Dear LAMMPS Community!
I've been experiencing some crashes using the latest version of LAMMPS
(10 Aug 2015) when starting a simulation using fix nvt and kspace
pppm/tip4p.
When running the simulation LAMMPS dies with the message ERROR: Out of
range atoms - cannot compute PPPM (../pppm_tip4p.cpp:107).
I've attached a sample configuration and an input file to demonstrate
the problem. A couple of notes:
- Turning of the fix rattle has no effect on the problem
- If you change fix nvt to fix nve there is no problem
- I suspect that the problem is related to particles 5 and 11 being
directly on the edge of the box. These particles are slightly shifted to
positions outside the box before the PPPMTIP4P::particle_map method is
called, leading to the averaging stencil not being within the current
processors simulation box.
So my question is, is there a way to fix this without changing the
initial condition? Also I am a little concerned that this problem
appears by chance during a longer simulation run.
Thanks in advance for your help!
Best regards
Clemens Moritz
initial_config.data (3.35 KB)
md.test.lmp (731 Bytes)
Dear LAMMPS Community!
I've been experiencing some crashes using the latest version of LAMMPS
(10 Aug 2015) when starting a simulation using fix nvt and kspace
pppm/tip4p.
When running the simulation LAMMPS dies with the message ERROR: Out of
range atoms - cannot compute PPPM (../pppm_tip4p.cpp:107).
I've attached a sample configuration and an input file to demonstrate
the problem. A couple of notes:
- Turning of the fix rattle has no effect on the problem
- If you change fix nvt to fix nve there is no problem
- I suspect that the problem is related to particles 5 and 11 being
directly on the edge of the box. These particles are slightly shifted to
positions outside the box before the PPPMTIP4P::particle_map method is
called, leading to the averaging stencil not being within the current
processors simulation box.
no. the problem is due to the bad choice of time constant in fix nvt. it
is *far* too short for your choice of units and will lead to extreme forces
which leads to the instability you observe. please note, that you can still
compute PPPM at the first step; thing go sour only in the second step.
axel.
Ahh,
I thought the time units are picoseconds. Thanks!
Regards
Clemens