Dear lammps users,
I am trying to test a system of RDX with Argon filled inside a spherical void. To simulate this system I am using hybrid style reax/c for C,H,O, N atoms and lj/cut for Ar and Ar + C,H,O,N atoms.
I am testing this sytsem on Cetus on 161212 cores in x,y and z directions.
The CG minimisation works completely fine without any problems. But as soon as I give any velocity corresponding to a very low temperature of 10K, the system gives the folllowing error:
step30-hbondchk failed: H=21 end(H)=3499 str(H+1)=3495
To check if this error is an artefact of my system I have performed the following checks:
-
Verified the atomic positions and they seem fine.
-
Varied the sigma, epsilon for LJ interaction between Ar and C,H,O,N, but no matter what I do it always fails giving the same error as highlighted in bold.
-
To test whether lj interaction is a problem between Ar and C,H,O,N atoms, I completely removed this interaction and even then it fails giving the same error as highlighted in bold.
-
Tried very very low time steps to limit any movement during initial time, but nothing changes.
-
Tried different fixes such as nvt and nve but the result stays the same.
-
Tried different number of cores in all the directions, but nothing works.
So, I do not know whether my system has a problem, or are there any compilation issues which I made while compiling lammps on bgq.
I used the standard lammps makefile provided under machine directory for bgq.
Please find attached my input file as well as my log file for this work.
Data read by the read_data command can be found at this link
https://drive.google.com/file/d/1hoT3I1y8WBFSTt2gT_OytpmMH2JZ8Moo/view?usp=sharing
Excutable can be found at this link
https://drive.google.com/file/d/1RYH4rGzHctjxGuDTBXx17F8X-Embnt-D/view?usp=sharing
Thanks,
Ankit
ffield.reax (15.6 KB)
lmp_control (1001 Bytes)
in.rdx (2.13 KB)
log (8.75 KB)