Code: WARP - Description
From EVOCD
Warp 2001 - Atomistic Stress Simulator, Copyright 1998-2001
Authored by Steve Plimpton
(505) 845-7873, sjplimp@sandia.gov
Dept 9233, MS 1111, Sandia National Labs, Albuquerque, NM 87185-1111
To download this description right click on this link and select "Save Link As...".
Contents |
Synopsis
Warp 2001 - Copyright 1998-2001 - January 2001 release
This distribution contains the Warp code which is a parallel molecular dynamics simulation for modeling stress and strain in materials using embedded atom method (EAM) and Lennard-Jones (LJ) potentials. It is written in F90 and performs message-passing via MPI calls. Thus it should be portable to virtually any parallel or single-processor machine. This code is copyrighted property of Sandia National Laboratories. Please contact the author(s) for a copy of the source code or further information.
Compiling WARP
Type make to see a list of supported machines. You can add a new machine to the list by creating an appopriate Makefile.machine file and adding the new machine to the target list in Makefile. The "serial" target does not require MPI be installed on your workstation. The other targets require MPI.
If the -DCRACK compiler flag is set in Makefile.machine, Warp adds the command "no neighbor" that enables crack propagation simulations.
Running WARP
Warp is run by redirecting an input script to stdin, e.g.
warp_machine < in.file mpirun -np 16 warp_machine < in.file
Input file
The input script consists of a series of single-line commands. Most of the input commands take one or more parameters. The keyword for each command should begin in the leftmost column and all characters in the command should be in lower-case. Parameters can be separated by arbitrary numbers of spaces and/or tabs (so long as the command fits on one line).
The order of commands for performing a simulation are as follows:
(1) set parameters only those different than defaults (2) create box mandatory Either: (3) orient and origin as needed (4) select region mandatory (5) define cutout as needed (6) create atoms mandatory (loop on 3-6 as desired) (7) check for atom overlap as needed (8) select and create types and eltypes as needed (9) select and create vels as needed Or: (3) scale read as needed (4) type translate as needed (5) read atoms mandatory (6) check for atom overlap as needed (7) select and create types and eltypes as needed (8) select and create vels as needed Or: (3) read restart mandatory (10) select and create fixes as needed (11) apply temp and pres controls as needed (12) run can perform multiple times, changing parameters as desired Or: (13) relax minimize potential energy
Note that several parameters must be defined (if not default) before the global simulation box is created and memory is allocated in step (2). These settings include units, extra memory, potential, neighbor, processor grid, periodicity, and lattice.
The following is a complete list of valid Warp inputs and the "type" of parameters that should be used (i=integer,r=real,c=character string) in the input script. More details are given below. Some of the commands are only available with certain compiler flags.
# comment
units lj (c)
extra memory 2.0 1.5 2.0 2.5 (rrrr)
timestep 0.001 (r)
augment t1 1 (i)
use interp 1 (i)
use dimer 0 (i)
gsmooth factor 99.0 (r)
potential eam 0 ../potentials/cuu7 (cic)
potential lj/cut 2.5 (cr)
potential lj/smooth (c)
potential meam matlibfile meamcardfile (ccc)
table 10000 (i)
neighbor 0.3 10 (ri)
no neighbor 2 3 (ii)
processor grid 10 10 10 (iii)
thermo 0 (i)
snapshot 1000 file xmovie (icc)
snap thresh centro 2.0 (cr)
snap column 2 (i)
snap limits inf inf inf inf -1.0 1.0 (rrrrrr)
restart 1000 file (ic)
write atoms 1000 file (ic)
write dynamo 1000 file (ic)
write meamstat 1000 file (ic)
diagnostic name 100 diagfile 2 10.0 ... (cicirrrrr)
numatoms 1000 (i)
periodicity 0 1 0 (iii)
lattice fcc 1.0 (cr)
create box -10.0 10.0 0.0 20.0 0.0 20.0 (rrrrrr)
orient x 1 0 0 (ciii)
orient y 0 1 0 (ciii)
orient z 0 0 1 (ciii)
origin 0.0 0.0 0.0 (rrr)
select region all
select region 0.0 inf 0.0 inf inf inf (rrrrrr)
select type 1 (i)
set eltype 1 (i)
define cutout 0 region -5.0 5.0 inf inf -3.0 inf (icrrrrrr)
define cutout 1 cylinder x y z 3.0 4.0 2.0 -10.0 inf (iccccrrrrr)
define cutout 0 sphere 0.0 0.0 2.0 1.0 (icrrrr)
define cutout 1 cylinder x y z 3.0 4.0 2.0 -10.0 inf (iccccrrrrr)
create atoms 1 (i)
scale read 1.0 1.0 1.0 (rrr)
type translate 28 1 (ii)
type translate 1 3 (ii)
box toler 0.00005 (r)
read atoms atomfile (c)
check overlap 0.25 1 (ir)
create types 1 (i)
create eltypes 1 (i)
displace atoms 100.0 xyz 1.0 0.0 0.0 (rcrrr)
displace atoms 100.0 ramp vy 0.0 5.0 x -10.0 10.0 (rccrrcrr)
displace atoms 100.0 twist 0.0 1.0 x y z 0.0 5.0 -10.0 10.0
(rcrrcccrrrr)
displace atoms 100.0 expand 1 1 0 0.0 0.0 0.0 1000.0 (rciiirrrr)
check vels 1 (i)
create vels thermal 300.0 39480 (cri)
create vels ramp vy 0.0 5.0 x -10.0 10.0 (ccrrcrr)
