2006-09-22 00:22:34 +08:00
|
|
|
<HTML>
|
|
|
|
|
<CENTER><A HREF = "http://lammps.sandia.gov">LAMMPS WWW Site</A> - <A HREF = "Manual.html">LAMMPS Documentation</A> - <A HREF = "Section_commands.html#comm">LAMMPS Commands</A>
|
|
|
|
|
</CENTER>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
<HR>
|
|
|
|
|
|
|
|
|
|
<H3>temper command
|
|
|
|
|
</H3>
|
|
|
|
|
<P><B>Syntax:</B>
|
|
|
|
|
</P>
|
|
|
|
|
<PRE>temper N M temp fix-ID seed1 seed2 index
|
|
|
|
|
</PRE>
|
|
|
|
|
<UL><LI>N = total # of timesteps to run
|
|
|
|
|
<LI>M = attempt a tempering swap every this many steps
|
|
|
|
|
<LI>temp = initial temperature for this ensemble
|
|
|
|
|
<LI>fix-ID = ID of the fix that will control temperature during the run
|
|
|
|
|
<LI>seed1 = random # seed used to decide on adjacent temperature to partner with
|
|
|
|
|
<LI>seed2 = random # seed for Boltzmann factor in Metropolis swap
|
|
|
|
|
<LI>index = which temperature (0 to N-1) I am simulating (optional)
|
|
|
|
|
</UL>
|
|
|
|
|
<P><B>Examples:</B>
|
|
|
|
|
</P>
|
|
|
|
|
<PRE>temper 100000 100 $t tempfix 0 58728
|
|
|
|
|
temper 40000 100 $t tempfix 0 32285 $w
|
|
|
|
|
</PRE>
|
|
|
|
|
<P><B>Description:</B>
|
|
|
|
|
</P>
|
|
|
|
|
<P>Run a parallel tempering (replica exchange) simulation of multiple
|
2010-10-01 04:38:58 +08:00
|
|
|
ensembles (replicas) of a system on multiple partitions of processors.
|
|
|
|
|
</P>
|
|
|
|
|
<P>The processor partitions are defined using the -partition command-line
|
|
|
|
|
switch (see <A HREF = "Section_start.html#2_6">this section</A>). Each partition
|
|
|
|
|
(replica) can be assigned one or more processors. Note that if you
|
|
|
|
|
have MPI installed, you can run a multi-partition simulation on more
|
|
|
|
|
partitions than you have physical processors, e.g you can run a
|
|
|
|
|
10-replica simulation on one or two processors. You will simply
|
|
|
|
|
not get the performance speed-up you would typically see with one or
|
|
|
|
|
more physical processors per replica. See <A HREF = "Section_howto.html#4_5">this
|
|
|
|
|
section</A> of the manual for further discussion.
|
|
|
|
|
</P>
|
|
|
|
|
<P>Each ensemble's temperature is controlled at a different value by a
|
|
|
|
|
fix with ID <I>fix-ID</I> that controls temperature. Possible fix styles
|
|
|
|
|
are <A HREF = "fix_nh.html">nvt</A>, <A HREF = "fix_nh.html">temp/berendsen</A>,
|
2008-03-24 22:45:46 +08:00
|
|
|
<A HREF = "fix_langevin.html">langevin</A> and <A HREF = "fix_temp_rescale.html">temp/rescale</A>.
|
|
|
|
|
The desired temperature is specified by <I>temp</I>, which is typically a
|
|
|
|
|
variable previously set in the input script, so that each partition is
|
|
|
|
|
assigned a different temperature. See the <A HREF = "variable.html">variable</A>
|
2010-10-01 04:38:58 +08:00
|
|
|
command for more details. For example:
|
2006-09-22 00:22:34 +08:00
|
|
|
</P>
|
2010-10-01 04:38:58 +08:00
|
|
|
<PRE>variable t world 300.0 310.0 320.0 330.0
|
|
|
|
|
fix myfix all nvt $t $t 100.0
|
|
|
|
|
temper 100000 100 $t myfix 3847 58382
|
2006-09-22 00:22:34 +08:00
|
|
|
</PRE>
|
2010-10-01 04:38:58 +08:00
|
|
|
<P>would define 4 temperatures, and assign one of them to the thermostat
|
|
|
|
|
used by each replica, and to the temper command.
|
|
|
|
|
</P>
|
2006-09-22 00:22:34 +08:00
|
|
|
<P>As the tempering simulation runs for <I>N</I> timesteps, a swap between
|
|
|
|
|
adjacent ensembles will be attempted every <I>M</I> timesteps. If <I>seed1</I>
|
|
|
|
|
is 0, then the swap attempts will alternate between odd and even
|
|
|
|
|
pairings. If <I>seed1</I> 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 Boltzmann-weighted Metropolis criterion which uses
|
|
|
|
|
<I>seed2</I> in the random number generator.
