RAM problem

LAMMPS LAMMPS Mailing List Mirror
1

dear axel
I want to simulate movement of water through CNT
but i got this error that after almost 10000 step my RAM becomes full
and it stops
these lines are in my input :
neighbor 1.0 bin
neigh_modify delay 10 every 2# check yes
boundary s s p
…
pair_style hybrid lj/cut/coul/cut 10.0

pair_coeff 1 2 lj/cut/coul/cut 0.105 3.1905
pair_coeff 1 3 lj/cut/coul/cut 0.000 0.000
pair_coeff 1 1 lj/cut/coul/cut 0.0700 3.5500 #ca
pair_coeff 2 2 lj/cut/coul/cut 0.1521 3.5365 #OW
pair_coeff 3 3 lj/cut/coul/cut 0.0000 0.0000 #HW
pair_coeff 2 3 lj/cut/coul/cut 0.0000 0.000

bond_style harmonic

bond_coeff 1 305.000 1.3750
bond_coeff 2 450.000 0.9572
angle_style harmonic

angle_coeff 1 40.000 120.00
angle_coeff 2 55.000 104.5200
…
compute 3 all rdf 18 1 2

fix 2 all nvt temp 298.0 298.0 100

thermo_style custom step press pe ke etotal pzz

thermo 100

dump 1 wat custom 1000 mahdi1.dat id mol type x y z vx vy vz fz

dump 2 all xyz 1000 structure1.xyz

Any help would be appreciated

thanks alot

2

dear axel
I want to simulate movement of water through CNT
but i got this error that after almost 10000 step my RAM becomes full
and it stops
these lines are in my input :

[...]

Any help would be appreciated

nobody can help with that based on so little and incomplete information.

3

dear axel

this is my entire input file and the amount od RAM usage increase with time from 758 up to 3650(almost) and after that run stops
i run it on 4 core in ubuntu
thanks alot

units real
atom_style full
dimension 3
neighbor 1.0 bin
neigh_modify delay 10 every 2# check yes
boundary s s p
#read_restart cnt1.restart.40000
read_data ladata.txt
#pair_style coul/long 10
#pair_style lj/charmm/coul/charmm 10 12
pair_style hybrid lj/cut/coul/cut 10.0

pair_coeff 1 2 lj/cut/coul/cut 0.105 3.1905
pair_coeff 1 3 lj/cut/coul/cut 0.000 0.000
pair_coeff 1 1 lj/cut/coul/cut 0.0700 3.5500 #ca
pair_coeff 2 2 lj/cut/coul/cut 0.1521 3.5365 #OW
pair_coeff 3 3 lj/cut/coul/cut 0.0000 0.0000 #HW
pair_coeff 2 3 lj/cut/coul/cut 0.0000 0.000

bond_style harmonic

bond_coeff 1 305.000 1.3750
bond_coeff 2 450.000 0.9572 #OW_HW OW HW 1200.0 0.9572
angle_style harmonic

angle_coeff 1 40.000 120.00 # 35.00 2.41620
angle_coeff 2 55.000 104.5200 #HOW OW HOW

#kspace_style pppm 1.0e-4

group cnt type 1
group wat type 2 3

fix 1 cnt setforce 0.0 0.0 0.0

min_style cg
minimize 1e-5 1e-5 500000 1000000

#compute 2 wat msd
compute 3 all rdf 18 1 2

fix 2 all nvt temp 298.0 298.0 100

thermo_style custom step press pe ke etotal pzz

thermo 100
restart 5000 cnt1.restart

dump 1 wat custom 1000 mahdi1.dat id mol type x y z vx vy vz fz
#dump 4 wat custom 1000 dump2.dat id type c_2
dump 2 all xyz 1000 structure1.xyz

#timestep 0.001
#run 5000
#timestep 0.4
#minimize 100000
#run 1000
timestep 1.0
run 100000

dear axel
I want to simulate movement of water through CNT
but i got this error that after almost 10000 step my RAM becomes full
and it stops
these lines are in my input :

[…]

Any help would be appreciated

nobody can help with that based on so little and incomplete information.

