ICODE-ADC/lammps
4
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

Resolving merge conflicts

Browse Source
This commit is contained in:
Joel Clemmer
2020-12-26 13:58:24 -07:00
parent a88fab7558 ab5a1f229e
commit 795bd40318
102 changed files with 5407 additions and 954 deletions
Download Patch File Download Diff File Expand all files Collapse all files

8
doc/src/Build_extras.rst
View File

@ -352,7 +352,7 @@ KIM Extra unit tests (CMake only)
During development, testing, or debugging, if
:doc:`unit testing <Build_development>` is enabled in LAMMPS, one can also
enable extra tests on :doc:`KIM commands <kim_commands>` by setting the
``KIM_EXTRA_UNITTESTS`` to *yes* (or *on*).
``KIM_EXTRA_UNITTESTS`` to *yes* (or *on*).
Enabling the extra unit tests have some requirements,
@ -367,9 +367,9 @@ Enabling the extra unit tests have some requirements,
*conda-forge* channel as ``conda install kim-property`` if LAMMPS is built in
Conda. More detailed information is available at:
`kim-property installation <https://github.com/openkim/kim-property#installing-kim-property>`_.
* It is also necessary to install
``EAM_Dynamo_Mendelev_2007_Zr__MO_848899341753_000``, and
``EAM_Dynamo_ErcolessiAdams_1994_Al__MO_123629422045_005`` KIM models.
* It is also necessary to install
``EAM_Dynamo_Mendelev_2007_Zr__MO_848899341753_000``, and
``EAM_Dynamo_ErcolessiAdams_1994_Al__MO_123629422045_005`` KIM models.
See `Obtaining KIM Models <http://openkim.org/doc/usage/obtaining-models>`_
to learn how to install a pre-build binary of the OpenKIM Repository of
Models or see

2
doc/src/Commands_fix.rst
View File

@ -221,6 +221,8 @@ OPT.
* :doc:`temp/rescale <fix_temp_rescale>`
* :doc:`temp/rescale/eff <fix_temp_rescale_eff>`
* :doc:`tfmc <fix_tfmc>`
* :doc:`tgnpt/drude <fix_tgnh_drude>`
* :doc:`tgnvt/drude <fix_tgnh_drude>`
* :doc:`thermal/conductivity <fix_thermal_conductivity>`
* :doc:`ti/spring <fix_ti_spring>`
* :doc:`tmd <fix_tmd>`

1
doc/src/Developer.rst
View File

@ -13,6 +13,7 @@ of time and requests from the LAMMPS user community.
Developer_org
Developer_flow
Developer_write
Developer_notes
Developer_unittest
Classes
Developer_utils

105
doc/src/Developer_notes.rst Normal file
View File

@ -0,0 +1,105 @@
Notes for Developers and Code Maintainers
-----------------------------------------
This section documents how a few large sections of code with LAMMPS
work at a conceptual level. Comments on code in source files
typically document what a variable stores, what a small section of
code does, or what a function does or its input/outputs. The topics
on this page are intended to document code at a higher level.
KSpace PPPM FFT grids
^^^^^^^^^^^^^^^^^^^^^
The various :doc:`KSpace PPPM <kspace_style>` styles in LAMMPS use
FFTs to solve Poisson's equation. This subsection describes:
* how FFT grids are defined
* how they are decomposed across processors
* how they are indexed by each processor
* how particle charge and electric field values are mapped to/from
the grid
An FFT grid cell is a 3d volume; grid points are corners of a grid
cell and the code stores values assigned to grid points in vectors or
3d arrays. A global 3d FFT grid has points indexed 0 to N-1 inclusive
in each dimension.
Each processor owns two subsets of the grid, each subset is
brick-shaped. Depending on how it is used, these subsets are
allocated as a 1d vector or 3d array. Either way, the ordering of
values within contiguous memory x fastest, then y, z slowest.
For the ``3d decomposition`` of the grid, the global grid is
partitioned into bricks that correspond to the sub-domains of the
simulation box that each processor owns. Often, this is a regular 3d
array (Px by Py by Pz) of bricks, where P = number of processors =
Px * Py * Pz. More generally it can be a tiled decomposition, where
each processor owns a brick and the union of all the bricks is the
global grid. Tiled decompositions are produced by load balancing with
the RCB algorithm; see the :doc:`balance rcb <balance>` command.
For the ``FFT decompostion`` of the grid, each processor owns a brick
that spans the entire x dimension of the grid while the y and z
dimensions are partitioned as a regular 2d array (P1 by P2), where P =
P1 * P2.
The following indices store the inclusive bounds of the brick a
processor owns, within the global grid:
.. parsed-literal::
nxlo_in,nxhi_in,nylo_in,nyhi_in,nzlo_in,nzhi_in = 3d decomposition brick
nxlo_fft,nxhi_fft,nylo_fft,nyhi_fft,nzlo_fft,nzhi_fft = FFT decomposition brick
nxlo_out,nxhi_out,nylo_out,nyhi_out,nzlo_out,nzhi_out = 3d decomposition brick + ghost cells
The ``in`` and ``fft`` indices are from 0 to N-1 inclusive in each
dimension, where N is the grid size.
The ``out`` indices index an array which stores the ``in`` subset of
the grid plus ghost cells that surround it. These indices can thus be
< 0 or >= N.
The number of ghost cells a processor owns in each of the 6 directions
is a function of:
.. parsed-literal::
neighbor skin distance (since atoms can move outside a proc subdomain)
qdist = offset or charge from atom due to TIP4P fictitious charge
order = mapping stencil size
shift = factor used when order is an even number (see below)
Here is an explanation of how the PPPM variables ``order``,
``nlower`` / ``nupper``, ``shift``, and ``OFFSET`` work. They are the
relevant variables that determine how atom charge is mapped to grid
points and how field values are mapped from grid points to atoms:
.. parsed-literal::
order = # of nearby grid points in each dim that atom charge/field are mapped to/from
nlower,nupper = extent of stencil around the grid point an atom is assigned to
OFFSET = large integer added/subtracted when mapping to avoid int(-0.75) = 0 when -1 is the desired result
The particle_map() method assigns each atom to a grid point.
If order is even, say 4:
.. parsed-literal::
atom is assigned to grid point to its left (in each dim)
shift = OFFSET
nlower = -1, nupper = 2, which are offsets from assigned grid point
window of mapping grid pts is thus 2 grid points to left of atom, 2 to right
If order is odd, say 5:
.. parsed-literal::
atom is assigned to left/right grid pt it is closest to (in each dim)
shift = OFFSET + 0.5
nlower = 2, nupper = 2
if point is in left half of cell, then window of affected grid pts is 3 grid points to left of atom, 2 to right
if point is in right half of cell, then window of affected grid pts is 2 grid points to left of atom, 3 to right
These settings apply to each dimension, so that if order = 5, an
atom's charge is mapped to 125 grid points that surround the atom.

5
doc/src/Howto_drude.rst
View File

@ -42,10 +42,11 @@ screening. It may be necessary to use the *extra/special/per/atom*
keyword of the :doc:`read_data <read_data>` command. If using :doc:`fix shake <fix_shake>`, make sure no Drude particle is in this fix
group.
There are two ways to thermostat the Drude particles at a low
There are three ways to thermostat the Drude particles at a low
temperature: use either :doc:`fix langevin/drude <fix_langevin_drude>`
for a Langevin thermostat, or :doc:`fix drude/transform/\* <fix_drude_transform>` for a Nose-Hoover
thermostat. The former requires use of the command :doc:`comm_modify vel yes <comm_modify>`. The latter requires two separate integration
thermostat, or :doc:`fix tgnvt/drude <fix_tgnh_drude>` for a temperature-grouped Nose-Hoover thermostat.
The first and third require use of the command :doc:`comm_modify vel yes <comm_modify>`. The second requires two separate integration
fixes like *nvt* or *npt*\ . The correct temperatures of the reduced
degrees of freedom can be calculated using the :doc:`compute temp/drude <compute_temp_drude>`. This requires also to use the
command *comm_modify vel yes*.

46
doc/src/Howto_drude2.rst
View File

@ -221,6 +221,14 @@ modification of forces but no position/velocity updates), the fix
fix NVE all nve
To avoid the flying ice cube artifact, where the atoms progressively freeze and the
center of mass of the whole system drifts faster and faster, the *fix momentum*
can be used. For instance:
.. code-block:: LAMMPS
fix MOMENTUM all momentum 100 linear 1 1 1
Finally, do not forget to update the atom type elements if you use
them in a *dump_modify ... element ...* command, by adding the element
type of the DPs. Here for instance
@ -376,14 +384,7 @@ For our phenol example, the groups would be defined as
Note that with the fixes *drude/transform*\ , it is not required to
specify *comm_modify vel yes* because the fixes do it anyway (several
times and for the forces also). To avoid the flying ice cube artifact
:ref:`(Lamoureux and Roux) <Lamoureux2>`, where the atoms progressively freeze and the
center of mass of the whole system drifts faster and faster, the *fix
momentum* can be used. For instance:
.. code-block:: LAMMPS
fix MOMENTUM all momentum 100 linear 1 1 1
times and for the forces also).
It is a bit more tricky to run a NPT simulation with Nose-Hoover
barostat and thermostat. First, the volume should be integrated only
@ -404,6 +405,31 @@ instructions for thermostatting and barostatting will look like
fix NVT DRUDES nvt temp 1. 1. 20
fix INVERSE all drude/transform/inverse
Another option for thermalizing the Drude model is to use the
temperature-grouped Nose-Hoover (TGNH) thermostat proposed by :ref:`(Son) <TGNH-SON>`.
This is implemented as :doc:`fix tgnvt/drude <fix_tgnh_drude>` and :doc:`fix tgnpt/drude <fix_tgnh_drude>`.
It separates the kinetic energy into three contributions:
the molecular center of mass (COM) motion, the motion of atoms or atom-Drude pairs relative to molecular COMs,
and the relative motion of atom-Drude pairs.
An independent Nose-Hoover chain is applied to each type of motion.
When TGNH is used, the temperatures of molecular, atomic and Drude motion can be printed out with :doc:`thermo_style` command.
NVT simulation with TGNH thermostat
.. code-block:: LAMMPS
comm_modify vel yes
fix TGNVT all tgnvt/drude temp 300. 300. 100 1. 20
thermo_style custom f_TGNVT[1] f_TGNVT[2] f_TGNVT[3]
NPT simulation with TGNH thermostat
.. code-block:: LAMMPS
comm_modify vel yes
fix TGNPT all tgnpt/drude temp 300. 300. 100 1. 20 iso 1. 1. 500
thermo_style custom f_TGNPT[1] f_TGNPT[2] f_TGNPT[3]
----------
**Rigid bodies**
@ -480,3 +506,7 @@ NPT ensemble using Nose-Hoover thermostat:
**(SWM4-NDP)** Lamoureux, Harder, Vorobyov, Roux, MacKerell, Chem Phys
Let, 418, 245-249 (2006)
.. _TGNH-Son:
**(Son)** Son, McDaniel, Cui and Yethiraj, J Phys Chem Lett, 10, 7523 (2019).

6
doc/src/comm_modify.rst
View File

@ -85,15 +85,15 @@ information is available, then also a heuristic based on that bond length
is computed. It is used as communication cutoff, if there is no pair
style present and no *comm_modify cutoff* command used. Otherwise a
warning is printed, if this bond based estimate is larger than the
communication cutoff used.
communication cutoff used.
The *cutoff/multi* option is equivalent to *cutoff*\ , but applies to
communication mode *multi* instead. Since in this case the communication
cutoffs are determined per atom type, a type specifier is needed and
cutoff for one or multiple types can be extended. Also ranges of types
using the usual asterisk notation can be given. For granular pairstyles,
using the usual asterisk notation can be given. For granular pair styles,
the default cutoff is set to the sum of the current maximum atomic radii
for each type.
for each type.
The *cutoff/bytype* option applies to *multi* and automtically sets communication
cutoffs for each particle type based on the largest interaction distance

2
doc/src/fix.rst
View File

@ -364,6 +364,8 @@ accelerated styles exist.
* :doc:`temp/rescale <fix_temp_rescale>` - temperature control by velocity rescaling
* :doc:`temp/rescale/eff <fix_temp_rescale_eff>` - temperature control by velocity rescaling in the electron force field model
* :doc:`tfmc <fix_tfmc>` - perform force-bias Monte Carlo with time-stamped method
* :doc:`tgnvt/drude <fix_tgnh_drude>` - NVT time integration for Drude polarizable model via temperature-grouped Nose-Hoover
* :doc:`tgnpt/drude <fix_tgnh_drude>` - NPT time integration for Drude polarizable model via temperature-grouped Nose-Hoover
* :doc:`thermal/conductivity <fix_thermal_conductivity>` - Muller-Plathe kinetic energy exchange for thermal conductivity calculation
* :doc:`ti/spring <fix_ti_spring>` -
* :doc:`tmd <fix_tmd>` - guide a group of atoms to a new configuration

21
doc/src/fix_bond_react.rst
View File

@ -457,6 +457,23 @@ example, the molecule fragment could consist of only the backbone
atoms of a polymer chain. This constraint can be used to enforce a
specific relative position and orientation between reacting molecules.
By default, all constraints must be satisfied for the reaction to
occur. In other words, constraints are evaluated as a series of
logical values using the logical AND operator "&&". More complex logic
can be achieved by explicitly adding the logical AND operator "&&" or
the logical OR operator "||" after a given constraint command. If a
logical operator is specified after a constraint, it must be placed
after all constraint parameters, on the same line as the constraint
(one per line). Similarly, parentheses can be used to group
constraints. The expression that results from concatenating all
constraints should be a valid logical expression that can be read by
the :doc:`variable <variable>` command after converting each
constraint to a logical value. Because exactly one constraint is
allowed per line, having a valid logical expression implies that left
parentheses "(" should only appear before a constraint, and right
parentheses ")" should only appear after a constraint and before any
logical operator.
Once a reaction site has been successfully identified, data structures
within LAMMPS that store bond topology are updated to reflect the
post-reacted molecule template. All force fields with fixed bonds,
@ -599,8 +616,8 @@ reset_mol_ids = yes, custom_charges = no, molecule = off
.. _Gissinger:
**(Gissinger)** Gissinger, Jensen and Wise, Polymer, 128, 211 (2017).
**(Gissinger)** Gissinger, Jensen and Wise, Polymer, 128, 211-217 (2017).
.. _Gissinger2020:
**(Gissinger)** Gissinger, Jensen and Wise, Macromolecules (2020, in press).
**(Gissinger)** Gissinger, Jensen and Wise, Macromolecules, 53, 22, 9953–9961 (2020).

4
doc/src/fix_drude.rst
View File

@ -41,12 +41,12 @@ Restrictions
""""""""""""
This fix should be invoked before any other commands that implement
the Drude oscillator model, such as :doc:`fix langevin/drude <fix_langevin_drude>`, :doc:`fix drude/transform <fix_drude_transform>`, :doc:`compute temp/drude <compute_temp_drude>`, :doc:`pair_style thole <pair_thole>`.
the Drude oscillator model, such as :doc:`fix langevin/drude <fix_langevin_drude>`, :doc:`fix tgnvt/drude <fix_tgnh_drude>`, :doc:`fix drude/transform <fix_drude_transform>`, :doc:`compute temp/drude <compute_temp_drude>`, :doc:`pair_style thole <pair_thole>`.
Related commands
""""""""""""""""
:doc:`fix langevin/drude <fix_langevin_drude>`, :doc:`fix drude/transform <fix_drude_transform>`, :doc:`compute temp/drude <compute_temp_drude>`, :doc:`pair_style thole <pair_thole>`
:doc:`fix langevin/drude <fix_langevin_drude>`, :doc:`fix tgnvt/drude <fix_tgnh_drude>`, :doc:`fix drude/transform <fix_drude_transform>`, :doc:`compute temp/drude <compute_temp_drude>`, :doc:`pair_style thole <pair_thole>`
Default
"""""""

18
doc/src/fix_temp_csvr.rst
View File

@ -124,16 +124,16 @@ temperature is calculated taking the bias into account, bias is
removed from each atom, thermostatting is performed on the remaining
thermal degrees of freedom, and the bias is added back in.
An important feature of these thermostats is that they have an
An important feature of these thermostats is that they have an
associated effective energy that is a constant of motion.
The effective energy is the total energy (kinetic + potential) plus
the accumulated kinetic energy changes due to the thermostat. The
latter quantity is the global scalar computed by these fixes. This
feature is useful to check the integration of the equations of motion
against discretization errors. In other words, the conservation of
the effective energy can be used to choose an appropriate integration
:doc:`timestep <timestep>`. This is similar to the usual paradigm of
checking the conservation of the total energy in the microcanonical
The effective energy is the total energy (kinetic + potential) plus
the accumulated kinetic energy changes due to the thermostat. The
latter quantity is the global scalar computed by these fixes. This
feature is useful to check the integration of the equations of motion
against discretization errors. In other words, the conservation of
the effective energy can be used to choose an appropriate integration
:doc:`timestep <timestep>`. This is similar to the usual paradigm of
checking the conservation of the total energy in the microcanonical
ensemble.

305
doc/src/fix_tgnh_drude.rst Normal file
View File

@ -0,0 +1,305 @@
.. index:: fix tgnvt/drude
.. index:: fix tgnpt/drude
fix tgnvt/drude command
=======================
fix tgnpt/drude command
=======================
Syntax
""""""
.. parsed-literal::
fix ID group-ID style_name keyword values ...
* ID, group-ID are documented in :doc:`fix <fix>` command
* style_name = *tgnvt/drude* or *tgnpt/drude*
* one or more keyword/values pairs may be appended
.. parsed-literal::
keyword = *temp* *iso* or *aniso* or *tri* or *x* or *y* or *z* or *xy* or *yz* or *xz* or *couple* or *tchain* or *pchain* or *mtk* or *tloop* or *ploop* or *nreset* or *scalexy* or *scaleyz* or *scalexz* or *flip* or *fixedpoint*
*temp* values = Tstart Tstop Tdamp Tdrude Tdamp_drude
Tstart, Tstop = external temperature at start/end of run (temperature units)
Tdamp = temperature damping parameter (time units)
Tdrude = desired temperature of Drude oscillators (temperature units)
Tdamp_drude = temperature damping parameter for Drude oscillators (time units)
*iso* or *aniso* or *tri* values = Pstart Pstop Pdamp
Pstart,Pstop = scalar external pressure at start/end of run (pressure units)
Pdamp = pressure damping parameter (time units)
*x* or *y* or *z* or *xy* or *yz* or *xz* values = Pstart Pstop Pdamp
Pstart,Pstop = external stress tensor component at start/end of run (pressure units)
Pdamp = stress damping parameter (time units)
*couple* = *none* or *xyz* or *xy* or *yz* or *xz*
*tchain* value = N
N = length of thermostat chain (1 = single thermostat)
*pchain* value = N
N length of thermostat chain on barostat (0 = no thermostat)
*mtk* value = *yes* or *no* = add in MTK adjustment term or not
*tloop* value = M
M = number of sub-cycles to perform on thermostat
*ploop* value = M
M = number of sub-cycles to perform on barostat thermostat
*nreset* value = reset reference cell every this many timesteps
*scalexy* value = *yes* or *no* = scale xy with ly
*scaleyz* value = *yes* or *no* = scale yz with lz
*scalexz* value = *yes* or *no* = scale xz with lz
*flip* value = *yes* or *no* = allow or disallow box flips when it becomes highly skewed
*fixedpoint* values = x y z
x,y,z = perform barostat dilation/contraction around this point (distance units)
Examples
""""""""
.. code-block:: LAMMPS
comm_modify vel yes
fix 1 all tgnvt/drude temp 300.0 300.0 100.0 1.0 20.0
fix 1 water tgnpt/drude temp 300.0 300.0 100.0 1.0 20.0 iso 0.0 0.0 1000.0
fix 2 jello tgnpt/drude temp 300.0 300.0 100.0 1.0 20.0 tri 5.0 5.0 1000.0
fix 2 ice tgnpt/drude temp 250.0 250.0 100.0 1.0 20.0 x 1.0 1.0 0.5 y 2.0 2.0 0.5 z 3.0 3.0 0.5 yz 0.1 0.1 0.5 xz 0.2 0.2 0.5 xy 0.3 0.3 0.5 nreset 1000
Description
"""""""""""
These commands are variants of the Nose-Hoover fix styles :doc:`fix nvt
<fix_nh>` and :doc:`fix npt <fix_nh>` for thermalized Drude polarizable
models. They apply temperature-grouped Nose-Hoover thermostat (TGNH)
proposed by :ref:`(Son) <tgnh-Son>`. When there are fast vibrational
modes with frequencies close to Drude oscillators (e.g. double bonds or
out-of-plane torsions), this thermostat can provide better kinetic
energy equipartitioning.
The difference between TGNH and the original Nose-Hoover thermostat is that,
TGNH separates the kinetic energy of the group into three contributions:
molecular center of mass (COM) motion,
motion of COM of atom-Drude pairs or non-polarizable atoms relative to molecular COM,
and relative motion of atom-Drude pairs.
An independent Nose-Hoover chain is applied to each type of motion.
The temperatures for these three types of motion are denoted as
molecular translational temperature (:math:`T_\mathrm{M}`), real atomic temperature (:math:`T_\mathrm{R}`) and Drude temperature (:math:`T_\mathrm{D}`),
which are defined in terms of their associated degrees of freedom (DOF):
.. math::
T_\mathrm{M}=\frac{\Sigma_{i}^{N_\mathrm{mol}} M_i V_i^2}{3 \left ( N_\mathrm{mol} - \frac{N_\mathrm{mol}}{N_\mathrm{mol,sys}} \right ) k_\mathrm{B}}
.. math::
T_\mathrm{R}=\frac{\Sigma_{i}^{N_\mathrm{real}} m_i (v_i-v_{M,i})^2}{(N_\mathrm{DOF} - 3 N_\mathrm{mol} + 3 \frac{N_\mathrm{mol}}{N_\mathrm{mol,sys}} - 3 N_\mathrm{drude}) k_\mathrm{B}}
.. math::
T_\mathrm{D}=\frac{\Sigma_{i}^{N_\mathrm{drude}} m_i^{\prime} v_i^{\prime 2}}{3 N_\mathrm{drude} k_\mathrm{B}}
Here :math:`N_\mathrm{mol}` and :math:`N_\mathrm{mol,sys}` are the numbers of molecules in the group and in the whole system, respectively.
:math:`N_\mathrm{real}` is the number of atom-Drude pairs and non-polarizable atoms in the group.
:math:`N_\mathrm{drude}` is the number of Drude particles in the group.
:math:`N_\mathrm{DOF}` is the DOF of the group.
:math:`M_i` and :math:`V_i` are the mass and the COM velocity of the i-th molecule.
:math:`m_i` is the mass of the i-th atom-Drude pair or non-polarizable atom.
:math:`v_i` is the velocity of COM of i-th atom-Drude pair or non-polarizable atom.
:math:`v_{M,i}` is the COM velocity of the molecule the i-th atom-Drude pair or non-polarizable atom belongs to.
:math:`m_i^\prime` and :math:`v_i^\prime` are the reduced mass and the relative velocity of the i-th atom-Drude pair.
.. note::
These fixes require that each atom knows whether it is a Drude particle or
not. You must therefore use the :doc:`fix drude <fix_drude>` command to
specify the Drude status of each atom type.
Because the TGNH thermostat thermostats the molecular COM motion,
all atoms belonging to the same molecule must be in the same group.
That is, these fixes can not be applied to a subset of a molecule.
For this fix to act correctly, ghost atoms need to know their velocity.
You must use the :doc:`comm_modify <comm_modify>` command to enable this.
These fixes assume that the translational DOF of the whole system is removed.
It is therefore recommended to invoke :doc:`fix momentum <fix_momentum>` command so that the :math:`T_\mathrm{M}` is calculated correctly.
----------
The thermostat parameters are specified using the *temp* keyword.
The thermostat is applied to only the translational DOF
for the particles. The translational DOF can also have
a bias velocity removed before thermostatting takes place; see the
description below. The desired temperature for molecular and real atomic motion is a
ramped value during the run from *Tstart* to *Tstop*\ . The *Tdamp*
parameter is specified in time units and determines how rapidly the
temperature is relaxed. For example, a value of 10.0 means to relax
the temperature in a timespan of (roughly) 10 time units (e.g. :math:`\tau`
or fs or ps - see the :doc:`units <units>` command).
The parameter *Tdrude* is the desired temperature for Drude motion at each timestep.
Similar to *Tdamp*, the *Tdamp_drude* parameter determines the relaxation speed for Drude motion.
Fix group are the only ones whose velocities and positions are updated
by the velocity/position update portion of the integration.
Other thermostat-related keywords are *tchain*\ and *tloop*\ ,
which are detailed in :doc:`fix nvt <fix_nh>`.
.. note::
A Nose-Hoover thermostat will not work well for arbitrary values
of *Tdamp*\ . If *Tdamp* is too small, the temperature can fluctuate
wildly; if it is too large, the temperature will take a very long time
to equilibrate. A good choice for many models is a *Tdamp* of around
100 timesteps. A smaller *Tdamp_drude* value would be required
to maintain Drude motion at low temperature.
.. code-block:: LAMMPS
fix 1 all nvt temp 300.0 300.0 $(100.0*dt) 1.0 $(20.0*dt)
----------
The barostat parameters for fix style *tgnpt/drude* is specified
using one or more of the *iso*\ , *aniso*\ , *tri*\ , *x*\ , *y*\ , *z*\ , *xy*\ ,
*xz*\ , *yz*\ , and *couple* keywords. These keywords give you the
ability to specify all 6 components of an external stress tensor, and
to couple various of these components together so that the dimensions
they represent are varied together during a constant-pressure
simulation. Other barostat-related keywords are *pchain*\ , *mtk*\ , *ploop*\ ,
*nreset*\ , *scalexy*\ , *scaleyz*\ , *scalexz*\ , *flip*\ and *fixedpoint*.
The meaning of barostat parameters are detailed in :doc:`fix npt <fix_nh>`.
Regardless of what atoms are in the fix group (the only atoms which
are time integrated), a global pressure or stress tensor is computed
for all atoms. Similarly, when the size of the simulation box is
changed, all atoms are re-scaled to new positions.
.. note::
Unlike the :doc:`fix temp/berendsen <fix_temp_berendsen>` command
which performs thermostatting but NO time integration, these fixes
perform thermostatting/barostatting AND time integration. Thus you
should not use any other time integration fix, such as :doc:`fix nve <fix_nve>` on atoms to which this fix is applied.
Likewise, these fixes should not be used on atoms that also
have their temperature controlled by another fix - e.g. by :doc:`fix langevin/drude <fix_langevin_drude>` command.
See the :doc:`Howto thermostat <Howto_thermostat>` and :doc:`Howto barostat <Howto_barostat>` doc pages for a discussion of different
ways to compute temperature and perform thermostatting and
barostatting.
----------
Like other fixes that perform thermostatting, these fixes can
be used with :doc:`compute commands <compute>` that calculate a
temperature after removing a "bias" from the atom velocities.
This is not done by default, but only if the :doc:`fix_modify <fix_modify>` command
is used to assign a temperature compute to this fix that includes such
a bias term. See the doc pages for individual :doc:`compute commands <compute>` to determine which ones include a bias. In
this case, the thermostat works in the following manner: the current
temperature is calculated taking the bias into account, bias is
removed from each atom, thermostatting is performed on the remaining
