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

Merge branch 'master' into reset-molecules

Browse Source
This commit is contained in:
Axel Kohlmeyer
2020-07-06 14:07:48 -04:00
parent a1011b606e 6bf329098e
commit 58cd6a160c
189 changed files with 13723 additions and 8435 deletions
Download Patch File Download Diff File Expand all files Collapse all files

7
doc/src/Commands_all.rst
View File

@ -46,13 +46,6 @@ An alphabetic list of all general LAMMPS commands.
* :doc:`dimension <dimension>`
* :doc:`displace_atoms <displace_atoms>`
* :doc:`dump <dump>`
* :doc:`dump atom/adios <dump_adios>`
* :doc:`dump custom/adios <dump_adios>`
* :doc:`dump image <dump_image>`
* :doc:`dump movie <dump_image>`
* :doc:`dump netcdf <dump_netcdf>`
* :doc:`dump netcdf/mpiio <dump_netcdf>`
* :doc:`dump vtk <dump_vtk>`
* :doc:`dump_modify <dump_modify>`
* :doc:`dynamical_matrix <dynamical_matrix>`
* :doc:`echo <echo>`

2
doc/src/Commands_compute.rst
View File

@ -80,6 +80,7 @@ KOKKOS, o = USER-OMP, t = OPT.
* :doc:`ke/eff <compute_ke_eff>`
* :doc:`ke/rigid <compute_ke_rigid>`
* :doc:`mesont <compute_mesont>`
* :doc:`mliap <compute_mliap>`
* :doc:`momentum <compute_momentum>`
* :doc:`msd <compute_msd>`
* :doc:`msd/chunk <compute_msd_chunk>`
@ -122,6 +123,7 @@ KOKKOS, o = USER-OMP, t = OPT.
* :doc:`smd/tlsph/strain/rate <compute_smd_tlsph_strain_rate>`
* :doc:`smd/tlsph/stress <compute_smd_tlsph_stress>`
* :doc:`smd/triangle/vertices <compute_smd_triangle_vertices>`
* :doc:`smd/ulsph/effm <compute_smd_ulsph_effm>`
* :doc:`smd/ulsph/num/neighs <compute_smd_ulsph_num_neighs>`
* :doc:`smd/ulsph/strain <compute_smd_ulsph_strain>`
* :doc:`smd/ulsph/strain/rate <compute_smd_ulsph_strain_rate>`

4
doc/src/compute.rst
View File

@ -225,6 +225,7 @@ The individual style names on the :doc:`Commands compute <Commands_compute>` doc
* :doc:`ke/atom/eff <compute_ke_atom_eff>` - per-atom translational and radial kinetic energy in the electron force field model
* :doc:`ke/eff <compute_ke_eff>` - kinetic energy of a group of nuclei and electrons in the electron force field model
* :doc:`ke/rigid <compute_ke_rigid>` - translational kinetic energy of rigid bodies
* :doc:`mliap <compute_mliap>` - gradients of energy and forces w.r.t. model parameters and related quantities for training machine learning interatomic potentials
* :doc:`momentum <compute_momentum>` - translational momentum
* :doc:`msd <compute_msd>` - mean-squared displacement of group of atoms
* :doc:`msd/chunk <compute_msd_chunk>` - mean-squared displacement for each chunk
@ -268,12 +269,13 @@ The individual style names on the :doc:`Commands compute <Commands_compute>` doc
* :doc:`smd/tlsph/strain/rate <compute_smd_tlsph_strain_rate>` -
* :doc:`smd/tlsph/stress <compute_smd_tlsph_stress>` - per-particle Cauchy stress tensor for SPH particles
* :doc:`smd/triangle/vertices <compute_smd_triangle_vertices>` -
* :doc:`smd/ulsph/effm <compute_smd_ulsph_effm>` - per-particle effective shear modulus
* :doc:`smd/ulsph/num/neighs <compute_smd_ulsph_num_neighs>` -
* :doc:`smd/ulsph/strain <compute_smd_ulsph_strain>` -
* :doc:`smd/ulsph/strain/rate <compute_smd_ulsph_strain_rate>` -
* :doc:`smd/ulsph/stress <compute_smd_ulsph_stress>` - per-particle Cauchy stress tensor and von Mises equivalent stress in Smooth Mach Dynamics
* :doc:`smd/vol <compute_smd_vol>` - per-particle volumes and their sum in Smooth Mach Dynamics
* :doc:`snap <compute_sna_atom>` - bispectrum components and related quantities for a group of atoms
* :doc:`snap <compute_sna_atom>` - gradients of SNAP energy and forces w.r.t. linear coefficients and related quantities for fitting SNAP potentials
* :doc:`sna/atom <compute_sna_atom>` - bispectrum components for each atom
* :doc:`snad/atom <compute_sna_atom>` - derivative of bispectrum components for each atom
* :doc:`snav/atom <compute_sna_atom>` - virial contribution from bispectrum components for each atom

