I’m using Atomsk to generate a cylindrical sample and reading it into LAMMPS using the read_data command. The model should contain around 1.6 million atoms, but occasionally, LAMMPS reads a completely nonsensical number of atoms—something like 140730947889715—and then hangs indefinitely.
Strangely, if I cancel the job and re-run the exact same script and data file, it often works perfectly without any changes. This suggests the issue is intermittent and not a problem with the data file itself.
To help debug this, I’ve attached:
• The LAMMPS input script (in.creatGB_cyln_CoAl_fccAl.txt)
• The LAMMPS output file from a failed run (Outrerunpolycrystal.o4619145)
in.creatGB_cyln_CoAl_fccAl.txt (3.3 KB)
Outrerunpolycrystal.o4619145 (438 Bytes)
Here is an excerpt of the cylindrical_sample_coal1.lmp data file generated by Atomsk:
Bcc CoAl alloy oriented X=[001] Y=[1-10] Z=[110].
1692799 atoms
2 atom types
0.000000000000 235.832000000000 xlo xhi
0.000000000000 333.516812840000 ylo yhi
0.000000000000 467.736993620000 zlo zhi
Masses
1 58.93319500 # Co
2 26.98153860 # Al
Atoms # atomic
1 1 103.412091451360 50.676292171009 1.928437528366
2 2 104.853965941040 50.679617333633 -0.104155673739
3 1 106.290854942240 50.677853029693 1.930397346369
4 2 107.732510108160 50.681094813114 -0.102167791516
5 1 109.169491083840 50.679423893881 1.932478775991
6 2 110.610936359280 50.682575627763 -0.100086361895
7 1 112.047981009600 50.680984752566 1.934653753011
8 2 113.489230544480 50.684049772076 -0.097930094354
Has anyone else experienced this kind of erratic behavior with read_data in LAMMPS?
Any insight would be appreciated. Thanks!