[lammps-users] the problem about time step

If I use the following condition for NVT simulation, it is completed successfully.

unit metal

timestep 0.001

fix 1 all temp/berendsen 298. 298. 0.5
fix 2 all nve

However, if I change them to timestep 0.0005 and fix 1 all temp/berendsen 298. 298. 0.25, it stops during the simulation.
I’ve just changed the time step to half value with modifying relaxation time to half of original vllue (500 fold of time step) consequently.

What’s wrong with it?

Best regards,

Luke

2010/8/4 JhonY. I. <[email protected]...>:

If I use the following condition for NVT simulation, it is completed
successfully.

unit metal
...
timestep 0.001
...
fix 1 all temp/berendsen 298. 298. 0.5
fix 2 all nve
...

However, if I change them to timestep 0.0005 and fix 1 all temp/berendsen
298. 298. 0.25, it stops during the simulation.
I've just changed the time step to half value with modifying relaxation time
to half of original vllue (500 fold of time step) consequently.

why change the relaxation time? it is very short already.
reducing the time step doesn't change the speed of
any relaxation that is described on an absolute time scale.

What's wrong with it?

who can tell? you don't provide any meaningful information,
that would allow to analyze what is going wrong.

most likely, some other part of your input is wrong and
the first simulation just worked by chance.

cheers,
     axel.

Dear all:

First thanks for your concerns on my question, Dr. Axel Kohlmeyer, udayagiri sai babu, and Ismail Ahmed E really.
It stopped without error message during simulation but it seems that it stopped at the identical step every time.
I am checking now whether it may be concerned to my computer system and if I have a problem still, I will ask your helps again.

Best regards,

Luke