create vels twist 0.0 1.0 x y z 0.0 5.0 -10.0 10.0 (crrcccrrrr)
create vels expand 1 1 0 0.0 0.0 0.0 1000.0 (ciiirrrr)
create fixes none (c)
create fixes xyz 0 1 -1 (ciii)
create fixes twist x y z 0.0 2.0 1 1 1 (ccccrriii)
create fixes external 0.1 0.0 0.0 (crrr)
temp type 0 (i)
temp control none (c)
temp control rescale 300.0 10 10.0 (crir)
temp control hoover 300.0 1.0 (crr)
temp adjust none (c)
temp adjust xyz 0 1 1 (ciii)
temp adjust ramp vy 0.0 5.0 x -10.0 10.0 (ccrrcrr)
temp adjust twist 0.0 1.0 x y z 0.0 5.0 -10.0 10.0 (crrcccrrrr)
pres control volume 0.0 0.5 (crr)
read restart filename (c)
reset timestep 0 (i)
run 1000 (i)
relax 0 (i)
relaxp 1 1500 4 40 1e-8 1e-6 (iiiirr)
----------------------------------------------------------------------------
# comments
blank lines are ignored
everything on a line after last parameter is ignored
lines starting with a # are echoed into the log file
----------------------------------------------------------------------------
# units
kind of units to use in the simulation
options: lj Lennard Jones reduced units
real Angstroms, degrees K, eV, etc
must be specified before "create box" command
default = lj
----------------------------------------------------------------------------
# extra memory
1st parameter = extra_own = padding factor on allocation of
owned atom arrays
2nd parameter = extra_ghost = padding factor on allocation of
ghost atom arrays
3rd parameter = extra_neigh = padding factor on allocation of
neighbor lists
4th parameter = extra_buf = padding factor on allocation of
communication buffers
factors that affect how much extra memory is allocated when a problem is setup
factor of 1.0 means no padding, factor of 2.0 means 2x longer arrays
typically don't need to change default settings unless Warp tells you
to "boost" some factor at run-time
must be specified before "create box" command
default = 2.0 2.0 2.0 3.0
----------------------------------------------------------------------------
# timestep
timestep size for MD run
for lj units, is in units of tau
for real units, is in picoseconds
default = 0.001 for both lj and real
----------------------------------------------------------------------------
# augment t1
flag specifying whether augmentation of t1 coefficient is needed,
only significant for MEAM
0 = don't augment t1
1 = do augment t1
if don't augment t1, t1 value from the input file is not modified
if augment t1, then 3/5*t3 is added to t1
default = 1
----------------------------------------------------------------------------
# use interp
flag specifying whether interpolation of phi is used for energy calculation
only significant for MEAM
0 = don't use interpolation
1 = do use interpolation
if not use interp then phi is calculated precisely
# and -100 is used for small log values (as Dynamo)
if use interp, polynomial interpolation of phi is used (faster)
default = 1
----------------------------------------------------------------------------
# use dimer
flag specifying whether dimer is used to calculate reference density and phi
only significant for MEAM
0 = don't use dimer
1 = do use dimer
if use dimer then density (and phi) is calculated the same way as in Dynamo
otherwise dimer is not used
default = 0
----------------------------------------------------------------------------
# gsmooth factor
factor determining the length of the G smoothing region
must be positive number
only significant for MEAM if ibar is 0 or 4
if gsmooth factor is larger, G smoothing region is smaller, sharper step
99.0 = short smoothing region, sharp step
0.5 = long smoothing region, smooth step
power of the smoothing function will be -(gsmooth_factor/2)
default = 99.00
----------------------------------------------------------------------------
# potential
kind of potential energy and force interactions to use between all
pairs of interacting atoms
examples:
potential eam 0 filename
potential lj/cut 2.5
potential lj/smooth
potential meam matlibfile meamcardfile
eam option means use a EAM potential from the specified file, the cutoff
for the potential is contained in the file
for the eam option, if the integer flag is 0, the EAM file is assumed
to be in the DYNAMO single-element funcfl format
for the eam option, if the integer flag is 1, the EAM file is assumed
to be in the DYNAMO setfl alloy format, and the potentials are assigned
to element type numbers in the order in which they appear in the file
lj/cut option means use a cutoff LJ potential with the specified
cutoff in units of sigma
lj/smooth option means use a LJ potential that goes smoothly to zero
in both the energy and force at a hard-wired cutoff value of
approximately 1.71 sigma
must be specified before "create box" command
default = lj/cut with 2.5 sigma cutoff
meam option means use a MEAM (modified EAM) potential, reading material
parameters from the matlibfile and the meam card data from meamcardfile.
See the file MEAM.README for an overview of what data is needed in the
meam card.
----------------------------------------------------------------------------
# table
pre-tabulate the potential energy and force functions in a table of the
specified size to speed force computations
the force and energy routines perform linear interpolations between
adjacent table values to estimate a force and energy
the specified size is the length of the table, the larger the number
the more accurate the table and the more memory used
10000 is a typical table length
a value of 0 means no tables are used, the forces and energies are
evaluated in the usual analytic fashion
default: 0
----------------------------------------------------------------------------
# neighbor
1st parameter = skin thickness of neighbor shell
2nd parameter = delay check for neighbor re-build
until this many steps since last build
factors that affect how and when neighbor lists are constructed
communication is done to acquire all atoms within distance cutoff + skin
after delay # of steps, check is done every timestep based on
atom movement to see if neighbor list should be rebuilt -
e.g. delay = 5 means start checking 5 timesteps after last rebuild
must be specified before "create box" command
default: skin = 0.3, delay = 10
----------------------------------------------------------------------------
# no neighbor (ONLY available with CRACK compiler flag)
1st parameter = type A
2nd parameter = type B
turn-off force interactions between atom pairs of types A and B
can be used to create an initial slit crack
works because neighbor-list formation routine checks for these interactions
setting both types to 0 essentially turns off this command
default = 0 0
----------------------------------------------------------------------------
# processor grid
parameters = # of processors in x,y,z dimensions
specify 3-d grid of processors to map to physical simulation domain
total # of processors must equal product of 3 parameters
must be specified before "create box" command
if you do not specify this, code will make best guess of how to map
procs to the box, so as to minimize the surface area of proc sub-domains
if the 1st param is zero, then the code reverts to its default assignment
and you don't need the other 2 parameters
default = code matches procs to simulation box
----------------------------------------------------------------------------
# thermo
print thermodynamic info to screen and log file every this many timesteps
value of 0 means print only at beginning and end of a run
default = 0
----------------------------------------------------------------------------
# snapshot
1st parameter = call snapshot routine every this many steps
2nd parameter = filename to dump shapshots to
3rd parameter = "xmovie" or "pds" or "ensight6bin" or "ensight6text" or
"ensight6binc"
dump snapshot of atom positions to a file every this many timesteps
value of 0 means never dump
non-zero value also writes snapshot at start and end of run
any previous snapshot file is closed
new filename can exist, will be overwritten
snapshot info is dumped in one of 4 formats:
"xmovie" format, which is ASCII text with descriptors for timestep,
number of atoms, bounding box, and one line per atom
"pds" format writes snapshots in Sandia's PDS format, which is
portable binary format, the tools/pds2xmovie.f file can be used
to extract XMOVIE snapshots from a PDS file
"ensight6bin"/"ensight6binc" format writes case and variable file
for the visualization software Ensight. Variable file is written
in Fortran/C binary form, one variable file for all time steps
"ensight6text" writes case and variable files for the visualization
software Ensight, one variable file for one time step
if the 3rd parameter is omitted, default "ensight6bin" format is used
info in snapshot file is affected by "snap thresh" and "snap column" commands
default = 0
----------------------------------------------------------------------------
# snap thresh
1st parameter = "none" or "energy" or "centro"
2nd parameter = value
only dump an atom to the snapshot file if a value associated with the atom
is larger than the specified threshhold value
choices for the value are "none" = all atoms are dumped out in which case
the 2nd parameter need not be specified, "energy" = potential energy
of the atom, and "centro" = centrosymmetry value of the atom
energy and distance^2 units for the threshhold value are in appropriate
real or lj units
can be used to throttle the output to only defect atoms
default = none
----------------------------------------------------------------------------
# snap column
1st parameter = 0,1,2,3,4,5,6,7
add columns of output to the snapshot file for each dumped atom
for parameter = 0, no additional info beyond the atom type and xyz coords is
dumped
for parameter = 1, the atom's potential energy is also dumped
for parameter = 2, the atom's centrosymmetry value is also dumped
for parameter = 3, both the atom's pot eng and centrosymmetry values are dumped
for parameter = 4, the atom's stress is dumped
for parameter = 5, the atom's stress and pot eng are dumped
for parameter = 6, the atom's stress and centrosymmetry are dumped
for parameter = 7, the atom's stress, centrosymmetry and pot eng are dumped
default = 0
----------------------------------------------------------------------------
# snap limits
6 parameters = xlo xhi ylo yhi zlo zhi (in cubic lattice units)
only dump an atom to the snapshot file if its coordinates lie within the
specified region.
for a simulation with variable domain size (e.g. constant pressure), this
region grows in proportion with the box size
default = inf inf inf inf inf inf
----------------------------------------------------------------------------
# restart
1st parameter = write a restart file every this many steps
2nd parameter = prefix file name to write to
create a restart file every this many timesteps
full filename is prefix.timestep
value of 0 means never create one (don't need filenames in this case)
restart file stores atom positions and velocities in binary form
allows program to restart from where it left off via "read restart" commmand
new restart files overwrite any previous files
restart files are written in one of 2 formats, depending on which
file is compiled into the code: restart.f or restart_pds.f
the restart.f file writes in a machine-specific binary format
the restart_pds.f file writes in Sandia's PDS format, which is
portable binary format
default = 0
----------------------------------------------------------------------------
# write atoms
1st parameter = write an atom file every this many steps
2nd parameter = prefix file name to write to
create an atom file every this many timesteps
full filename is prefix_timestep.atoms
value of 0 means never create one (don't need filenames in this case)
atom file stores header with time step, number of atoms and box bounds
also, element type, atom type and position are stored for every atom
new atom files overwrite any previous files
default = 0
----------------------------------------------------------------------------
# write dynamo
1st parameter = write a dynamo format restart file every this many steps
2nd parameter = prefix file name to write to
create a dynamo format file every this many timesteps
full filename is prefix_timestep.dynamo
value of 0 means never create one (don't need filenames in this case)
dynamo format restart file stores:
comment line
numatoms, numelementypes, constant, angle
box bounds
atomic masses and atomic numbers
also, for each atom it stores:
position, velocity and type
new dynamo files overwrite any previous files
default = 0
----------------------------------------------------------------------------
# write meamstat
1st parameter = write a dynamo format meamstat file every this many steps
2nd parameter = prefix file name to write to
create a dynamo format meamstat file every this many timesteps
full filename is prefix_timestep.meamstat
value of 0 means never create one (don't need filenames in this case)
dynamo format meamstat file stores:
atomic data:number, type, positions, energy
velocities
forces
el. density, f(rho), phi(r)
rho0 rho(1)^2 rho(2)^2 rho(3)^2
new meamstat files overwrite any previous files
default = 0
----------------------------------------------------------------------------
# numatoms
number of atoms - if specified before create box, it saves memory, otherwise
maximum number of atoms is estimated from lattice constant
and box size
default = 0
----------------------------------------------------------------------------
# diagnostic
1st parameter = identifier name of routine
2nd parameter = call diagnostic routine every this many steps
3rd parameter = filename to dump diagnostic results to
4th parameter = # of remaining parameters (maximum of 5)
5th-9th parameters = parameters to pass to diagnostic routine
call a user-specified diagnostic routine and write results into file
1st parameter is a string that is checked inside diagnostic.f
in the diagalloc routine when a run is performed
if 2nd parameter = 0, turn off diagnositcs (no need for rest of parameters)
non-zero value also calls routine at start and end of run
any previous diagnostic file is closed
new filename can exist, will be overwritten
4th parameter can be from 0 to 5
5th-9th parameters are visible by diagnostic routine
default = 0
----------------------------------------------------------------------------
# diag types
whether to write a separate diagnostic file for each type
0 = write one diagnostic file providing average for all types
1 = write separate diagnostic files for each atom type
default = 0
----------------------------------------------------------------------------
# periodicity
whether each x,y,z dimension is periodic or not
0 = non-periodic, 1 = periodic
if non-periodic, forces do not interact across the boundary,
atoms do not migrate across the boundary, and ghost atoms are not
created across the boundary
if non-periodic, atoms will be "lost" if they move outsize the simulation
box specified by the "create box" command
must be specified before "create box" command
default = 1 1 1
----------------------------------------------------------------------------
# lattice
1st parameter = fcc
2nd parameter = density or lattice constant of cubic unit cell
what type and what spacing will be used for the fundamental lattice of atoms
fcc = 4 atoms per cubic unit cell
2nd parameter determines how closely the atoms are packed on the lattice
for lj units, 2nd parameter is reduced density, from which a lattice
constant is derived
for real units, 2nd parameter is the cubic lattice constant in Angstroms
must be specified before "create box" command
default = fcc and 1.0
----------------------------------------------------------------------------
# create box
parameters = xlo xhi ylo yhi zlo zhi
bounds of the global simulation box in cubic lattice units
for example, 0 10 0 10 0 10 = a box of size 10x10x10 cubic unit cells of atoms
which means 4000 atoms (in a 100 orientation) will fit in the box
for a fcc lattice (4 atoms/cubic unit cell)
if non-periodic boundaries are specified, you must insure the box is
large enough that atoms will never leave the box during the course
of the simulation, else they will be lost
unlike the other "create" commands, this one does not work in conjunction
with "select" commands, all 6 parameters must be explicitly specified
when this command is executed, the processor grid is layed out, the global
box is partitioned across procs, neighbor list and communication data
structures are setup, and all memory is allocated for the simulation
re-issuing this command essentially starts a simulation from scratch, all old
information (including atoms) is deleted
----------------------------------------------------------------------------
# orient
1st parameter = "x" or "y" or "z" = box direction
2nd-4th parameters = lattice orientation that is along this box direction
(e.g. 111 or 110 or 1,1,-2)
specify orientation of lattice (in cubic sense) along box directions (xyz)
3 basis vectors must be mutually orthogonal and be a right-handed system
such that (X cross Y) is in same direction as Z
each orientation vector should be in irreducible form (minimum integers)
these vectors are used when the "create atoms" command generates a lattice
of atoms
default: x = 100, y = 010, z = 001
----------------------------------------------------------------------------
# origin
x,y,z origin of a cubic lattice in lattice units
for example, 0.5 0.5 0.0 means a lattice displaced by 1/2 a cubic cell
in both the x and y box dimensions
this origin is used as a starting point when the "create atoms" command
generates a lattice of atoms
default = 0.0 0.0 0.0
----------------------------------------------------------------------------
# select region
one parameter = all (entire simulation box)
6 parameters = xlo xhi ylo yhi zlo zhi (in cubic lattice units)
this command must be used prior to each "create atoms" or "create types"
command to specify the volume of interest
if used prior to a "create atoms" command, means to select a
rectangular volume to create atoms within
if used prior to a "create types" command, means to select all
previously created atoms within a rectangular volume
if a single parameter "all" is used, this selects the entire box
if 6 parameters are used, they specify the bounds of the rectangular region
in cubic lattice units (NOT in reduced units of sigma or in Angstroms)
if "inf" is used as one of the 6 parameters, it means extend the bound
to the global simulation box boundary in that direction
a specified parameter must be contained in the simulation box
this command must be used after the "create box" command
----------------------------------------------------------------------------
# select type
this command must be used prior to each "create vels" or "create fixes"
command to specify the atoms of interest
selects all previously created atoms of the specified type
----------------------------------------------------------------------------
# set eltype
this command may be used prior to "create atoms" to set the element type
of the atoms to be created.
----------------------------------------------------------------------------
# define cutout
if desired, this command must be used prior to each "create atoms" or
"create types" command to define a cutout volume (inside or outside)
that will be intersected with the selected region to form a new complex
volume that will be used by the "create atoms" and "create types" commands
examples:
define cutout 0 region xlo xhi ylo yhi zlo zhi
define cutout 1 cylinder x y z c0 c1 rad zlo zhi
define cutout 0 sphere cx cy cz rad
define cutout 1 cylshell x y z c0 c1
radouter radinner zlo zhi
the 1st argument is 0 if the "inside" of the cutout
volume should be included and the "outside" excluded, the 1st argument
is 1 if the "outside" of the cutout volume should be included and the
"inside" excluded
for the region option, a rectangular volume is defined as the cutout volume,
the meaning of the 6 parameters is the same as for the "select region"
command, "inf" can be used for any of the 6 parameters
for the cylinder option, a cylindrical volume is defined as the cutout
volume, the x y z parameters define the orientation of the cylinder
with respect to the simulation box directions and the other 5 parameters
define the size of the cylinder: c0 and c1 are the coordinates of the
center axis of the cylinder in the 1st 2 dimensions, rad is its radius,
and the cylinder extends along its axis in the 3rd dimension from
zlo to zhi where either of zlo,zhi may be "inf" which means extend
to the simulation box boundary - e.g. "cylinder y z x 0.0 2.0 1.0 -5.0 inf"
means the cylinder axis lies along the x direction of the simulation box
at y = 0.0 and z = 2.0, the cylinder has a radius of 1.0, and it extends
from x = -5.0 to the upper +x box boundary
for the sphere option, a spherical volume is defined as the cutout volume,
the next 3 parameters are the origin of the sphere, and the last parameter
is the radius
for the cylshell option, the meaning of all arguments is the same as
for the cylinder option, except that 2 radii are defined, radouter
and radinner, the created atoms lie between (or outside of) the 2 radii
default = none
----------------------------------------------------------------------------
# create atoms
parameter = type of atoms to be created
create a lattice of atoms so as to tile the entire volume specified by
the "select region" and "define cutout" commands
the atoms are created at an orientation and from an origin specified
by the most recent "orient" and "origin" commands
the created atoms are assigned a velocity of 0.0
----------------------------------------------------------------------------
# scale read
parameters = x, y, and z direction scale factors
scale all atomic coordinates read in by the given factor in each
direction. Note that this is only applied to read-in, not created
atoms, and must be specified before the atom file is read.
