README neurosys/neurondiz

The files in this directory can be used to build a parallel program
for viewing the output of the solver.  

1. Prerequisites
   You must have X-Windows and Argonne National Labs MPE extensions
   to MPI.  The MPE extensions can be downloaded from

       http://www.mcs.anl.gov/mpi/mpich.

2. Compiling
   a) Define the following macros in the Makefile:  CFLAGS, LDFLAGS,
      INCLUDE, and LIB.  Define the CLICK_STEP CFLAG if you want to
      step through the displays generated by neurondiz.  If not,
      define the SLEEP_INTERVAL macro in const.h -- this controls the 
      length of time the processes sleep before updating the display.
      1 second is probably as long as you'll want this.
   b) make

3. Execution
   a) Start the program as you would any mpi program.
   b) The program prompts for a source of input, which can be
      either stdin or a file.  In either case, the following
      input is required:
          the number of columns of neurons
          the number of rows of neurons
          the number of variables per neuron (currently 1)
          the number of timesteps
          the display frequency
          the type of the data:  "txt" for text, "bin" for binary
          the "root" of the input filename
   c) Input data(s).  The input data should have been produced by
      the solver compiled with the DIST_OUTPUT option -- even if
      you only used 1 process.  The program "splitter" in the ../utils
      directory can be used to convert data produced by the solver 
      without DIST_OUTPUT to the proper format.
   d) The program waits for a mouseclick after drawing the initial
      display -- regardless of whether the CLICK_STEP macro was
      defined.
   e) After the second display has been drawn, the program will
      either wait to draw subsequent displays (CLICK_STEP defined)
      or it will proceed to draw the remaining displays.
