Scientific workflow is an important technique used in many simulation and data analysis applications. In particular, workflows automate high-throughput / high-complexity computing applications, enable code and data reuse and provenance, provide methods for validation and error tracking, and exploit application concurrency using distributed computing resources. The goal of this tutorial is to learn composing and running workflow applications using the FireWorks workflow environment (https://materialsproject.github.io/fireworks/). After an introduction to the FireWorks, the participants will learn to construct workflows using a library of existing Firetasks. The composed workflows will be verified, visualized and then executed and monitored. The exercises include managing data and control flow dynamically using FWAction. The goal of Exercise 5 is to learn writing custom Firetasks to match more specific application requirements.

Basic knowledge of using the bash shell is required. For Exercise 5 basic knowledge of Python is required.