170
doc/src/compute_mliap.rst Normal file
View File

@ -0,0 +1,170 @@
.. index:: compute mliap
compute mliap command
=====================
Syntax
""""""
.. code-block:: LAMMPS
compute ID group-ID mliap ... keyword values ...
* ID, group-ID are documented in :doc:`compute <compute>` command
* mliap = style name of this compute command
* two keyword/value pairs must be appended
* keyword = *model* or *descriptor*
.. parsed-literal::
*model* values = style Nelems Nparams
style = *linear* or *quadratic*
Nelems = number of elements
Nparams = number of parameters per element
*descriptor* values = style filename
style = *sna*
filename = name of file containing descriptor definitions
Examples
""""""""
.. code-block:: LAMMPS
compute mliap model linear 2 31 descriptor sna Ta06A.mliap.descriptor
Description
"""""""""""
Compute style *mliap* provides a general interface to the gradient
of machine-learning interatomic potentials w.r.t. model parameters.
It is used primarily for calculating the gradient of energy, force, and
stress components w.r.t. model parameters, which is useful when training
:doc:`mliap pair_style <pair_mliap>` models to match target data.
It provides separate
definitions of the interatomic potential functional form (*model*)
and the geometric quantities that characterize the atomic positions
(*descriptor*). By defining *model* and *descriptor* separately,
it is possible to use many different models with a given descriptor,
or many different descriptors with a given model. Currently, the
compute supports just two models, *linear* and *quadratic*,
and one descriptor, *sna*, the SNAP descriptor used by
:doc:`pair_style snap <pair_snap>`, including the linear, quadratic,
and chem variants. Work is currently underway to extend
the interface to handle neural network energy models,
and it is also straightforward to add new descriptor styles.
The compute *mliap* command must be followed by two keywords
*model* and *descriptor* in either order.
The *model* keyword is followed by a model style, currently limited to
either *linear* or *quadratic*. In both cases,
this is followed by two arguments. *nelems* is the number of elements.
It must be equal to the number of LAMMPS atom types. *nparams*
is the number of parameters per element for this model i.e.
the number of parameter gradients for each element. Note these definitions
are identical to those of *nelems* and *nparams* in the
:doc:`pair_style mliap <pair_mliap>` model file.
The *descriptor* keyword is followed by a descriptor style, and additional arguments.
Currently the only descriptor style is *sna*, indicating the bispectrum component
descriptors used by the Spectral Neighbor Analysis Potential (SNAP) potentials of
:doc:`pair_style snap <pair_snap>`.
The \'p\' in SNAP is dropped, because keywords that match pair_styles are silently stripped
out by the LAMMPS command parser. A single additional argument specifies the descriptor filename
containing the parameters and setting used by the SNAP descriptor.
The descriptor filename usually ends in the *.mliap.descriptor* extension.
The format of this file is identical to the descriptor file in the
:doc:`pair_style mliap <pair_mliap>`, and is described in detail
there.
.. note::
The number of LAMMPS atom types (and the value of *nelems* in the model)
must match the value of *nelems* in the descriptor file.
Compute *mliap* calculates a global array containing gradient information.
The number of columns in the array is :math:`nelems \times nparams + 1`.
The first row of the array contain the derivative of potential energy w.r.t. to
each parameter and each element. The last six rows
of the array contain the corresponding derivatives of the
virial stress tensor, listed in Voigt notation: *pxx*, *pyy*, *pzz*,
*pyz*, *pxz*, *pxy*. In between the energy and stress rows are
the 3\*\ *N* rows containing the derivatives of the force components.
See section below on output for a detailed description of how
rows and columns are ordered.
The element in the last column of each row contains
the potential energy, force, or stress, according to the row.
These quantities correspond to the user-specified reference potential
that must be subtracted from the target data when training a model.
The potential energy calculation uses the built in compute *thermo_pe*.
The stress calculation uses a compute called *mliap_press* that is
automatically created behind the scenes, according to the following
command:
.. code-block:: LAMMPS
compute mliap_press all pressure NULL virial
See section below on output for a detailed explanation of the data
layout in the global array.
Atoms not in the group do not contribute to this compute.
Neighbor atoms not in the group do not contribute to this compute.
The neighbor list needed to compute this quantity is constructed each
time the calculation is performed (i.e. each time a snapshot of atoms
is dumped). Thus it can be inefficient to compute/dump this quantity
too frequently.
.. note::
If the user-specified reference potentials includes bonded and
non-bonded pairwise interactions, then the settings of
:doc:`special_bonds <special_bonds>` command can remove pairwise
interactions between atoms in the same bond, angle, or dihedral. This
is the default setting for the :doc:`special_bonds <special_bonds>`
command, and means those pairwise interactions do not appear in the
neighbor list. Because this fix uses the neighbor list, it also means
those pairs will not be included in the calculation. The :doc:`rerun <rerun>`
command is not an option here, since the reference potential is required
for the last column of the global array. A work-around is to prevent
pairwise interactions from being removed by explicitly adding a
*tiny* positive value for every pairwise interaction that would otherwise be
set to zero in the :doc:`special_bonds <special_bonds>` command.
----------
**Output info:**
Compute *mliap* evaluates a global array.
The columns are arranged into
*nelems* blocks, listed in order of element *I*\ . Each block
contains one column for each of the *nparams* model parameters.
A final column contains the corresponding energy, force component
on an atom, or virial stress component. The rows of the array appear
in the following order:
* 1 row: Derivatives of potential energy w.r.t. each parameter of each element.
* 3\*\ *N* rows: Derivatives of force components. x, y, and z components of force on atom *i* appearing in consecutive rows. The atoms are sorted based on atom ID.
* 6 rows: Derivatives of virial stress tensor w.r.t. each parameter of each element. The ordering of the rows follows Voigt notation: *pxx*, *pyy*, *pzz*, *pyz*, *pxz*, *pxy*.
These values can be accessed by any command that uses a global array
from a compute as input. See the :doc:`Howto output <Howto_output>` doc
page for an overview of LAMMPS output options. To see how this command
can be used within a Python workflow to train machine-learning interatomic
potentials, see the examples in `FitSNAP <https://github.com/FitSNAP/FitSNAP>`_.
Restrictions
""""""""""""
This compute is part of the MLIAP package. It is only enabled if
LAMMPS was built with that package. In addition, building LAMMPS with the MLIAP package
requires building LAMMPS with the SNAP package.
See the :doc:`Build package <Build_package>` doc page for more info.
Related commands
""""""""""""""""
:doc:`pair_style mliap <pair_mliap>`
**Default:** none

