Learning Resources for SSRDE
Social Sciences Research and Development Environment
Our compute cluster is designed to be as simple as possible, so the learning overhead is low.
If you find that our cluster is not capable of doing the advanced cluster work you might need, or you have out grown our environment(we have done our job correctly!!!) we can contact SDSC and obtain access to the much larger cluster resources at SDSC.
You need to SSH to ssrde.ucsd.edu head node first (this sets up your account the first time you log in).
Take a look at the message of the Day when you log in . . there is useful information there!
The following Video might help you if you are unfamiliar with command line computing.
Intro to command line video tutorial
You can try basic commands on the head node and test your code and batch submission files.
Our guide might be useful for basic instruction SSRDE Quickstart guide
On the head node you can use the following command to see the current job queue - Yes the queue is sometimes very full, and job wait times are impossible to control, however the queue knows who has jobs in the queue and will add new users jobs as a priority over users that have lots of jobs queued up.
Detailed squeue commands - squeue -o "%.18i %.18P %.18j %.12u %.5t %.10M %.6D %.6C %R %N"
On the head node you can use the following command to see the status of the nodes.
Detailed sinfo commands - sinfo -o "%20P %5a %.10l %.12s %12r %5h %10g %.5D %11T %R %N"
You will almost certainly need to load modules to run the software versions you need. You can test module loading on the head node and basic code testing.
Take a look here for more detailed info Modules: Managing software on SSRDE
At some point you are going to need to move your data to your SSRDE home directory OR have your lab share mounted by SSCF.
We are compiling a list of Batch file examples to help everyone to get started - It is a work in progress. If you have something to contribute, email us at the email address at the top of the page.
9. After you submit your first job, use the sinfo and squeue commands above to see how your job is interacting with the cluster.