Comprehensive Learning for End-users to Effectively Utilize Cyberinfrastructure (COMPLECS)
Advanced cyberinfrastructure (CI), comprising resources with capabilities beyond those of desktop computing, plays an important role in nearly every scientific and engineering discipline. Most CI users no longer need to write their own software, relying instead on existing robust 3rd party applications. Unfortunately, much of the available training still emphasizes software development topics rather than the skills other than programming that are needed to effectively use advanced CI. COMPLECS addresses this gap by creating a coherent training program spanning foundational knowledge to the practical skills that enable these CI users to make best use of these specialized resources. The training consists of three tiers (multi-day workshops, webinars, and self-paced study), which gives participants a variety of options that best fit into their schedules.
Training topics include parallel computing concepts, intermediate Linux and shell scripting, hardware essentials, batch computing, data management, best practices in securing data and research, interactive computing, getting help and how to work effectively with user support, code migration and installation, high throughput computing and Linux tools for file processing. Except for several foundational topics, the program is designed so that webinar and self-paced study participants can take the training in any order. Materials are available through GitHub repositories, Google drives and captioned recordings of webinars.
Topics covered in this program:
- Parallel computing concepts**
- Intermediate Linux and shell scripting**
- Linux tools for file processing*
- HPC Security and getting help*
- Code migration*
- Getting Started with Batch Job Scheduling: Slurm Edition
- HPC hardware overview*
- Interactive computing*
- Data management: Part 1 & 2
- High-Throughput and Many Task Computing Worksflows: Slurm Edition
(*)topic covered 2x/year
(**)topic covered 3x/year
COMPLECS is supported by U.S. National Science Foundation award #2320934.
