I wonder simulation results are different if machines are different.
I run two simulations with the completely same setting.
Surprisingly the results are not identical at all.
I run 12000step, and in the first 50steps the results are the same as each other.
Is it true?
I think a very very small difference (less than 0.0001) accumulated over the 12000step and caused a big difference finally.
You can see the same results in step1 but different results in step12000.
Step1
Machine1
Code: Select all
INSTANT 0.100000000E+00 TIME STEP NUMBER 1
=============================================================
** Valeurs calculées sur cellules
------------------------------
criterion minimum maximum moy. ensemble moy. spatiale
--------------- -------------- -------------- -------------- --------------
Courant/Fourier 3.1321 3967.8 204.92 10.034
Code: Select all
INSTANT 0.100000000E+00 TIME STEP NUMBER 1
=============================================================
** Computed values on cells
------------------------
criterion minimum maximum set mean spatial mean
--------------- -------------- -------------- -------------- --------------
Courant/Fourier 3.1321 3967.8 204.92 10.034
Step12000
Machine1
Code: Select all
INSTANT 0.120000000E+03 TIME STEP NUMBER 12000
=============================================================
** Valeurs calculées sur cellules
------------------------------
criterion minimum maximum set mean spatial mean
--------------- -------------- -------------- -------------- --------------
Courant/Fourier 0.013982 310.3 8.8389 2.4452
Code: Select all
INSTANT 0.120000000E+03 TIME STEP NUMBER 12000
=============================================================
** Computed values on cells
------------------------
criterion minimum maximum moy. ensemble moy. spatiale
--------------- -------------- -------------- -------------- --------------
Courant/Fourier 0.0090879 314.62 8.6604 2.369