SIParCS 2019 - Thomas Johnson III
Thomas Hilton Johnson III, Elizabeth City State University
Development of Interactive HPC Documentation
The Cheyenne High Performance Computing (HPC) system and Casper data analysis and visualization cluster are heavily used for their resources by scientists across the world. The Computational and Information Systems Lab (CISL) Consulting Group (CSG) is tasked with ensuring that users properly receive the assistance necessary to utilize these HPC and cluster resources. To do so, CSG develops documentation to provide users with information on how utilize software effectively and updates CISL webpages with content that will assist users. To add to the current documentation, three Jupyter Notebooks have been added to introduce interactive documentation to further instruct users as to how to use the software and resources for the HPC system and data analysis and visualization cluster. Jupyter Notebooks are runnable pages of code that are built on IPython and can support a large selection of programming languages. These Jupyter notebooks include an introduction to how to use Cheyenne, how to implement TensorFlow and how to use the tools installed on the HPC and cluster systems to monitor the memory usage of scheduled jobs. An Lmod web application has also been crafted to allow users to view the software modules available on both the Cheyenne and Casper systems without logging in. The Lmod web application uses Python Flask and JavaScript to develop a visual tree of the available software as well as display help information for said software. Efforts are underway to deploy the application within CISL’s existing HPC and cluster documentation websites.
Mentors: Brian Vanderwende, Davide Del Vento, Kevin Paul