[lammps-users] deformed box after npt run with pressure control style "xyz"

Thanks Steve. I’ve add the print statement. The results show that

  1. The results showed by the statement is the same to the output of the thermo lx ly lz.

For example: step 990 lx ly lz = 28.924478 28.924478 74.863908

BOX DILATE: 0 1 28.9245

BOX DILATE: 1 1 28.9245

BOX DILATE: 2 1 74.8639

  1. The radio of the box length a b c:

The beginning: 28.8397:28.8397:74.6745=0.3862:0.3862:1

The end : 28.9245:28.9245:74.8639= 0.3864:0.3864:1

The ratio changes from 0.3862 to 0.3864.

  1. The re-running repeats my previous calculation. So I think this error will be larger as the running time goes on as what I’ve gotten after 36000 running even if it changes so slowly.

I just wonder this error is caused by my special model or by some tiny wrong in the source code. Following is the output:

0 2.0232614 2.9235608 5.5379974 62108.72 -5.5149004 3.0347186 28.839654 28.839654 74.674529 5.6099543 5.7951015 5.2089364 -0.057484868 0.14193013 -0.017237353

BOX DILATE: 0 1 28.8397

BOX DILATE: 1 1 28.8397

BOX DILATE: 2 1 74.6745

830 2.5896789 3.7471515 7.2988867 62526.094 -4.1267752 3.8842963 28.907375 28.907375 74.824528 7.3699137 7.3986651 7.1280815 -0.056022161 0.055412561 0.013413788

BOX DILATE: 0 1 28.9074

BOX DILATE: 1 1 28.9074

BOX DILATE: 2 1 74.8245

990 2.4283236 3.5475842 7.2844631 62633.048 -4.0144855 3.6422772 28.924478 28.924478 74.863908 7.3827125 7.4243636 7.0463133 0.029408006 -0.033111426 -0.11857669

BOX DILATE: 0 1 28.9245

BOX DILATE: 1 1 28.9245

BOX DILATE: 2 1 74.8639

I think you need to run a lot longer and print
out to higher precision. I don't think that
appliying the same dilation factor to a rectangular
box can change the aspect ratio more than epsilon.

Try

if (comm->me == 0) printf("BOX DILATE: %d %15.12g %g\n, ...);

and you can run in parallel.

Steve

2009/2/11 Kitty Ji <[email protected]>: