Dear LAMMPS users
I am new to lammps, and i use the lammps of the latest version. I am encountering segmentation fault when doing deposition simulation.
Using 28 cpu to do simulation can deposit 20 molecules at 10000 steps to insert a molecule. If increase the area of adding molecule, it can
deposit more molecules at 50 steps to insert a molecule. Using 4 or 8 cpu to do simulation can deposit more molecules at 10000 steps to
add a molecule. But simulation all suffer from segmentation fault. I have read the mail list about the segmentation fault, but i still do not get
rid of it . I guess this is parallel computing problem. Do you have any recommendation?
thanks in advance.
your input is too large and you are requiring too many processors to
try reproduce and debug this.
can you try to reproduce the segfault with a setup using less atoms
and less processors? the smaller,the better. that will enable using
several detailed debugging methods which slow down calculations quite
significantly.
thanks,
axel.
Dear LAMMPS users
I am new to lammps, and i use the lammps of the latest version. I am
encountering segmentation fault when doing deposition simulation.
Using 28 cpu to do simulation can deposit 20 molecules at 10000 steps to
insert a molecule. If increase the area of adding molecule, it can
deposit more molecules at 50 steps to insert a molecule. Using 4 or 8 cpu
to do simulation can deposit more molecules at 10000 steps to
add a molecule. But simulation all suffer from segmentation fault. I have
read the mail list about the segmentation fault, but i still do not get
rid of it . I guess this is parallel computing problem. Do you have any
recommendation?
the problem is elsewhere. you seem to be triggering an issue with
sorting dump files by atom id.
if you comment out the dump_modify lines, your input does not crash.