tx258/lammps-gran-kokkos
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

small changes to temper/grem commands

Browse Source
This commit is contained in:
Steve Plimpton
2016-11-22 08:47:44 -07:00
parent d7b542101a
commit b95f255af4
6 changed files with 81 additions and 78 deletions
Download Patch File Download Diff File Expand all files Collapse all files

5
doc/src/Section_commands.txt
View File

@ -531,7 +531,8 @@ package"_Section_start.html#start_3.
"dump nc"_dump_nc.html,
"dump nc/mpiio"_dump_nc.html,
"group2ndx"_group2ndx.html,
"ndx2group"_group2ndx.html :tb(c=3,ea=c)
"ndx2group"_group2ndx.html,
"temper/grem"_temper_grem.html :tb(c=3,ea=c)
:line
@ -631,7 +632,7 @@ USER-INTEL, k = KOKKOS, o = USER-OMP, t = OPT.
"rigid/npt (o)"_fix_rigid.html,
"rigid/nve (o)"_fix_rigid.html,
"rigid/nvt (o)"_fix_rigid.html,
"rigid/small (o)"_fix_rigid.html,
<"rigid/small (o)"_fix_rigid.html,
"rigid/small/nph"_fix_rigid.html,
"rigid/small/npt"_fix_rigid.html,
"rigid/small/nve"_fix_rigid.html,

72
doc/src/fix_grem.txt
View File

@ -31,52 +31,51 @@ fix fxgREM all grem 502 -0.15 -80000 fxnvt :pre
This fix implements the molecular dynamics version of the generalized
replica exchange method (gREM) originally developed by "(Kim)"_#Kim,
which uses non-Boltzmann ensembles to sample over first order phase
transitions. The is done by defining replicas with an enthalpy dependent
effective temperature
transitions. The is done by defining replicas with an enthalpy
dependent effective temperature
:c,image(Eqs/fix_grem.jpg)
with {eta} negative and steep enough to only intersect the characteristic
microcanonical temperature (Ts) of the system once, ensuring a unimodal
enthalpy distribution in that replica. {Lambda} is the intercept and
effects the generalized ensemble similar to how temperature effects a
Boltzmann ensemble. {H0} is a reference enthalpy, and is typically set
as the lowest desired sampled enthalpy. Further explanation can be
found in our recent papers "(Malolepsza)"_#Malolepsza.
with {eta} negative and steep enough to only intersect the
characteristic microcanonical temperature (Ts) of the system once,
ensuring a unimodal enthalpy distribution in that replica. {Lambda} is
the intercept and effects the generalized ensemble similar to how
temperature effects a Boltzmann ensemble. {H0} is a reference
enthalpy, and is typically set as the lowest desired sampled enthalpy.
Further explanation can be found in our recent papers
"(Malolepsza)"_#Malolepsza.
This fix requires a Nose-Hoover thermostat fix reference passed to the
grem as {thermostat-ID}. Two distinct temperatures exist in this generalized
ensemble, the effective temperature defined above, and a kinetic
temperature that controls the velocity distribution of particles as
usual. Either constant volume or constant pressure algorithms can be
used.
grem as {thermostat-ID}. Two distinct temperatures exist in this
generalized ensemble, the effective temperature defined above, and a
kinetic temperature that controls the velocity distribution of
particles as usual. Either constant volume or constant pressure
algorithms can be used.
The fix enforces a generalized ensemble in a single replica
only. Typically, this ideaology is combined with replica
exchange with replicas differing by {lambda} only
for simplicity, but this is not required. A multi-replica
simulation can be run within the LAMMPS environment using the
"grem"_temper_grem.html command. This utilizes LAMMPS partition
mode and requires the number of available processors be
on the order of the number of desired replicas. A 100-replica
simulation would require at least 100 processors (1 per world
at minimum). If a many replicas are needed on a small number of
processors, multi-replica runs can be run outside of LAMMPS.
An example of this can be found in examples/USER/misc/grem and
has no limit on the number of replicas per processor. However,
this is very inefficient and error prone and should be avoided
if possible.
only. Typically, this ideaology is combined with replica exchange with
replicas differing by {lambda} only for simplicity, but this is not
required. A multi-replica simulation can be run within the LAMMPS
environment using the "temper/grem"_temper_grem.html command. This
utilizes LAMMPS partition mode and requires the number of available
processors be on the order of the number of desired replicas. A
100-replica simulation would require at least 100 processors (1 per
world at minimum). If a many replicas are needed on a small number of
processors, multi-replica runs can be run outside of LAMMPS. An
example of this can be found in examples/USER/misc/grem and has no
limit on the number of replicas per processor. However, this is very
inefficient and error prone and should be avoided if possible.
In general, defining the generalized ensembles is unique for every
system. When starting a many-replica simulation without any knowledge of
the underlying microcanonical temperature, there are several tricks we
have utilized to optimize the process. Choosing a less-steep {eta}
system. When starting a many-replica simulation without any knowledge
of the underlying microcanonical temperature, there are several tricks
we have utilized to optimize the process. Choosing a less-steep {eta}
yields broader distributions, requiring fewer replicas to map the
microcanonical temperature. While this likely struggles from the same
sampling problems gREM was built to avoid, it provides quick insight to
Ts. Initially using an evenly-spaced {lambda} distribution identifies
regions where small changes in enthalpy lead to large temperature
changes. Replicas are easily added where needed.
sampling problems gREM was built to avoid, it provides quick insight
to Ts. Initially using an evenly-spaced {lambda} distribution
identifies regions where small changes in enthalpy lead to large
temperature changes. Replicas are easily added where needed.
:line
@ -97,7 +96,8 @@ LAMMPS"_Section_start.html#start_3 section for more info.
[Related commands:]
"grem"_temper_grem.html, "fix nvt"_fix_nh.html, "fix npt"_fix_nh.html, "thermo_modify"_thermo_modify.html
"temper/grem"_temper_grem.html, "fix nvt"_fix_nh.html, "fix
npt"_fix_nh.html, "thermo_modify"_thermo_modify.html
[Default:] none

