strange behavior of thermo

Hi

the compilation warnings are fine. For output stuff I have both single precision and double precision versions with a runtime if(sizeof(..)) branch. So one of the branches will actually not be executed, and thats the one where it complains about incompatibility.

Anyway I gonna look into it tomorrow.

Christian

Hi, i think i localized problem.

The strange happens when direction with shrink-wrapped boundaries (seems only fixed, not m) is divided into several processes. I tested it with different configurations of processors and boundaries.

On log below you can see 3 processes on Y direction, s boundary, and extremely raised Py after 1 step with all zeroed forces in system(so potential parameters is not matter, just for test).

feeling as if it turns into a periodic boundary…

LAMMPS (14 Mar 2013)
USER-CUDA mode is enabled (lammps.cpp:394)
package cuda gpu/node/special 3 0 1 2 pinned 0
processors 1 3 1
units metal
boundary p s p
atom_style atomic
newton on

lattice bcc 2.855312
Lattice spacing in x,y,z = 2.85531 2.85531 2.85531
region box block -10 10 -10 10 -10 10
create_box 1 box
Created orthogonal box = (-28.5531 -28.5531 -28.5531) to (28.5531 28.5531 28.5531)
1 by 3 by 1 MPI processor grid
create_atoms 1 region box
Created 16400 atoms

pair_style sw
pair_coeff * * Si.sw Si
mass * 50
neigh_modify delay 1 every 1 check no
thermo 1
thermo_style custom step pxx pyy pzz lx ly lz vol
thermo_modify lost warn
timestep 0.001
fix 1 all nve
fix 2 all setforce 0 0 0
run 5
Memory usage per processor = 3.70642 Mbytes
Step Pxx Pyy Pzz Lx Ly Lz Volume
0 6067363.7 5999486.3 6067363.7 57.10624 57.117661 57.10624 186267.7
1 7145088.9 8.2200286e+13 7145088.9 57.10624 57.117661 57.10624 186267.7
2 7145088.9 8.2200286e+13 7145088.9 57.10624 57.117661 57.10624 186267.7
3 7145088.9 8.2200286e+13 7145088.9 57.10624 57.117661 57.10624 186267.7
4 7145088.9 8.2200286e+13 7145088.9 57.10624 57.117661 57.10624 186267.7
5 7145088.9 8.2200286e+13 7145088.9 57.10624 57.117661 57.10624 186267.7
Loop time of 0.118994 on 3 procs for 5 steps with 16400 atoms

Pair time () = 0.0371122 (31.1883) Neigh time () = 0.0222937 (18.7352)
Comm time () = 0.0482981 (40.5887) Outpt time () = 0.00147907 (1.24298)
Other time (%) = 0.00981092 (8.24489)

Nlocal: 5466.67 ave 5600 max 5200 min
Histogram: 1 0 0 0 0 0 0 0 0 2
Nghost: 19207.7 ave 19616 max 18391 min
Histogram: 1 0 0 0 0 0 0 0 0 2
Neighs: 0 ave 0 max 0 min
Histogram: 3 0 0 0 0 0 0 0 0 0
FullNghs: 360267 ave 374000 max 332800 min
Histogram: 1 0 0 0 0 0 0 0 0 2

Total # of neighbors = 1080800
Ave neighs/atom = 65.9024
Neighbor list builds = 5
Dangerous builds = 0

25.03.2013, 06:55, “Trott, Christian Robert (-EXP)” <crtrott@…3…>: