Important: The trajectory is generated according to a user-provided potential energy routine which recieves a coordinate vector Q and returns a potential value V(Q). The routine must be callable through Python. If the routine is written in Fortran we suggest using the f2py Fortran to Python interface generator available at the NumPy/SciPy homepage. Another possibility could be a wrapper routine which sends the coordinate vector via pipes to a background process which passes the potential value back to the wrapper. See also 'usersource.py'.
The trajectory generated by chkpes.py resambles a Boltzman distribution η(Q) = exp(-βV(Q)) with β-1 being the temperature.
----------------------------------------------------------------------------
Purpose: Use a Monte-Carlo sampling to check a sampled PES file against an analytic potential routine.
Usage: chkpes.py [-D -n -mnd -deb -w -t -ver -h -? ] inputfile.inp
-D <dir> Denotes the directory where files are written to (name in .inp file ignored). -n <int> Use <int> Monte-Carlo steps (value in .inp is ignored). -mnd Make name directory. -deb Run in debug mode (more detailed logging). -w Allow to overwrite existing data. -t <file> Run with a given trajectory stored in <file>. -ver Print version info. -h -? Print this help text. ----------------------------------------------------------------------------
|
||
name = S | The 'name' directory to which output files are written. | |
user-source = S | The path (relative or absolute) to the module
containing the potential energy routine. See
using the 'user-source' keyword. |
|
temperature=R(,S) | The 'temperature' (including the Botlzmann constant) used to generate the Monte-Carlo trajectory. S may be a unit: "cm-1", "eV" or "au". Default: S = "au". | |
steps = I | The number of accepted Monte-Carlo steps, i.e, the length of the trajectory. | |
pesdir = S | The directory containing the PES file generated by MCTDH. 'pesdir' can be omitted if 'walk-only' is set. In this case 'dvrdir' has to be given, | |
dvrdir = S | The directory containing the DVR file generated by MCTDH. If 'pesdir' is given, 'dvrdir' defaults to 'pesdir'. Note: For generating the trajectory there is no check whether a given DVR is compatatible with a given PES file. | |
|
||
potential-routine = S | The name of the potential routine in the module set by the 'user-source' keyword (see above). Deufault: S = "potential" | |
vminmax-exe = S | The name of the executable of the MCTDH analysis program 'vminmax'. Default: S = "vminmax84". | |
nham = I | Number of the Hamiltonian within the PES file. Default: I = 1 (system). | |
trajectory = S | Path to a previously calculated trajectory (same as option -t). If 'trajectory' is set, the Monte-Carlo run is omitted. Instead, ony the 'energies' file in the name directory is updated and, if required, vminmax is invoked. | |
overwrite | Allow overwrite of existng files in the 'name' directory. Similar is option -w in the command line. | |
walk-only | Just obtain a Monte-Carlo trajectory and stop thereafter. Do not invoke vminmax. | |
wrcoord | Write the (real) coordinate vector for each step to the 'coordinates' file (the tracectory is usually stored as the interger grid index in file 'dvrindex'). |
Example: A system containing four DOF, (Q1, Q2, Q3, Q4), where the initial dvr-index of Q1 is 7, the dvr-index of Q2 is 6, etc. In Addition, the dvr-index of Q4 is set to 13 and kept constand during the calculation.
INIT-SECTION
#coord.label initial point const?
Q1 7
Q2 6
Q3 6
Q4 13 fix
end-init-section
Note: Once the files 'dvrindex' and 'energies' are created, vminmax is used in a standard way and could also be run from the command line (see options -t and -c of vminmax).