gparm is a program for generating oodles of text-based input files for other programs. It facilitates development and management of parametric analyses.

The general use case is where you need to run a program---say, a simulation of some kind---dozens, thousands, or even millions of times, using a variety of inputs.

gparm is Perl script embodied in a single file. Put it wherever you want. In Linux, it should probably go in a bin directory somewhere (e.g., /usr/local/bin or $HOME/bin).

You will need a working Perl 5 interpreter on your machine, and may need to add two modules that are sometimes not part of the base Perl installation:

• Text::Template
• Parallel::ForkManager

In Linux it's easy to make a script file executable, so you generally only have to name the file to run it, just like any other command, as long as the script file is in a directory on your search path. E.g.,

	gparm [arguments...]

If you don't want to put it in a bin directory, just run it from wherever you put it. E.g.,

	./gparm [arguments...]

or

	$HOME/tools/gparm [arguments...]

In Windows, you'll have to name gparm as an argument to the perl command:

	perl gparm [arguments...]

To do parametric analysis with gparm, create two files: 1) a template file formatted in the proper syntax for your simulation program, but having replaceable parameters in place of any real input values you want to vary, and 2) a parameter file that holds all your input sets. Then, combine the two with gparm:

	gparm -t my.template.file -p my.parmameter.file

The template file is simply a text file in the format required by your software, with specific inputs replaced by special tags that use replaceable parameters. For example (completely contrived):

	set iterations       {$iterations}
	set use_fast_library {$use_fastlib}
	set initial_estimate {$start_value <= 0 ? 1.0 : $start_value}

The tags inside the curlies {} are simply snippets of Perl code that may contain nothing more than the name of a scalar variable or may contain programmed logic of arbitrary complexity. The values in the variables get defined by the parameter file.

The parameter file is simply a CSV file holding a rectangular block of data, one column per parameter, one row per run set. The first row holds the parameter names (as they're used in the template file). E.g.,

	ID, iterations, use_fastlib, start_value
	1, 12, yes, 1
	2, 12, yes, 10
	3, 20, yes, 1
	4, 20, yes, 10
	5, 12, no, 1
	6, 12, no, 10
	7, 20, no, 1
	8, 20, no, 10

The gparm user manual covers many other topics and describes gparm's capabilities in detail:

• Changing snippet delimiters from {} to other characters
• Assigning default parameter values in templates
• Control over generated filenames
• Automatic compression of generated files
• Breaking up templates into multiple files
• Multiprocessing for speed on multicore computers
• Use of persistent storage to accommodate large datasets in templates
• Use of special (non-parameter) variables inside templates
• Availability of Perl functions in templates
• Error messages and debugging of templates