Input files

The following files are required in the directory where PATHSAMPLE is executed.

• pathdata: contains parameters to control the stationary point sampling and/or the kinetic analysis. See §4.
• odata.* files: the odata file is the control file for OPTIM. One odata file is needed for each type of OPTIM job performed by PATHSAMPLE. See §5.

• commit.data. This file is updated every time a committor probability calculation is performed. Initial committor probabilities are read from it if the file is detected. Otherwise, the initial values are set to zero and one as appropriate. If a PATHSAMPLE run ends abnormally it is possible for the number of entries in commit.data to be less than the number of local minima in min.data. In this case the missing values are automatically set to zero. This procedure does not work with some compilers if they behave incorrectly on reaching the end of file. In this case the missing zeros can simply be added using a text editor.

The following files may be required, depending upon the keywords set in pathdata:

• min.A, min.B and min.data: lists of A, B and intermediate minima, respectively. The first line of min.A is the number of A minima (free format), $N_{\rm A}$; on the following lines there must be $N_{\rm A}$ integers (free format), which specify all the local minima belonging to the A set. Hence, the A region can be changed simply by editing this file. min.B has the same format for the B minima. The numbering of the local minima must be consistent with the order of their coordinates in points.min (see below). However, unlike older versions of PATHSAMPLE, the A and B minima can appear in any order, facilitating regrouping schemes. The file min.data has the format:

  energy    frequency    pgorder    itx    ity    itz

  for each minimum, where $energy$ is the potential energy (as calculated by OPTIM), $frequency$ is the natural log of the product of positive eigenvalues from the mass-weighted Hessian (used to evaluate partition functions and rates constants), $pgorder$ is the order of the point group, and $itx, ity, itz$ are the (sorted) eigenvalues of the inertia tensor. The order of entries in min.data must agree with the min.A, min.B and points.min files.

• The file ts.data is analogous to the min.data file, but contains data for the transition states. The format is:

  energy    frequency    pgorder    min1    min2    itx    ity    itz

  where the entries are the same as for the minima, with the addition of min1 and min2, which identify the minima that the transition state connects. This file must be consistent with points.ts (see below).

• points.min: contains the Cartesian coordinates of all the minima that are specified in min.data in the same order. It is an unformatted direct access file, written with the Fortran specifier RECL=8*3*NATOMS.

• points.ts: as for points.min, but contains the transition state coordinates.

• nodes.info: must be present for runs on a distributed memory machine. The first line must contain an integer equal to the number of nodes available, and the names of these nodes must follow below. On the final line the userid must be provided. An example script for the qsub command could look like:
  

  #!/bin/csh

  #PBS -q h4

  cd $PBS_O_WORKDIR

  cat $PBS_NODEFILE $>$& output

  wc output $>$ nodes.info

  cat $PBS_NODEFILE $>>$ nodes.info

  echo $USER $>>$ nodes.info

  /home/bin/pathsample $>>$& output

  
  

• path.info.startup: an OPTIM path.info file from a connect run that connects an A minimum to a B minimum. Note that an output file is not required, unlike older versions of PATHSAMPLE. This file is needed to provide an initial path, and the file name startup is specified in the pathdata file by the STARTFROMPATH keyword. If the files min.A and min.B are absent, then STARTFROMPATH will create them, with a single minimum specified in each region, corresponding to the entries specified on the STARTFROMPATH line after the file extension startup. The rest of the minima read from the path.info.startup file will be placed in the intervening set. Note that for path.info files in the new OPTIM DUMPALLPATHS format, the two end point minima will generally not be the first and last entries. Since min.data contains only unique minima, whereas the path.info file may have duplicates, the positions of the starting and finishing minimum may not be obvious. A quick way to check these values is to run PATHSAMPLE first with CYCLES 0, find the positions of the two minima in min.data and correct the values in pathdata, remove min.*, points.* and *.data, and start the real PATHSAMPLE run.For a setup where the A and B sets are predefined in min.A and min.B a connecting pathway can be read using the keyword ADDPATH add in pathdata. A path will then be read from file add.

The min.data, ts.data, points.min and points.ts files are updated by PATHSAMPLE as new minima and transition states are found.