The pedagogical approach of this workshop will be modeled on that of the Software and Data Carpentry communities and will utilize the following interactive components:
The content will be presented through a combination of speaking and live coding in front of tutorial participants. Presenters will both talk about HPC concepts as described in the outline, and demonstrate them in front of participants, with the expectation that participants follow along and run the same commands as the presenter.
Example: Logging onto a large-scale computing system
Presenter draws a diagram of a user's local computer and the login node of a remote computing system, and describes what is happening when users log in. She then demonstrates logging into the provided cluster using the "ssh" command; all participants use the same command to log in to the same cluster.
In addition to the content presentation, the tutorial will contain regular exercises where participants are expected to work individually or in groups on a set of challenges. There will be an exercise or independent activity for about every 15-20 minutes of instruction.
Example Exercise: Submitting jobs to a batch scheduler
submit-job.pbssubmits a request to run our script on the test filetest.csv:
python analyze.py test.csvModify the script
submit-job.pbsto instead submit a job to run the analysis on our actual input data,temperatures.csv. This job will need more than 1 GB, so you will also need to change the amount of memory thatsubmit-job.pbsis requesting.
Throughout the workshop, participants will be provided with two colors of sticky notes to serve as a visual indicator of progress: one color to indicate success or understanding, the other to indicate confusion or a question. This allows the lead presenter to respond to the audience at a glance. In addition, when presenters are not leading a session, they will circulate among participants to answer questions or address immediate issues. Finally, an online note-taking page (etherpad) will be set up for participants to take collaborative notes and chat with each other.
Exercises and demonstrations will be fully tested on the planned resources before the workshop.
The materials developed for this workshop will be similar in format to those developed by Software Carpentry. For an example of what the final materials will look like see:
- Software Carpentry lesson on the Unix Shell: http://swcarpentry.github.io/shell-novice/
To see an example of how lessons will incorporate live commands, diagrams, and follow-on exercises, see:
This tutorial requires access to a large scale computing system that supports remote login and uses a batch scheduler. We will have temporary accounts created for all partipants on the Oklahoma State University computing cluster for this purpose. We are also applying for an education allocation on XSEDE which, if granted, will complement the resources available through Oklahoma State.
Participants will need to bring their own laptop or portable computing device that can use WiFi. The tutorial will require using a bash shell (or similar tool); installation instructions will be part of the workshop materials.