HPC everywhere 2.0: IU SCA continues to lower barriers to IU supercomputing

The Scalable Compute Archive (IU SCA) team strives to make IU’s world class cyberinfrastructure accessible to a wide audience of researchers and scientists. The HPC everywhere web portal, unveiled in fall 2018, has seen a steady increase in its user base and available functionality. This fall, as part of efforts to move the web application from beta to production, the SCA team unveiled HPC everywhere 2.0, featuring several significant updates to its many tools and uses. The updated web application enables an improved user experience and significantly faster load times.

After evaluating the Angular, React, and Vue.js frameworks, the team rewrote the web user interface in the modern Vue.js framework this summer. The original web user interface (UI) which was developed in the AngularJS framework, will only be supported through July 2021.  “Vue.js offered the best combination for our team in terms of developmental learning curve, intuitive design, time to production, ability to create reusable components, and tighter integration with UI libraries like Bootstrap and FontAwesome,” said Michael D. Young, lead developer and architect. Young presented HPC everywhere 2.0 at the 2019 SuperComputing conference in Denver, offering a demonstration of the site and its many capabilities.

While some existing tools (Pathfinder, HPC Applications, MyHPC) remain unchanged in functionality, new and updated features include the following:

1. The Wait-time Visualizer has been integrated into the Job Script Generator tool, which is also now accessible directly within the MyHPC tool.  
2. A new Script Doctor tool assists users in evaluating job scripts for common error conditions. 
3. The new Job Script Generator and Script Doctor tools support the Slurm job manager, enabling a smoother transition for users. As IU supercomputers migrate away from the PBS job manager, a new tool allowing users to convert their PBS jobs scripts to the Slurm syntax has also been made available. 
4. The new Big Red 3 and Carbonate Deep Learning clusters have been integrated into various tools.
5. The user’s SDA (tape archive), Slate (High Performance File System), and Geode (home directory) storage usage are displayed on the home page, with improvements to the respective backend data source structures and near real-time updates.

The 2.0 update also includes administrative tools that enable Research Technologies staff to offer improved support to users, and the SCA team to monitor the web application’s overall health.

“We are thrilled with the improvements made to HPC everywhere 2.0; big props to Mike Young and my sincere thanks to RT colleagues including HPFS, Research Storage, and HPS for their contributions,” said SCA manager Arvind Gopu.