Moab Advanced Linux Batch
From Cornell CAC Documentation
Revision as of 00:55, 4 March 2010 by Ajd27
- Matlab In Batch - How to run a Matlab program on the cluster.
- Programmatically Submitting Jobs and Checking Whether They are Done
- Batch Jobs for Apps Started by Graphical Interfaces
- Scripts to Run IPython in Batch - This is a unique way to take advantage of a cluster. You type commands at your console on your computer, but they execute on the remote cluster. For certain types of analysis, this could be very cool.
- Ways to run a single application many times.
- Farm-out-work - Use MPI to execute every line of a text file on a set of nodes. When one task ends, another starts behind it. This is appropriate if there are many tasks that take varying amounts of time to run.
- Run Multiple Copies of the Same Application - Start a set of processes simultaneously, using MPI, and wait for them to finish.
- Python Scripts for Multiple Tasks on a Node - This is equivalent to farm-out-work, but it runs on only one node at a time, sharing the multiple cores to run lots of tasks.
- Using IPython to Farm Work - This is a version of farm-out-work that uses IPython. It's cute and friendly for those who like Python, but it can take time and attention to install IPython.