lammps Segmentation fault by using 'fix rigid' command

LAMMPS LAMMPS Mailing List Mirror
1

Dear All

I’m trying to use the “fix rigid” command to fix a big particle and two small particles as a molecule.However,it

turns out an error’Segmentation fault(core dumped)’ after ‘setting up run…’.
I use baidu(like google) to search it.The problem is about the ‘virtual address area’.But I don’t know how to

modify the code to make it on the right way.

The input manuscript is attached as below:

2

Dear All

I'm trying to use the "fix rigid" command to fix a big particle and two
small particles as a molecule.However,it

turns out an error'Segmentation fault(core dumped)' after 'setting up
run...'.
I use baidu(like google) to search it.The problem is about the 'virtual
address area'.But I don't know how to

modify the code to make it on the right way.

yes, this is a bug in the rigid body integrator styles. this is
problem that needs to be corrected by a person who knows the source of
LAMMPS well. please stay tuned. thanks for providing a good, complete
and easy to follow example demonstrating the issue. that helps a lot.

axel.

3

please try this modification as a workaround and let us know, if there
are still problems. thanks in advance. axel.

[[email protected]... src]$ git diff
diff --git a/src/RIGID/fix_rigid_small.cpp b/src/RIGID/fix_rigid_small.cpp
index 1e44b34..81b5582 100644
--- a/src/RIGID/fix_rigid_small.cpp
+++ b/src/RIGID/fix_rigid_small.cpp
@@ -1739,6 +1739,10 @@ void FixRigidSmall::setup_bodies_static()
     if (atom->tri_flag) orientflag = 4;
     if (atom->mu_flag) dorientflag = 1;

+ // grow_arrays() has been called before with "extended" always set to 0