];
   ztmp = x[i][2];
   itype = type[i];
   jlist = firstneigh[i];
   jnum = numneigh[i];
   for (jj = 0; jj < jnum; jj++)
   {
    j = jlist[jj];
    factor_lj = special_lj[sbmask(j)];
    factor_coul = special_coul[sbmask(j)];
    j &= NEIGHMASK;
    delx = xtmp - x[j][0];
    dely = ytmp - x[j][1];
    delz = ztmp - x[j][2];
    rsq = delx*delx + dely*dely + delz*delz;
    jtype = type[j];
    if (rsq < cutsq[itype][jtype])
    {
     eng = pair->single(i, j, itype, jtype, rsq, factor_coul, factor_lj, fpair);
     potential_single[0] += eng;
    }
   }
  }
  MPI_Allreduce(&potential_single[0], &potential_all[0], 1, MPI_DOUBLE, MPI_SUM, world);

I am not sure why different results can be generated for different potential, Morse and EAM,
and is there any difference to refer the same function single() by using different potentials.

Thanks!


On Tue, Dec 19, 2017 at 9:45 AM, Axel Kohlmeyer <akohlmey@...24...> wrote:


On Tue, Dec 19, 2017 at 3:33 AM, Xiao Jia <xiaoj@...1508...> wrote:
Dear LAMMPS users,

I am referring the function single () defined in pair.cpp to calculate the pairwise force coefficient fpair and the potential energy.
eng=pair->single (i, j, itype, jtype, rsq, factor_coul, factor_lj, fpair);
The current potential I am using is EAM potential.

The weird observation I have is that my summation of the potential energy (eng value of each pair) from all the pairs does not match with the output from log file. The value I get is -21936114 and the value from log file is -63005273, the difference is close to 3 times. 

The calculation of pressure based on fpair seems to be consistent with the value from log file, but the difference in potential energy is hard to understand. 

​your e-mail is missing two important pieces of information.

1) which LAMMPS version you are using exactly. what you are reporting sounds a lot like a bug that was fixed not so long ago.
2) an example input deck that allows to quickly reproduce your issue (in case this is a real bug and still exists with the latest patch release).

axel.​


 

I will appreciate it so much if anyone can give me any idea about this issue.
Thanks!


------------------------------------------------------------------------------
Check out the vibrant tech community on one of the world's most
engaging tech sites, Slashdot.org! http://sdm.link/slashdot
_______________________________________________
lammps-users mailing list
lammps-users@...655....net
https://lists.sourceforge.net/lists/listinfo/lammps-users




--
Dr. Axel Kohlmeyer  akohlmey@...24...  http://goo.gl/1wk0
College of Science & Technology, Temple University, Philadelphia PA, USA
International Centre for Theoretical Physics, Trieste. Italy.