[lammps-users] about the fix deposit

Hi,
I use the latest version (2010.1018th) lammps to simulate deposition. the input file below works well while using the 2008.5.21th version. But using the latest version (2010.1018th) lammps ,it could not run in the right way. In the first 500 steps (relaxtion of the atom system) the software works right,when begin to deposit,it will be error. the command used is : mpirun -np 30 ./lmp_g++<in.deposit
,and the error information is :

Nlocal: 174.933 ave 352 max 0 min
Histogram: 12 0 0 0 0 6 0 0 0 12
Nghost: 550.733 ave 1245 max 0 min
Histogram: 12 0 0 0 4 2 0 6 0 6
Neighs: 0 ave 0 max 0 min
Histogram: 30 0 0 0 0 0 0 0 0 0
25 - MPI_ALLREDUCE : Message truncated
[25] Aborting program !
[25] Aborting program!
p25_17870: p4_error: : 14
15 - MPI_ALLREDUCE : Message truncated
[15] Aborting program !
[15] Aborting program!
p15_17800: p4_error: : 14
5 - MPI_ALLREDUCE : Message truncated
[5] Aborting program !
[5] Aborting program!
p5_17730: p4_error: : 14
rm_l_25_17875: (15.925781) net_send: could not write to fd=5, errno = 32
rm_l_15_17805: (16.410156) net_send: could not write to fd=5, errno = 32
p11_17772: p4_error: : 14
p17_17814: p4_error: net_recv read: probable EOF on socket: 1
p14_17793: p4_error: net_recv read: probable EOF on socket: 1
p4_17723: p4_error: net_recv read: probable EOF on socket: 1
p20_17835: p4_error: net_recv read: probable EOF on socket: 1
rm_l_11_17777: (16.535156) net_send: could not write to fd=5, errno = 32
p24_17863: p4_error: net_recv read: probable EOF on socket: 1
p9_17758: p4_error: net_recv read: probable EOF on socket: 1
p19_17828: p4_error: net_recv read: probable EOF on socket: 1
rm_l_17_17819: (16.343750) net_send: could not write to fd=5, errno = 32
p28_17891: p4_error: net_recv read: probable EOF on socket: 1
rm_l_20_17840: (16.164062) net_send: could not write to fd=5, errno = 32
、、、、、

the input file::

echo both

initilization

units metal
dimension 3
boundary p p fm
atom_style atomic
newton on
#atom definition
lattice diamond 3.5667
region box1 block 0 8 0 8 0 20
create_box 4 box1
#neighbor 2.0 bin
#neigh_modify delay 0 every 1 check yes
#setting
mass 1 12.011150 # the mass of C
mass 2 12.011150 # the mass of C
mass 3 12.011150 # the mass of C
mass 4 12.011150 # the mass of C
#EAM potentials
#pair_style airebo 3.0 1 1
pair_style tersoff
#pair_coeff * * CH.airebo C C C C
pair_coeff * * SiC.tersoff C C C C
region box2 block 0 8 0 8 0 10
create_atoms 1 region box2

region 1 block INF INF INF INF INF 8 units box
region 2 block INF INF INF INF 8 25 units box
region 3 block INF INF INF INF 25 36 units box
group lower region 1
group constant region 2
group newdun region 3
group mobile union constant newdun

set group constant type 2
set group newdun type 3
#minimize 1.0e-6 1.0e-6 1000 1000
velocity mobile create 300.0 5812775 sum no
fix 1 all nve
fix 2 lower setforce 0.0 0.0 0.0
fix 3 constant temp/rescale 5 300.0 300.0 0.01 1.0
#compute 4 mobile temp
thermo_style custom step temp pe ke etotal press vol
#thermo_modify temp 4
thermo 10
run 500
group 44 type 4
group whole1 union mobile 44
thermo 1
region box3 block 1 27 1 27 69 70 units box
fix 4 44 deposit 1000 4 1000 29494 region box3 vx 0.0 0.0 vy 0.0 0.0 vz -500.0 -500.0 attempt 50 global 10.0 10.0 units box
fix 6 all wall/reflect zlo zhi
compute 1 44 ke
compute 2 newdun temp
compute 3 constant temp
thermo_style custom step temp etotal c_1 c_2 c_3
timestep 0.0005
run 1000000
dump 1 all xyz 50 cual.xyz
dump 2 all atom 50 dump.cual
run 30000

I think this error was fixed by the 19Oct10 patch.
It had to do with the full neighbor list stats output between
runs - not with fix deposit itself. Note that you
are running a small problem on 30 procs which
will likely be inefficient - some of your procs have no
atoms.

Steve

2010/10/15 黄德明 <[email protected]...>:

Hi:

I'm running a simulation with pairwise, bonded, and angle interactions. I need to use the fix bond/break command to remove bonds in the system, but this does not affect the angle interactions.

Just wondering if there is a workaround (however inefficient) that can turn off the angle interactions between atoms that include the broken bond, but only when the bond breaks.

Thanks.

Gopinath.

There is no workaround. Somone would have to add logic and code
to fix bond/break to detect and break the angles.

Steve