[lammps-users] an error in Neighbour::build_one

Dear Steve,
Lammps has some error when the Neighbours list build in "compute cna/atom"
command.
Error message is:
[claster:21984] *** Process received signal ***
[claster:21984] Signal: Segmentation fault (11)
[claster:21984] Signal code: Address not mapped (1)
[claster:21984] Failing at address: 0x11eda4ee4
[claster:21984] [ 0] /lib64/libpthread.so.0 [0x34aa40eb10]
[claster:21984] [ 1]
/home/sas/bin/lmp_linux(_ZN9LAMMPS_NS8Neighbor8full_binEPNS_9NeighListE+0x2c1)
[0x55ab5f]
[claster:21984] [ 2]
/home/sas/bin/lmp_linux(_ZN9LAMMPS_NS8Neighbor9build_oneEi+0x192)
[0x555168]
[claster:21984] [ 3]
/home/sas/bin/lmp_linux(_ZN9LAMMPS_NS14ComputeCNAAtom15compute_peratomEv+0x111)
[0x46e05d]
[claster:21984] [ 4]
/home/sas/bin/lmp_linux(_ZN9LAMMPS_NS10DumpCustom5countEv+0x13d)
[0x4a6861]
[claster:21984] [ 5]
/home/sas/bin/lmp_linux(_ZN9LAMMPS_NS4Dump5writeEv+0x139) [0x4a2999]
[claster:21984] [ 6]
/home/sas/bin/lmp_linux(_ZN9LAMMPS_NS6Output5writeEl+0xb6) [0x5663ba]
[claster:21984] [ 7]
/home/sas/bin/lmp_linux(_ZN9LAMMPS_NS6Verlet3runEi+0x3f5) [0x6033fb]
[claster:21984] [ 8]
/home/sas/bin/lmp_linux(_ZN9LAMMPS_NS3Run7commandEiPPc+0x701) [0x5e6065]
[claster:21984] [ 9]
/home/sas/bin/lmp_linux(_ZN9LAMMPS_NS5Input15execute_commandEv+0xfe0)
[0x537486]
[claster:21984] [10]
/home/sas/bin/lmp_linux(_ZN9LAMMPS_NS5Input4fileEv+0x2b6) [0x537ea8]
[claster:21984] [11] /home/sas/bin/lmp_linux(main+0x5f) [0x54023b]
[claster:21984] [12] /lib64/libc.so.6(__libc_start_main+0xf4) [0x34a981d994]
[claster:21984] [13] /home/sas/bin/lmp_linux(__gxx_personality_v0+0x1f9)
[0x453b29]
[claster:21984] *** End of error message ***
Segmentation fault

Input files are in attachements. Have You any suggestion to fix the problem?

Aleexander Vorontsov

in.1 (817 Bytes)

Cu_u3.eam (35.6 KB)

crd.lmp (69.8 KB)

The problem is you are waiting too long to reneighbor, due to this setting:

neigh_modify every 20 delay 0 check yes

And 1 or more atoms have moved far outside the box. If you use every 1,
it runs fine. I'm still checking why the code crashed and didn't just flag
this as an error.

Steve