4

dear axel

this is my entire input file and the amount od RAM usage increase with time from 758 up to 3650(almost) and after that run stops
i run it on 4 core in ubuntu
thanks alot

units real
atom_style full
dimension 3
neighbor 1.0 bin
neigh_modify delay 10 every 2# check yes
boundary s s p
#read_restart cnt1.restart.40000
read_data ladata.txt
#pair_style coul/long 10
#pair_style lj/charmm/coul/charmm 10 12
pair_style hybrid lj/cut/coul/cut 10.0

pair_coeff 1 2 lj/cut/coul/cut 0.105 3.1905
pair_coeff 1 3 lj/cut/coul/cut 0.000 0.000
pair_coeff 1 1 lj/cut/coul/cut 0.0700 3.5500 #ca
pair_coeff 2 2 lj/cut/coul/cut 0.1521 3.5365 #OW
pair_coeff 3 3 lj/cut/coul/cut 0.0000 0.0000 #HW
pair_coeff 2 3 lj/cut/coul/cut 0.0000 0.000

bond_style harmonic

bond_coeff 1 305.000 1.3750
bond_coeff 2 450.000 0.9572 #OW_HW OW HW 1200.0 0.9572
angle_style harmonic

angle_coeff 1 40.000 120.00 # 35.00 2.41620
angle_coeff 2 55.000 104.5200 #HOW OW HOW

#kspace_style pppm 1.0e-4

group cnt type 1
group wat type 2 3

fix 1 cnt setforce 0.0 0.0 0.0

min_style cg
minimize 1e-5 1e-5 500000 1000000

#compute 2 wat msd
compute 3 all rdf 18 1 2

fix 2 all nvt temp 298.0 298.0 100

thermo_style custom step press pe ke etotal pzz

thermo 100
restart 5000 cnt1.restart

dump 1 wat custom 1000 mahdi1.dat id mol type x y z vx vy vz fz
#dump 4 wat custom 1000 dump2.dat id type c_2
dump 2 all xyz 1000 structure1.xyz

#timestep 0.001
#run 5000
#timestep 0.4
#minimize 100000
#run 1000
timestep 1.0
run 100000

dear axel

this is my entire input file and the amount od RAM usage increase with time from 758 up to 3650(almost) and after that run stops
i run it on 4 core in ubuntu
thanks alot

units real
atom_style full
dimension 3
neighbor 1.0 bin
neigh_modify delay 10 every 2# check yes
boundary s s p
#read_restart cnt1.restart.40000
read_data ladata.txt
#pair_style coul/long 10
#pair_style lj/charmm/coul/charmm 10 12
pair_style hybrid lj/cut/coul/cut 10.0

pair_coeff 1 2 lj/cut/coul/cut 0.105 3.1905
pair_coeff 1 3 lj/cut/coul/cut 0.000 0.000
pair_coeff 1 1 lj/cut/coul/cut 0.0700 3.5500 #ca
pair_coeff 2 2 lj/cut/coul/cut 0.1521 3.5365 #OW
pair_coeff 3 3 lj/cut/coul/cut 0.0000 0.0000 #HW
pair_coeff 2 3 lj/cut/coul/cut 0.0000 0.000

bond_style harmonic

bond_coeff 1 305.000 1.3750
bond_coeff 2 450.000 0.9572 #OW_HW OW HW 1200.0 0.9572
angle_style harmonic

angle_coeff 1 40.000 120.00 # 35.00 2.41620
angle_coeff 2 55.000 104.5200 #HOW OW HOW

#kspace_style pppm 1.0e-4

group cnt type 1
group wat type 2 3

fix 1 cnt setforce 0.0 0.0 0.0

min_style cg
minimize 1e-5 1e-5 500000 1000000

#compute 2 wat msd
compute 3 all rdf 18 1 2

fix 2 all nvt temp 298.0 298.0 100

thermo_style custom step press pe ke etotal pzz

thermo 100
restart 5000 cnt1.restart

dump 1 wat custom 1000 mahdi1.dat id mol type x y z vx vy vz fz
#dump 4 wat custom 1000 dump2.dat id type c_2
dump 2 all xyz 1000 structure1.xyz

#timestep 0.001
#run 5000
#timestep 0.4
#minimize 100000
#run 1000
timestep 1.0
run 100000

dear axel
I want to simulate movement of water through CNT
but i got this error that after almost 10000 step my RAM becomes full
and it stops
these lines are in my input :

[…]

Any help would be appreciated

nobody can help with that based on so little and incomplete information.