|
|
|
|
|
</P>
|
2010-08-24 22:39:37 +08:00
|
|
|
<P>As a tempering run proceeds, multiple log files and screen output
|
|
|
|
|
files are created, one per replica. By default these files are named
|
|
|
|
|
log.lammps.M and screen.M where M is the replica number from 0 to N-1,
|
|
|
|
|
with N = # of replicas. See the <A HREF = "Section_start.html#2_6">section on command-line
|
|
|
|
|
switches</A> for info on how to change these
|
|
|
|
|
names.
|
|
|
|
|
</P>
|
|
|
|
|
<P>The main screen and log file (log.lammps) will list information about
|
|
|
|
|
which temperature is assigned to each replica at each thermodynamic
|
|
|
|
|
output timestep. E.g. for a simulation with 16 replicas:
|
|
|
|
|
</P>
|
|
|
|
|
<PRE>Running on 16 partitions of processors
|
|
|
|
|
Step T0 T1 T2 T3 T4 T5 T6 T7 T8 T9 T10 T11 T12 T13 T14 T15
|
|
|
|
|
0 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15
|
|
|
|
|
500 1 0 3 2 5 4 6 7 8 9 10 11 12 13 14 15
|
|
|
|
|
1000 2 0 4 1 5 3 6 7 8 9 10 11 12 14 13 15
|
|
|
|
|
1500 2 1 4 0 5 3 6 7 9 8 10 11 12 14 13 15
|
|
|
|
|
2000 2 1 3 0 6 4 5 7 10 8 9 11 12 14 13 15
|
|
|
|
|
2500 2 1 3 0 6 4 5 7 11 8 9 10 12 14 13 15
|
|
|
|
|
...
|
|
|
|
|
</PRE>
|
|
|
|
|
<P>The column headings T0 to TN-1 mean which temperature is currently
|
|
|
|
|
assigned to the replica 0 to N-1. Thus the columns represent replicas
|
|
|
|
|
and the value in each column is its temperature (also numbered 0 to
|
|
|
|
|
N-1). For example, a 0 in the 4th column (column T3, step 2500) means
|
|
|
|
|
that the 4th replica is assigned temperature 0, i.e. the lowest
|
|
|
|
|
temperature. You can verify this time sequence of temperature
|
|
|
|
|
assignments for the Nth replica by comparing the Nth column of screen
|
|
|
|
|
output to the thermodynamic data in the corresponding log.lammps.N or
|
|
|
|
|
screen.N files as time proceeds.
|
|
|
|
|
</P>
|
|
|
|
|
<P>The last argument <I>index</I> in the temper command is optional and is
|
|
|
|
|
used when restarting a tempering run from a set of restart files (one
|
|
|
|
|
for each replica) which had previously swapped to new temperatures.
|
|
|
|
|
The <I>index</I> value (from 0 to N-1, where N is the # of replicas)
|
|
|
|
|
identifies which temperature the replica was simulating on the
|
|
|
|
|
timestep the restart files were written. Obviously, this argument
|
|
|
|
|
must be a variable so that each partition has the correct value. Set
|
|
|
|
|
the variable to the <I>N</I> values listed in the log file for the previous
|
|
|
|
|
run for the replica temperatures at that timestep. For example if the
|
|
|
|
|
log file listed the following for a simulation with 5 replicas:
|
2006-09-22 00:22:34 +08:00
|
|
|
</P>
|
|
|
|
|
<PRE>500000 2 4 0 1 3
|
|
|
|
|
</PRE>
|
|
|
|
|
<P>then a setting of
|
|
|
|
|
</P>
|
|
|
|
|
<PRE>variable w proc 2 4 0 1 3
|
|
|
|
|
</PRE>
|
|
|
|
|
<P>would be used to restart the run with a tempering command like the
|
|
|
|
|
example above with $w as the last argument.
|
|
|
|
|
</P>
|
2010-10-01 04:38:58 +08:00
|
|
|
<HR>
|
|
|
|
|
|
|
|
|
|
<P><B>Restrictions:</B>
|
|
|
|
|
</P>
|
|
|
|
|
<P>This command can only be used if LAMMPS was built with the "replica"
|
|
|
|
|
package. See the <A HREF = "Section_start.html#2_3">Making LAMMPS</A> section for
|
|
|
|
|
more info on packages.
|
2006-09-22 00:22:34 +08:00
|
|
|
</P>
|
|
|
|
|
<P><B>Related commands:</B>
|
|
|
|
|
</P>
|
2010-10-01 04:38:58 +08:00
|
|
|
<P><A HREF = "variable.html">variable</A>, <A HREF = "prd.html">prd</A>, <A HREF = "neb.html">neb</A>
|
2006-09-22 00:22:34 +08:00
|
|
|
</P>
|
|
|
|
|
<P><B>Default:</B> none
|
|
|
|
|
</P>
|
|
|
|
|
</HTML>
|
