#######################################################################
#                                                                     #
#  Example:  Find optimum step size                                   #
#                                                                     #
#######################################################################
#                                                                     #
#    XMD requires the user to choose a time step size when doing      #
#  dynamics.  The best step size is the largest one that is still     #
#  stable.  This optimum value can be found by trial and error.       #
#                                                                     #
#    This is done by running molecular dynamics with the              #
#  configuration and temperature that one wants to simulate.  A       #
#  small time step is used initially, and after 20 steps or so this   #
#  time step size is double.  This doubling repeats 5 times in the    #
#  following example.                                                 #
#                                                                     #
#    Meanwhile, during the molecular dynamics run, the clamp is set   #
#  to -1, and the energy is written out at every step.  This clamp    #
#  setting results in an adiabatic simulation, which means that the   #
#  total energy will remain constant for a stable time step.  When    #
#  the time step becomes too large, the total energy will instead     #
#  drift upward.                                                      #
#                                                                     #
#    You can see the value of the total energy by using the ESAVE     #
#  command, writing the energies out to a file at every step.  ESAVE  #
#  saves the energies in a text file.  In this example, the name of   #
#  the file is timestep.e.  Inside this text file, the first column   #
#  holds the step number, the second holds the total energy.  The     #
#  next two columns hold the potential and kinetic energy             #
#  respectively, which add together to give the total energy.         #
#                                                                     #
#######################################################################

# Trial and error to find optimum time step size

# Read potential for nial
read ../nial.txt

#  Use old neighbor search algorithm to accomodate small atomic array
nsearch sort

# Make repeating box and lattice (in units of a0)
box 6 6 6
particle 2
1 0.25 0.25 0.25
2 0.75 0.75 0.75
dup 5   1 0 0
dup 5   0 1 0
dup 5   0 0 1

# Scale up to units of angstroms (2.8712 unit cell)
scale 2.8712

# Save energies from every dynamics step in file "timestep.e'
esave 1 timestep.e

# Set particle masses (in atomic mass units)
select type 1
mass 58.71
select type 2
mass 26.982

# Set adiabtic simulation at starting temperature of 200K
clamp -1
itemp 200

# Vary time step

# Set initial timestep size variable
calc STEPSIZE = 4e-16


# Do 6 separate runs of 20 steps each
repeat 6

   # Set timestep size using dtime command
   dtime  STEPSIZE

   # 20 steps of dynamics (energies are written to tempstep.e file)
   cmd 20

   # Double the time step size
   calc   STEPSIZE = 2 * STEPSIZE

end