----------------------------------------------------------------------------
# type translate
translate a given atom type in the atom coord file to another type in Warp
1st parameter = atom type in atom coord file
2nd parameter = atom type in Warp
This command is convenient when the atom coord file lists atoms by atomic
number (e.g. Ni=28, Al=13, H=1) and these must be converted to another
enumeration (e.g. Ni=1, Al=2, H=3 for the nialhjea EAM potential file)
----------------------------------------------------------------------------
# box toler
distance determining the tolerance of simulation box size
must be positive number
only significant if atoms are read from file
must be specified before "read atoms" command
default = 1.0E-14
----------------------------------------------------------------------------
# read atoms
parameter = file to read atom coords from
read a single XMOVIE snapshot from a file to initialize all the atoms
for the simulations
takes the place of "create atoms" or "read restart" command
atom coords and types are extracted from the file
box size specified in snapshot is assumed to be equal or smaller
than box specified by "create box" command
all atom coords in file are used as-is and must fit in simulation box,
are not checked for periodic boundary conditions
atom coords are NOT in cubic lattice units, but in simulation box units,
e.g. Angstroms for real units or reduced units for lj units
the created atoms are assigned a velocity of 0.0
----------------------------------------------------------------------------
# check overlap
check all atoms for too-close an overlap with neighboring atoms
and delete one atom in the overlapping pair
1st parameter = overlap occurs within this distance (in cubic lattice units)
2nd parameter = type of atom to delete if overlap occurs
this command is useful if previous "create atoms" commands may have
created duplicate atoms on region boundaries or created 2 atoms too close
together (e.g. at a grain boundary)
the specified type is used to choose which atom of the overlapped pair
can be deleted
all atom pairs are checked and the code attempts to delete one atom in
an overlapping pair if it is of the specified type
if neither or both of the overlapping atoms can be deleted, various errors
and warnings may be generated if the code cannot uniquely choose which
atom to delete, the only solution for this may be to
use the "select" commands more carefully with epsilon increments to
prevent atom overlap at creation time
using a type of 0 will flag errors or warnings for all overlaps since
neither atom in the overlapped pair can be deleted
this command can be done anytime after atoms are created
----------------------------------------------------------------------------
# create types
parameter = type to be assigned
assign the specified type to all atoms within the entire volume specified by
the "select region" and "define cutout" commands
----------------------------------------------------------------------------
# create eltypes
parameter = element type to be assigned
assign the specified element type to all atoms within the entire volume
specified by the "select region" and "define cutout" commands
note that currently multiple element types are available only for the
MEAM potential
----------------------------------------------------------------------------
# displace atoms
displace all atoms selected by the "select type" command to new positions
useful for setting up an initial strain on the system
examples:
displace atoms dt xyz vx vy vz
displace atoms dt ramp vy vlo vhi x clo chi
displace atoms dt twist vlo vhi x y z c0 c1 clo chi
displace atoms dt expand 1 1 0 cx cy cz doubletime
for all options, a "select type" must be in effect to first select
atoms of a particular type
for all options, a velocity is computed for each selected atom, exactly
as it would be by the "create vels" command for that option,
the velocity is not assigned to the atom, rather the atom is moved
to a new position as if that velocity were applied for a time dt
for all options, the velocity parameters and their meaning and units are
exactly the same as they are for the corresponding "create vels" command
for the xyz option, the velocity for each atom is the specified vx,vy,vz
for the twist option, the new atom coordinate is computed by rotating
the atom to a new coord, not by moving the atom in a straight line
the units of dt are in tau (lj units) or picoseconds (real units)
for the xyz option, the velocity units are cubic lattice units
per time - e.g. lattice-spacings/tau or lattice-spacings/picosecond
----------------------------------------------------------------------------
# check vels
check or do not check velocities
0 = don't check velocities
1 = check velocities
if check velocities, vhi must be greater than vlo
if don't check velocities, vhi can be lower than vlo, thus compression
can be applied
default = 1
----------------------------------------------------------------------------
# snap types
whether to snapshot types or eltypes
0 = snapshot eltypes
1 = snapshot types
default = 0
----------------------------------------------------------------------------
# recalculate output
whether to rebuild neighbors and recalculate values before output
0 = don't reneighbor before output
1 = reneighbor and recalculate values before output
default = 0
----------------------------------------------------------------------------
# create vels
assign initial velocities to all atoms selected by the "select type" command
this command adds a new velocity to a current initial velocity for each atom,
thus multiple "create vels" commands can be used to build up a
desired velocity distribution - e.g. a ramp profile superposed on a
thermal background
examples:
create vels thermal 300.0 seed
create vels ramp vy vlo vhi x clo chi
create vels twist vlo vhi x y z c0 c1 clo chi
create vels expand 1 1 0 cx cy cz doubletime
for all options, a "select type" must be in effect to first select
atoms of a particular type
thermal option means create a Boltzmann distribution at the specified
temperature using a random number generator initialized with the
specified integer seed
for the thermal option the units of temperature are reduced units or
degrees Kelvin depending on the "units" command previously specified
for the thermal option, the center-of-mass motion of the selected atoms
will be zeroed
ramp option means assign velocities ranging from vlo -> vhi to a
specfied velocity component (vx or vy of vz) of the selected atoms
for the ramp option, the assigned velocity value depends on where the
specified coordinate (x or y or z) of the atom falls
within the specified coordinate clo -> chi bounds - e.g. for a clo -> chi