56
doc/src/compute_smd_ulsph_effm.rst Normal file
View File

@ -0,0 +1,56 @@
.. index:: compute smd/ulsph/effm
compute smd/ulsph/effm command
================================
Syntax
""""""
.. code-block:: LAMMPS
compute ID group-ID smd/ulsph/effm
* ID, group-ID are documented in :doc:`compute <compute>` command
* smd/ulsph/effm = style name of this compute command
Examples
""""""""
.. code-block:: LAMMPS
compute 1 all smd/ulsph/effm
Description
"""""""""""
Define a computation that outputs the effective shear modulus for
particles interacting via the updated Lagrangian SPH pair style.
See `this PDF guide <PDF/SMD_LAMMPS_userguide.pdf>`_ to using Smooth
Mach Dynamics in LAMMPS.
**Output info:**
This compute calculates a per-particle vector, which can be accessed
by any command that uses per-particle values from a compute as input.
See the :doc:`Howto output <Howto_output>` doc page for an overview of
LAMMPS output options.
The per-particle vector contains the current effective per atom shear
modulus as computed by the :doc:`pair smd/ulsph <pair_smd_ulsph>` pair
style.
Restrictions
""""""""""""
This compute is part of the USER-SMD package. It is only enabled if
LAMMPS was built with that package. See the :doc:`Build package <Build_package>` doc page for more info. This compute can
only be used for particles which interact with the updated Lagrangian
SPH pair style.
Related commands
""""""""""""""""
:doc:`pair smd/ulsph <pair_smd_ulsph>`
**Default:** none