thermal DOF, and the bias is added back in.
.. note::
However, not all temperature compute commands are valid to be used with these fixes.
Precisely, only temperature compute that does not modify the DOF of the group can be used.
E.g. :doc:`compute temp/ramp <compute_temp_ramp>` and :doc:`compute viscosity/cos <compute_viscosity_cos>`
compute the kinetic energy after remove a velocity gradient without affecting the DOF of the group,
then they can be invoked in this way.
In contrast, :doc:`compute temp/partial <compute_temp_partial>` may remove the DOF at one or more dimensions,
therefore it cannot be used with these fixes.
----------
Restart, fix_modify, output, run start/stop, minimize info
"""""""""""""""""""""""""""""""""""""""""""""""""""""""""""
These fixes writes the state of all the thermostat and barostat
variables to :doc:`binary restart files <restart>`. See the
:doc:`read_restart <read_restart>` command for info on how to re-specify
a fix in an input script that reads a restart file, so that the
operation of the fix continues in an uninterrupted fashion.
The :doc:`fix_modify <fix_modify>` *temp* and *press* options are
supported by these fixes. You can use them to assign a
:doc:`compute <compute>` you have defined to this fix which will be used
in its thermostatting or barostatting procedure, as described above.
If you do this, note that the kinetic energy derived from the compute
temperature should be consistent with the virial term computed using
all atoms for the pressure. LAMMPS will warn you if you choose to
compute temperature on a subset of atoms.
.. note::
If both the *temp* and *press* keywords are used in a single
thermo_modify command (or in two separate commands), then the order in
which the keywords are specified is important. Note that a :doc:`pressure compute <compute_pressure>` defines its own temperature compute as
an argument when it is specified. The *temp* keyword will override
this (for the pressure compute being used by fix npt), but only if the
*temp* keyword comes after the *press* keyword. If the *temp* keyword
comes before the *press* keyword, then the new pressure compute
specified by the *press* keyword will be unaffected by the *temp*
setting.
The :doc:`fix_modify <fix_modify>` *energy* option is supported by these
fixes to add the energy change induced by Nose/Hoover thermostatting
and barostatting to the system's potential energy as part of
:doc:`thermodynamic output <thermo_style>`.
These fixes compute a global scalar and a global vector of quantities,
which can be accessed by various :doc:`output commands <Howto_output>`.
The scalar value calculated by these fixes is "extensive"; the vector
values are "intensive".
The scalar is the cumulative energy change due to the fix.
The vector stores the three temperatures :math:`T_\mathrm{M}`, :math:`T_\mathrm{R}` and :math:`T_\mathrm{D}`.
These fixes can ramp their external temperature and pressure over
multiple runs, using the *start* and *stop* keywords of the
:doc:`run <run>` command. See the :doc:`run <run>` command for details of
how to do this.
These fixes are not invoked during :doc:`energy minimization <minimize>`.
----------
Restrictions
""""""""""""
These fixes are only available when LAMMPS was built with the USER-DRUDE package.
These fixes cannot be used with dynamic groups as defined by the :doc:`group <group>` command.
These fixes cannot be used in 2D simulations.
*X*\ , *y*\ , *z* cannot be barostatted if the associated dimension is not
periodic. *Xy*\ , *xz*\ , and *yz* can only be barostatted if the
simulation domain is triclinic and the second dimension in the keyword
(\ *y* dimension in *xy*\ ) is periodic. The :doc:`create_box <create_box>`,
:doc:`read data <read_data>`, and :doc:`read_restart <read_restart>`
commands specify whether the simulation box is orthogonal or
non-orthogonal (triclinic) and explain the meaning of the xy,xz,yz
tilt factors.
For the *temp* keyword, the final *Tstop* cannot be 0.0 since it would
make the external T = 0.0 at some timestep during the simulation which
is not allowed in the Nose/Hoover formulation.
The *scaleyz yes*\ , *scalexz yes*\ , and *scalexy yes* options
can only be used if the second dimension in the keyword is periodic,
and if the tilt factor is not coupled to the barostat via keywords
*tri*\ , *yz*\ , *xz*\ , and *xy*\ .
Related commands
""""""""""""""""
:doc:`fix drude <fix_drude>`, :doc:`fix nvt <fix_nh>`, :doc:`fix_npt <fix_nh>`,
:doc:`fix_modify <fix_modify>`
Default
"""""""
The keyword defaults are tchain = 3, pchain = 3, mtk = yes, tloop = 1,
ploop = 1, nreset = 0, couple = none,
flip = yes, scaleyz = scalexz = scalexy = yes if periodic in second
dimension and not coupled to barostat, otherwise no.
----------
.. _tgnh-Son:
**(Son)** Son, McDaniel, Cui and Yethiraj, J Phys Chem Lett, 10, 7523 (2019).