69
doc/src/temper_grem.txt
View File

@ -6,11 +6,11 @@
:line
grem command :h3
temper/grem command :h3
[Syntax:]
grem N M lambda fix-ID thermostat-ID seed1 seed2 index :pre
temper/grem N M lambda fix-ID thermostat-ID seed1 seed2 index :pre
N = total # of timesteps to run
M = attempt a tempering swap every this many steps
@ -23,25 +23,26 @@ index = which temperature (0 to N-1) I am simulating (optional) :ul
[Examples:]
grem 100000 1000 ${lambda} fxgREM fxnvt 0 58728
grem 40000 100 ${lambda} fxgREM fxnpt 0 32285 ${walkers} :pre
temper/grem 100000 1000 ${lambda} fxgREM fxnvt 0 58728
temper/grem 40000 100 ${lambda} fxgREM fxnpt 0 32285 ${walkers} :pre
[Description:]
Run a parallel tempering or replica exchange simulation in LAMMPS partition
mode using multiple generalized replicas (ensembles) of a system defined by
"fix grem"_fix_grem.html. Two or more replicas must be used.
Run a parallel tempering or replica exchange simulation in LAMMPS
partition mode using multiple generalized replicas (ensembles) of a
system defined by "fix grem"_fix_grem.html. Two or more replicas must
be used.
This command is a modification of the "temper"_temper.html command and has
the same dependencies, restraints, and input variables which are discussed
there in greater detail.
This command is a modification of the "temper"_temper.html command and
has the same dependencies, restraints, and input variables which are
discussed there in greater detail.
Instead of temperature, this command performs replica exchanges in lambda
as per the generalized ensemble enforced by "fix grem"_fix_grem.html.
The desired lambda is specified by {lambda}, which is typically a
variable previously set in the input script, so that each partition is
assigned a different temperature. See the "variable"_variable.html
command for more details. For example:
Instead of temperature, this command performs replica exchanges in
lambda as per the generalized ensemble enforced by "fix
grem"_fix_grem.html. The desired lambda is specified by {lambda},
which is typically a variable previously set in the input script, so
that each partition is assigned a different temperature. See the
"variable"_variable.html command for more details. For example:
variable lambda world 400 420 440 460
fix fxnvt all nvt temp 300.0 300.0 100.0
@ -49,27 +50,27 @@ fix fxgREM all grem ${lambda} -0.05 -50000 fxnvt
temper 100000 100 ${lambda} fxgREM fxnvt 3847 58382 :pre
would define 4 lambdas with constant kinetic temperature but unique
generalized temperature, and assign one of them to
"fix grem"_fix_grem.html used by each replica, and to the grem command.
generalized temperature, and assign one of them to "fix
grem"_fix_grem.html used by each replica, and to the grem command.
As the gREM simulation runs for {N} timesteps, a swap
between adjacent ensembles will be attempted every {M} timesteps. If
{seed1} is 0, then the swap attempts will alternate between odd and
even pairings. If {seed1} is non-zero then it is used as a seed in a
random number generator to randomly choose an odd or even pairing each
time. Each attempted swap of temperatures is either accepted or
rejected based on a Metropolis criterion, derived for gREM by
"(Kim)"_#Kim, which uses {seed2} in the random number generator.
As the gREM simulation runs for {N} timesteps, a swap between adjacent
ensembles will be attempted every {M} timesteps. If {seed1} is 0,
then the swap attempts will alternate between odd and even pairings.
If {seed1} is non-zero then it is used as a seed in a random number
generator to randomly choose an odd or even pairing each time. Each
attempted swap of temperatures is either accepted or rejected based on
a Metropolis criterion, derived for gREM by "(Kim)"_#Kim, which uses
{seed2} in the random number generator.
File management works identical to the "temper"_temper.html command.
Dump files created by this fix contain continuous trajectories and
require post-processing to obtain per-replica information.
The last argument {index} in the grem command is optional and is
used when restarting a run from a set of restart files (one
for each replica) which had previously swapped to new lambda.
This is done using a variable. For example if the
log file listed the following for a simulation with 5 replicas:
The last argument {index} in the grem command is optional and is used
when restarting a run from a set of restart files (one for each
replica) which had previously swapped to new lambda. This is done
using a variable. For example if the log file listed the following for
a simulation with 5 replicas:
500000 2 4 0 1 3 :pre
@ -77,9 +78,9 @@ then a setting of
variable walkers world 2 4 0 1 3 :pre
would be used to restart the run with a grem command like the
example above with ${walkers} as the last argument. This functionality
is identical to "temper"_temper.html.
would be used to restart the run with a grem command like the example
above with ${walkers} as the last argument. This functionality is
identical to "temper"_temper.html.
:line