region ranging from -10.0 to 10.0 in "x" and a vlo -> vhi specification
of 0.0 -> 5.0 for "vy", then a selected atom with an x-coord of 0.0 will
be assigned a vy velocity of 2.5
for the ramp option, an atom whose coordinate falls outside the specified
bounds will be given the min or max velocity
for the ramp option, if vlo = vhi then all the selected atoms will be
assigned the same velocity, however clo and chi must still be specified
for the ramp option, clo must be smaller than chi
for the ramp option the units of vlo and vhi are in cubic lattice units
per time - e.g. lattice-spacings/tau or lattice-spacings/picosecond -
and the units of clo and chi are in cubic lattice units as well
twist option means assign velocities to selected atoms in a cylinder so that
the cylinder initially spins about its axis at a rotation rate varying
from vlo -> vhi from one end to the other, the cylinder is centered
on c0,c1 in a coordinate system defined by x y z and the velocities
are applied along the cylinder's z axis from clo to chi
for the twist option the units of vlo and vhi are in rotations per time -
e.g. rotations/tau or rotations/picosecond depending on the "units" flag -
and the units of c0, c1, clo, and chi are in cubic lattice units
for the twist option the x,y,z specification defines which box directions
correspond to the x,y plane and z-axis of the cylinder
for the twist option, the assigned velocity depends on where the atom's
coordinates fall within the cylinder - e.g. for the command
"create vels twist 0.0 0.1 y z x 0.0 2.0 -10.0 10.0" the cylinder lies
along the x-axis from x = -10 to 10 and the circular cross-section
of the cylinder is the y-z plane with the center axis at y = 0.0, z = 2.0 -
each atom will be given 0.0 velocity in x, and a y-z velocity that will
rotate the atom around the cylinder axis at a rotational rate from 0.0 to
0.1 revolutions/time which depends on where the atom's x-coord lies between
-10.0 and 10.0
for the twist option, if vlo = vhi then all the selected atoms will be
assigned the same rotational velocity, however clo and chi must still
be specified
expand option means assign velocities to selected atoms in a radial direction
away from (or towards) a center point so that the system expands
(or contracts) at a uniform rate - useful for biaxial stress simulations
for the expand option, the 1st 3 parameters are flags on the x,y,z
dimensions for whether to include them or not in radial direction
computation - a value of 0 means do not include, a value of 1 means include
for the expand option, the radial distance r of each selected atom from the
center point (cx,cy,cz) is computed using only included dimensions -
the atom is given a velocity in the radial direction of magnitude
alpha*r where alpha is computed from doubletime factor (see units below)
for the expand option the units of cx, cy, cz are cubic lattice units
and the units of doubletime = time in tau (lj units) or picoseconds
(real time) to double the linear dimension of the system - e.g. a
value of doubletime = 1000.0 means each atom will be 2x farther from
the center point after 1000 tau or picoseconds
for the expand option, a negative doubletime means compression instead
of expansion, i.e. the time for the system to implode to a single point
----------------------------------------------------------------------------
# create fixes
assign a force fix to all atoms selected by the "select type" command
force fixes are applied each timestep after forces are computed and
communicated, but before they are used to update velocities
examples:
create fixes none
create fixes xyz 0 1 -1
create fixes twist x y z c0 c1 1 1 1
create fixes external 0.1 0.0 0.0
none option means erase all previously specified force fixes
for all other options, a "select type" must be in effect to first select
atoms of a particular type
each time the "create fixes" command is used, a new fix is added,
so the only way to change a previously specified fix, is to use the
"none" option to erase them all, then re-specify the desired fixes
xyz option means to not-change, zero-out, or average the force on
each x, y, z component of force for the selected atoms, a flag
is used to specify for each of the components whether it should be
unchanged (-1) or zeroed (0) or averaged (1) - e.g. the command
"create fixes xyz 0 -1 1" will zero the x component of force on each atom,
not alter the y force component, and compute the total z force on all
selected atoms, divide it by the number of selected atoms,
and then set the z force component of each atom to the same average force
for the twist option, the 3 final flags refer to the radial,
theta (azimuthal), and z-components of force for the specified cylinder
twist option means to add a centripetal force to each selected atom
so as to keep it rotating at its current angular velocity in the
specified cylinder, additionally a flag is used to specify for each of
the force components in the cylinder frame-of-reference whether it
should be unchanged (-1) or zeroed (0) or averaged (1)
for the twist option the x,y,z specification defines which box directions
correspond to the x,y plane and z-axis of the cylinder
for the twist option, unchanged (-1) means do not alter the component
unless it is the radial component which always has a centripetal term
added, zero (0) means zero out the component unless it is the radial
one in which case the centripetal term will remain, and average (1)
can only be applied to the z-component of the cylinder and means to
compute the total force on all selected atoms, divide it by the
number of selected atoms, and then set the force component along the
cylinder axis for each atom to the same average force
an example for the twist option is
"create fixes twist y z x 0.0 2.0 0 0 1" which will treat the line in the
the x-dimension at y = 0.0, z = 2.0 (in cubic lattice units) as the axis
of a cylinder, the Cartesian x,y,z force on each selected atom will be
recast in r,theta,z components depending on the atoms current position
relative to the cylinder axis, the atom's theta velocity will be computed,
each of the 3 recast force components will be zeroed if the corresponding
flag is set to 0 (in this case the radial and azimuthal forces will be
zeroed, the z-component along the cylinder axis will be averaged),
a centripetal v_theta**2/r term will be added to the radial force,
and the forces are mapped back to the Cartesian x,y,z form
external option means simply add the given external forces (input in the x, y,
and z directions) to the already-computed forces. Note that this does not
zero out the forces before adding the given force. Units are eV/A.
IMPORTANT NOTE: the "temp control" command can also alter
atom velocities, so may want to use "temp type" to exclude
fixed atoms from the control
----------------------------------------------------------------------------
# temp type
compute temperature only using atoms with types <= specified type
can be used to exclude constrained atoms from temperature calculation
and temperature control in "temp control" command
value of 0 means use all atoms to compute temperature
must be done after "create box" command
default = 0
----------------------------------------------------------------------------
# temp control
which style of temperature control to use
examples:
temp control none
temp control rescale 300.0 20 10.0
temp control hoover 300.0 10.0
none means constant NVE -> no temperature control
rescale means rescale the temperature to exactly the desired value
every so many timesteps
hoover means do true NVT temperature control via Nose/Hoover thermostat
1st parameter for rescale is desired temperature in LJ or real units
2nd parameter for rescale is every how many timesteps to do rescaling
3rd parameter for rescale is temperature window, rescaling is only
done if temperature is outside this +/- window
1st parameter for hoover is desired temperature in LJ or real units
2nd parameter for hoover is a frequency constant for the damping which
is like an inverse "piston" mass, it determines how rapidly the
temperature fluctuates in response to a restoring force,
large frequency -> small mass -> rapid fluctations
for hoover, units of frequency/damping constant are inverse time,
so a value of 0.1 means relax in a timespan on the
order of 10 tau (lj units) or 10 picoseconds (real units)
all controls are only performed on atoms with types <= specified type
in "temp type" command
must be done after "create box" command
default = none
----------------------------------------------------------------------------
# temp adjust
adjust the temperature calculation and control to compensate for
non-thermal components of velocity
example:
temp adjust none
temp adjust xyz 0 1 1
temp adjust ramp vy vlo vhi x clo chi
temp adjust twist vlo vhi x y z c0 c1 clo chi
before computing a temperature and before modifying an atom's velocity
via the "temp control" command, its velocities are adjusted to compensate
for non-thermal velocity component(s)
if the adjustment subtracts out a particular velocity component(s),
then the subtracted component(s) are added back in after the
"temp control" is performed
none means do no adjustment
xyz option means only the specified components contribute to the velocity
and are controlled, a 0 means do not include that component, a 1 means
do include that component
ramp option means subtract a particular velocity component based on the
atom's current coordinate between 2 bounds in a particular dimension,
an atom whose coordinate falls outside the specified bounds will be
adjusted by the min or max velocity
for the ramp option the parameters and their meaning and units are
exactly the same as they are for the "create vels ramp" command
twist option means subtract out angular velocity in the two components
orthogonal to the cylinder axis, depending on where the atom's current
coordinates lie along the cylinder axis, an atom whose coordinate falls
beyond the end of the cylinder will be adjusted by the min or max rotation
for the twist option the parameters and their meaning and units are
exactly the same as they are for the "create vels twist" command
all adjustments are only performed on atoms with types <= specified type
in "temp type" command
must be done after "create box" command
default = none
----------------------------------------------------------------------------
# pres control
apply pressure control
example:
pres control none
pres control rescale 0.0 0.5
none means constant volume -> no pressure control
rescale means rescale the box volume to maintain pressure at the desired
value. The volume is an extra degree of freedom with a corresponding
spring-mass-damper type equation.
1st parameter for rescale is desired pressure in bars.
2nd parameter for rescale is time constant in ps for the volume equation.
WARNING: currently pressure control cannot be used with restarts, since
the box size changes when pressure control is turned on. The box
size in the restart file will not match that in the command input
file.
----------------------------------------------------------------------------
# read restart
read atom coords, velocities, types from specified file
allows continuation of a previous run via "restart" command
file is binary to enable exact restarts, assuming simulation box is
created in the same fashion
do not have to restart on same # of processors, but can only do exact
restarts on same # of processors
must do a "create box" with all associated settings before reading a
restart file
after reading a restart file, do not use the "create atoms" and
"create vels" commands
still do "create fixes" after reading a restart file since fixes are
not stored in the file
----------------------------------------------------------------------------
# reset timestep
reset the global timestep # to the specified value
useful to do after equilibration run, before data-gathering run
----------------------------------------------------------------------------
# run
run or continue MD for specified # of timesteps
must have performed "create box" and "create atoms" commands first
----------------------------------------------------------------------------
# relax
minimize potential energy using a conjugate gradient solver; repeated for
a specified number of timesteps (between which atoms with fixed velocities
are updated, although solution is quasistatic)
----------------------------------------------------------------------------
# relaxp
relax with parameters
example:
relaxp 1 1500 4 40 1e-8 1e-6 1e-10
1st parameter is number of steps (atoms with fixed velocities are updated)
2nd parameter = "imax" is max number of cg iterations
3rd parameter = "jmax" is number of secant method iterations
4th parameter = "kres" number of iterations to restart nonlinear cg
5th parameter = "eps" cg error tolerance (< 1)
6th parameter = "sigma" secant method step parameter
7th parameter = "res_limit" smallest residual value worthy to improve
introduced to ensure numerical stability of relaxation
----------------------------------------------------------------------------
# relaxf
relax with parameters from a specified file
example:
relaxf 1 RELAX
1st parameter is number of steps (atoms with fixed velocities are updated)
other six parameters (2nd to 7-th parameter of relaxp) are read from a
specified file, in the same order as relaxp parameters
example RELAX file:
2000 ! imax
4 ! jmax
40 ! kres
1e-8 ! eps
1e-6 ! sigma
1e-9 ! res_limit
----------------------------------------------------------------------------
----------------------------------------------------------------------------