8
doc/src/neighbor.rst
View File

@ -49,7 +49,7 @@ sometimes be faster. Either style should give the same answers.
The *multi* style is a modified binning algorithm that is useful for
systems with a wide range of cutoff distances, e.g. due to different
size particles. For granular pairstyles, cutoffs are set to the
size particles. For granular pair styles, cutoffs are set to the
sum of the maximum atomic radii for each atom type.
For the *bin* style, the bin size is set to 1/2 of
the largest cutoff distance between any pair of atom types and a
@ -59,8 +59,10 @@ other type pairs have a much shorter cutoff. For style *multi* the
bin size is set to 1/2 of the shortest cutoff distance and multiple
sets of bins are defined to search over for different atom types.
This imposes some extra setup overhead, but the searches themselves
may be much faster for the short-cutoff cases. See the :doc:`comm_modify mode multi <comm_modify>` command for a communication option
that may also be beneficial for simulations of this kind.
may be much faster for the short-cutoff cases.
See the :doc:`comm_modify mode multi <comm_modify>` command for a
communication option that may also be beneficial for simulations of
this kind.
The *bytype* style is an extension of the *multi* style that was
presented by Shire, Hanley, and Stratford :ref:`(Shire) <bytype-Shire>`.

28
doc/src/package.rst
View File

@ -18,13 +18,16 @@ Syntax
*gpu* args = Ngpu keyword value ...
Ngpu = # of GPUs per node
zero or more keyword/value pairs may be appended
keywords = *neigh* or *newton* or *binsize* or *split* or *gpuID* or *tpa* or *device* or *blocksize*
keywords = *neigh* or *newton* or *pair/only* or *binsize* or *split* or *gpuID* or *tpa* or *device* or *blocksize*
*neigh* value = *yes* or *no*
yes = neighbor list build on GPU (default)
no = neighbor list build on CPU
*newton* = *off* or *on*
off = set Newton pairwise flag off (default and required)
on = set Newton pairwise flag on (currently not allowed)
*pair/only* = *off* or *on*
off = apply "gpu" suffix to all available styles in the GPU package (default)
on - apply "gpu" suffix only pair styles
*binsize* value = size
size = bin size for neighbor list construction (distance units)
*split* = fraction
@ -65,7 +68,7 @@ Syntax
*no_affinity* values = none
*kokkos* args = keyword value ...
zero or more keyword/value pairs may be appended
keywords = *neigh* or *neigh/qeq* or *neigh/thread* or *newton* or *binsize* or *comm* or *comm/exchange* or *comm/forward* or *comm/reverse* or *cuda/aware*
keywords = *neigh* or *neigh/qeq* or *neigh/thread* or *newton* or *binsize* or *comm* or *comm/exchange* or *comm/forward* or *comm/reverse* or *cuda/aware* or *pair/only*
*neigh* value = *full* or *half*
full = full neighbor list
half = half neighbor list built in thread-safe manner
@ -91,6 +94,9 @@ Syntax
*cuda/aware* = *off* or *on*
off = do not use CUDA-aware MPI
on = use CUDA-aware MPI (default)
*pair/only* = *off* or *on*
off = use device acceleration (e.g. GPU) for all available styles in the KOKKOS package (default)
on = use device acceleration only for pair styles (and host acceleration for others)
*omp* args = Nthreads keyword value ...
Nthread = # of OpenMP threads to associate with each MPI process
zero or more keyword/value pairs may be appended
@ -194,6 +200,14 @@ for compatibility with the package command for other accelerator
styles. Note that the newton setting for bonded interactions is not
affected by this keyword.
The *pair/only* keyword can change how any "gpu" suffix is applied.
By default a suffix is applied to all styles for which an accelerated
variant is available. However, that is not always the most effective
way to use an accelerator. With *pair/only* set to *on* the suffix
will only by applied to supported pair styles, which tend to be the
most effective in using an accelerator and their operation can be
overlapped with all other computations on the CPU.
The *binsize* keyword sets the size of bins used to bin atoms in
neighbor list builds performed on the GPU, if *neigh* = *yes* is set.
If *binsize* is set to 0.0 (the default), then bins = the size of the
@ -534,12 +548,20 @@ available (currently only possible with OpenMPI v2.0.0 or later), then
the *cuda/aware* keyword is automatically set to *off* by default. When
the *cuda/aware* keyword is set to *off* while any of the *comm*
keywords are set to *device*\ , the value for these *comm* keywords will
be automatically changed to *host*\ . This setting has no effect if not
be automatically changed to *no*\ . This setting has no effect if not
running on GPUs or if using only one MPI rank. CUDA-aware MPI is available
for OpenMPI 1.8 (or later versions), Mvapich2 1.9 (or later) when the
"MV2_USE_CUDA" environment variable is set to "1", CrayMPI, and IBM
Spectrum MPI when the "-gpu" flag is used.
The *pair/only* keyword can change how the KOKKOS suffix "kk" is applied
when using an accelerator device. By default device acceleration is
always used for all available styles. With *pair/only* set to *on* the
suffix setting will choose device acceleration only for pair styles and
run all other force computations concurrently on the host CPU.
The *comm* flags will also automatically be changed to *no*\ . This can
result in better performance for certain configurations and system sizes.
----------
The *omp* style invokes settings associated with the use of the