3
src/USER-MISC/README
View File

@ -29,6 +29,7 @@ bond_style harmonic/shift/cut, Carsten Svaneborg, science at zqex.dk, 8 Aug 11
compute ackland/atom, Gerolf Ziegenhain, gerolf at ziegenhain.com, 4 Oct 2007
compute basal/atom, Christopher Barrett, cdb333 at cavs.msstate.edu, 3 Mar 2013
compute temp/rotate, Laurent Joly (U Lyon), ljoly.ulyon at gmail.com, 8 Aug 11
compute pressure/grem, David Stelter, dstelter@bu.edu, 22 Nov 16
dihedral_style cosine/shift/exp, Carsten Svaneborg, science at zqex.dk, 8 Aug 11
dihedral_style fourier, Loukas Peristeras, loukas.peristeras at scienomics.com, 27 Oct 12
dihedral_style nharmonic, Loukas Peristeras, loukas.peristeras at scienomics.com, 27 Oct 12
@ -39,6 +40,7 @@ fix addtorque, Laurent Joly (U Lyon), ljoly.ulyon at gmail.com, 8 Aug 11
fix ave/correlate/long, Jorge Ramirez (UPM Madrid), jorge.ramirez at upm.es, 21 Oct 2015
fix flow/gauss, Joel Eaves (CU Boulder), Joel.Eaves@Colorado.edu, 23 Aug 2016
fix gle, Michele Ceriotti (EPFL Lausanne), michele.ceriotti at gmail.com, 24 Nov 2014
fix grem, David Stelter, dstelter@bu.edu, 22 Nov 16
fix imd, Axel Kohlmeyer, akohlmey at gmail.com, 9 Nov 2009
fix ipi, Michele Ceriotti (EPFL Lausanne), michele.ceriotti at gmail.com, 24 Nov 2014
fix pimd, Yuxing Peng (U Chicago), yuxing at uchicago.edu, 24 Nov 2014
@ -65,3 +67,4 @@ pair_style meam/sw/spline, Robert Rudd (LLNL), robert.rudd at llnl.gov, 1 Oct 12
pair_style morse/smooth/linear, Stefan Paquay (TU Eindhoven), stefanpaquay at gmail.com, 29 Feb 16
pair_style srp, Tim Sirk, tim.sirk at us.army.mil, 21 Nov 14
pair_style tersoff/table, Luca Ferraro, luca.ferraro@caspur.it, 1 Dec 11
temper/grem, David Stelter, dstelter@bu.edu, 22 Nov 16

6
src/USER-MISC/fix_grem.cpp
View File

@ -17,9 +17,9 @@
------------------------------------------------------------------------- */
/* ----------------------------------------------------------------------
Contributing author: Edyta Malolepsza (Broad Institute)
Contributing author: David Stelter (Boston University)
Contributing author: Tom Keyes (Boston University)
Contributing authors: Edyta Malolepsza (Broad Institute)
David Stelter (Boston University)
Tom Keyes (Boston University)
------------------------------------------------------------------------- */
#include <string.h>

4
src/USER-MISC/temper_grem.cpp
View File

@ -12,9 +12,7 @@
------------------------------------------------------------------------- */
/* ----------------------------------------------------------------------
Contributing author: Mark Sears (SNL)
Modified for gREM by David Stelter (BU)
Contributing author: David Stelter (BU)
------------------------------------------------------------------------- */
#include <math.h>