5

dear axel
this is my entire input file and the amount od RAM usage increase with time
from 758 up to 3650(almost) and after that run stops

758 to 3650 of what? and how do you determine that?

this is still not sufficient data (what LAMMPS version is this, where
is the data file?).

your neighbor list settings seem bogus. don't you get a *lot* of
dangerous builds?
also a 1 fs time step for flexible water is *way* too large.

the only way you can get that much of an increase of RAM use would be
if your force field parameters and or simulation settings are wrong
and thus your system is collapsing and thus your neighbor lists are
blowing up through the steadily increasing number of neighbors.
considering how bogus most of the input looks, that is quite possible.

as per usual in these cases, do not start a simulation of multiple
compounds until you have mastered the skill of setting up a correct
simulation for each compound that can properly reproduce documented
properties of the individual components. this doesn't look at all like
you have reached that skill level.

axel.

6

tnx for your advice and reply
my data file contain

20396 atoms
21541 bonds
32226 angles
0 dihedrals
0 impropers

3 atom types
2 bond types
2 angle types
0 dihedral types
0 improper types

-13.546 13.546 xlo xhi
-13.546 13.546 ylo yhi
5 301 zlo zhi

Masses

1 12.000
2 15.9994
3 1.0080

and i dont get any dangerous builds

Nlocal: 5099 ave 5124 max 5055 min
Histogram: 1 0 0 0 0 0 0 1 1 1
Nghost: 1914 ave 1964 max 1847 min
Histogram: 1 0 0 0 0 1 0 1 0 1
Neighs: 1.79239e+06 ave 1.81252e+06 max 1.7659e+06 min
Histogram: 1 0 0 0 1 0 0 0 1 1

Total # of neighbors = 7169560
Ave neighs/atom = 351.518
Ave special neighs/atom = 9.47705
Neighbor list builds = 0
Dangerous builds = 0

tnx for any help

dear axel
this is my entire input file and the amount od RAM usage increase with time
from 758 up to 3650(almost) and after that run stops

758 to 3650 of what? and how do you determine that?

this is still not sufficient data (what LAMMPS version is this, where
is the data file?).

your neighbor list settings seem bogus. don’t you get a lot of
dangerous builds?
also a 1 fs time step for flexible water is way too large.

the only way you can get that much of an increase of RAM use would be
if your force field parameters and or simulation settings are wrong
and thus your system is collapsing and thus your neighbor lists are
blowing up through the steadily increasing number of neighbors.
considering how bogus most of the input looks, that is quite possible.

as per usual in these cases, do not start a simulation of multiple
compounds until you have mastered the skill of setting up a correct
simulation for each compound that can properly reproduce documented
properties of the individual components. this doesn’t look at all like
you have reached that skill level.

axel.

7

tnx for your advice and reply
my data file contain

so now you have exhausted my patience. this is *still* incomplete and
useless. plenty of information and explanations that i asked for is
still missing.
of course you don't get dangerous builds for the minimization (which i
guess your post run info is taken from), since LAMMPS will reset the
neighbor settings to sane.

i don't have the time to play 20 questions here. this kind of system
has been set up by lots of people without a problem, so the obvious
conclusion is that your input is bad. you will have to debug it
yourself. i have already outlined how to go about it. if you need to
learn basic MD techniques and best practices, please refer to your
local colleague and particularly your adviser. those are the people
that are supposed to help you with that.

have a nice day,
     axel.