Introduction

OldMcData - The Data Farmer (OMD) is a software application designed to setup, start, and manage data farming runs, from running large simulation experiments on a distributed computer cluster to multiple replications of a single excursion on a single machine. For runs on a distributed computing cluster, OMD uses Condor (http://www.cs.wisc.edu/condor/), an open-source distributed computing environment.

This document, as the name implies, is a guide to get you started using OMD quickly. The first section gives instructions on how to install OMD, and the following section tells you how to run OMD, using a sample simulation. More details on how to setup OMD for other models can be found in the OMD User's Manual.

Installation

To install, just unzip the omd1.1.zip file to any directory (one with no spaces), preferably at the top level, e.g., c:\oldmcdata on Windows or /usr/oldmcdata on Linux.

Once unzipped, the file structure should look like this (where root is the directory you unzipped into):

• root -we'll also call this the "oldmcdata" directory
• oldmcdata.start.bat - a windows bat file to start OMD from the command line
• oldmcdata.start.sh - a unix shell script to start OMD from the command line
• oldmcdata.config.xml - OMD's configuration file
• lib - the library directory containing all the supporting jars used by OMD
• models
  • simpleplanner - a directory containing all the files needed by simpleplanner to run. simpleplanner is a simple task/resource simulation written in Scala (2.7.7)
• studies
  • test-simpleplanner - a "study" directory containing all the files needed for an OMD run

The only configuration needed is to modify the oldmcdata.config.xml file to reflect where you unzipped omd1.1.zip. Just change all the file paths in the config file from "/Users/stephenupton/Work/oldmcdata/1.1" to your oldmcdata directory. You will also need to change the command executable to a ".bat" file if you are running on windows. Currently, you do not need to worry about the Condor entries at the top of the config file, unless you have Condor installed and want to experiment with that configuration.

Running OMD1.1

To run, open up a command or terminal window, change directories to your oldmcdata directory and type:

./oldmcdata.start.sh <whereyouunzippedomd>/studies/test-simpleplanner study.xml

for unix/mac, and

oldmcdata.start.bat <whereyouunzippedomd>/studies/test-simpleplanner study.xml

for windows (note you need to use the forward slash here as well -something to fix in a later version)

After the run, you should have a directory called "Output" and a directory called "run" in the study directory (that's the <whereyouunzippedomd>/studies/test-simpleplanner directory). In the Output directory, you should have a bunch of output files called *.csv and *.tasks. In the run directory, you should have all the simpleplanner files, plus the excursion files generated from the study.csv file.

That's it! You have successfully tested your installation of OMD. More details can be found in the OMD User's Manual:

• on how to configure OMD for other models,
• how the study.xml file is structured,
• how to use your own experimental design,
• and a description of several other algorithms you can use for your own experiments.