|About this Abstract
||2016 TMS Annual Meeting & Exhibition
||ICME Infrastructure Development for Accelerated Materials Design: Data Repositories, Informatics, and Computational Tools
||Automated Convergence Checks with the Python Based Workbench PyIron
||Jan Janssen, