Relaxation with Reax/c

I want to relax a composite filled with Aluminium and Aluminia. The input file is as follow:

#Lammps 3d Al-Al2O3
units real
boundary p p p

atom_style charge
read_data al_al2o3_whole.txt

reax args: hbcut hbnewflag tripflag precision

pair_style reax/c control.file
pair_coeff * * ffield 7 3 7

neigh_modify every 10 delay 0 check no

velocity all create 300 27851123

fix f2 all nve
fix f3 all qeq/reax 1 0.0 10.0 1.0e-6 reax/c

thermo 100
thermo_style custom step temp epair etotal press lx ly lz vol
timestep 1.0
dump d1 all cfg 2000 Al-Al2O3.*.cfg id type xs ys zs id type q
dump_modify d1 element Al O Al
run 30000

The enclosure is the data file. Problem comes out when the calculation step is up to 700. The log file is as follow:

LAMMPS (2 Feb 2011)
Reading data file …
orthogonal box = (0 0 0) to (129.6 129.6 129.6)
3 by 2 by 4 processor grid
134167 atoms
Setting up run …
Memory usage per processor = 35.7511 Mbytes
Step Temp E_pair TotEng Press Lx Ly Lz Volume
0 300 -11170213 -11050236 -30366.295 129.6 129.6 129.6 2176782.3
100 290.40561 -11166831 -11050691 -21262.551 129.6 129.6 129.6 2176782.3
200 276.26514 -11161126 -11050641 -18772.264 129.6 129.6 129.6 2176782.3
300 291.94924 -11167319 -11050561 -18801.971 129.6 129.6 129.6 2176782.3
400 297.8291 -11169606 -11050497 -18675.134 129.6 129.6 129.6 2176782.3
500 290.70828 -11166707 -11050446 -18044.791 129.6 129.6 129.6 2176782.3
600 293.99937 -11168045 -11050468 -16219.97 129.6 129.6 129.6 2176782.3
700 292.38619 -11167439 -11050507 -14800.479 129.6 129.6 129.6 2176782.3
rank 15 in job 1 node10_47310 caused collective abort of all ranks
exit status of rank 15: return code 242
ERROR:failed to allocate 121375885632 bytes for array list:three_bodiesapplicatio called MPI_Abort(MPI_COMM_WORLD,-14) -process 15
WARNING: trying to allocate -206158430208 bytes for array list:three_bodies. returning NULL.

What is that ERROR mean?

al_al2o3_whole.txt (6.14 MB)

What version of lamps are you using? There has been an update in March 2011 that fixed many memory errors that I had with reax/c.

Shawn P Coleman

University of Arkansas
Mechanical Engineering

2011/4/1 meihai3166 <[email protected]…127…>