[lammps-users] Strange results

LAMMPS LAMMPS Mailing List Mirror
1

Dear Lammps user,

       Right now I calculate a rectangular block system (a,b,c are all
different. alpha=beta=gamma=90).
       I calcualte the elastic constants.
      But the results are different with the number of calculation processors.

      Here is the results of 1 processor
      Elastic Constant C11all = 142.8442601 GPa
Elastic Constant C22all = 414.0452621 GPa
Elastic Constant C33all = 203.0386778 GPa
Elastic Constant C12all = 84.49933935 GPa
Elastic Constant C13all = 70.07458349 GPa
Elastic Constant C23all = 89.61608803 GPa
Elastic Constant C44all = 33.86150426 GPa
Elastic Constant C55all = 30.25907884 GPa
Elastic Constant C66all = 15.18591479 GPa

   This is for 2 processors.

Elastic Constant C11all = 134.9056529 GPa
Elastic Constant C22all = 199.3279293 GPa
Elastic Constant C33all = 218.0071914 GPa
Elastic Constant C12all = -75.36347145 GPa
Elastic Constant C13all = 83.67093754 GPa
Elastic Constant C23all = 73.48298501 GPa
Elastic Constant C44all = 21.38525366 GPa
Elastic Constant C55all = 47.8379712 GPa
Elastic Constant C66all = 25.08291871 GPa

And this is for 4 processors.

Elastic Constant C11all = 178.620808 GPa
Elastic Constant C22all = 234.3595784 GPa
Elastic Constant C33all = 273.8363956 GPa
Elastic Constant C12all = 18.15445801 GPa
Elastic Constant C13all = 62.14474542 GPa
Elastic Constant C23all = 90.36618796 GPa
Elastic Constant C44all = 51.37238981 GPa
Elastic Constant C55all = 37.15761233 GPa
Elastic Constant C66all = 21.42501557 GPa

So what should i do next

thanks you.

by the way, lattice constants are never changed.

2

How did you calculate the elastic constants?
Aidan may want to comment.

Steve

1 Like
3

I recognize the output as coming from the script that I wrote. When you run
this script, you have to be sure that the minimization step converges
properly. You should see something like:

  Force two-norm initial, final = 0.000216202 3.06261e-13
  Force max component initial, final = 1.56031e-05 4.8243e-14

Also, your results should not be sensitive to the value of this variable

variable up equal 1.0e-6

For more info on this, see also:

http://lammps.sandia.gov/doc/Section_howto.html#4_18

Aidan

1 Like