Use prunner
-----------------------------------------------------------

Requirements
============

The following libraries are required
    
    * cherrypy (http://www.cherrypy.org/) (v3.0+)
    * cheetah (http://www.cheetahtemplate.org/) (v2.01+)
    * paramiko (http://www.lag.net/paramiko/) (v1.76+)
    * matplotlib (http://matplotlib.sourceforge.net/) (v0.99+)

Run prunner
============

The simplest example to run prunner is 

bin/prunner -c clusterpage.ini -r prunner.ini -f cmds.txt

where <clusterpage.ini> is a configuration file to setup CherryPy web server; 
<prunner.ini> is to configure the parallel runner and ssh key/password to connect 
with the remote clusters; <cmds.txt> is the file to specify which commands to run.

In <prunner.ini>, it also requires a text file to specify the cluster names. The 
format of this file is

cluster-1  <max # of threads>
cluster-2  <max # of threads>
...

The first column specifies the name of each cluster, the second column indicates 
the maximum number of jobs that can be run simultaneously on the corresponding 
cluster. For example, for multi-core cluster, you might want to run multiple 
single-thread jobs on it.

Please see the example configuration files (i.e. clusterpage.ini, prunner.ini, cmds.txt
and clusters.txt) for more details.