4
doc/src/compute_sna_atom.rst
View File

@ -391,7 +391,9 @@ of :math:`K N_{elem}^3` columns.
These values can be accessed by any command that uses per-atom values
from a compute as input. See the :doc:`Howto output <Howto_output>` doc
page for an overview of LAMMPS output options.
page for an overview of LAMMPS output options. To see how this command
can be used within a Python workflow to train SNAP potentials,
see the examples in `FitSNAP <https://github.com/FitSNAP/FitSNAP>`_.
Restrictions
""""""""""""

2
doc/src/package.rst
View File

@ -518,7 +518,7 @@ script are supported by the KOKKOS package. In this case data can stay
on the GPU for many timesteps without being moved between the host and
GPU, if you use the *device* value. If your script uses styles (e.g.
fixes) which are not yet supported by the KOKKOS package, then data has
to be move between the host and device anyway, so it is typically faster
to be moved between the host and device anyway, so it is typically faster
to let the host handle communication, by using the *host* value. Using
*host* instead of *no* will enable use of multiple threads to
pack/unpack communicated data. When running small systems on a GPU,

33
doc/src/pair_mliap.rst
View File

@ -8,7 +8,19 @@ Syntax
.. code-block:: LAMMPS
pair_style mliap
pair_style mliap ... keyword values ...
* two keyword/value pairs must be appended
* keyword = *model* or *descriptor*
.. parsed-literal::
*model* values = style filename
style = *linear* or *quadratic*
filename = name of file containing model definitions
*descriptor* values = style filename
style = *sna*
filename = name of file containing descriptor definitions
Examples
""""""""
@ -22,8 +34,8 @@ Examples
Description
"""""""""""
Pair style *mliap* provides a general interface to families of
machine-learning interatomic potentials. It provides separate
Pair style *mliap* provides a general interface to families of
machine-learning interatomic potentials. It allows separate
definitions of the interatomic potential functional form (*model*)
and the geometric quantities that characterize the atomic positions
(*descriptor*). By defining *model* and *descriptor* separately,
@ -34,6 +46,9 @@ and one descriptor, *sna*, the SNAP descriptor used by :doc:`pair_style snap <pa
and chem variants. Work is currently underway to extend
the interface to handle neural network energy models,
and it is also straightforward to add new descriptor styles.
In order to train a model, it is useful to know the gradient or derivative
of energy, force, and stress w.r.t. model parameters. This information
can be accessed using the related :doc:`compute mliap <compute_mliap>` command.
The pair_style *mliap* command must be followed by two keywords
*model* and *descriptor* in either order. A single
@ -46,10 +61,10 @@ where N is the number of LAMMPS atom types.
The *model* keyword is followed by a model style, currently limited to
either *linear* or *quadratic*. In both cases,
this is followed by a single argument specifying the model filename containing the
linear or quadratic coefficients for a set of elements.
this is followed by a single argument specifying the model filename containing the
parameters for a set of elements.
The model filename usually ends in the *.mliap.model* extension.
It may contain coefficients for many elements. The only requirement is that it
It may contain parameters for many elements. The only requirement is that it
contain at least those element names appearing in the
*pair_coeff* command.
@ -58,10 +73,10 @@ but follows a strict format after that. The first non-blank non-comment
line must contain two integers:
* nelems = Number of elements
* ncoeff = Number of coefficients
* nparams = Number of parameters
This is followed by one block for each of the *nelem* elements.
Each block consists of *ncoeff* coefficients, one per line.
Each block consists of *nparams* parameters, one per line.
Note that this format is similar, but not identical to that used
for the :doc:`pair_style snap <pair_snap>` coefficient file.
Specifically, the line containing the element weight and radius is omitted,
@ -131,6 +146,6 @@ See the :doc:`Build package <Build_package>` doc page for more info.
Related commands
""""""""""""""""
:doc:`pair_style snap <pair_snap>`,
:doc:`pair_style snap <pair_snap>`, :doc:`compute mliap <compute_mliap>`
**Default:** none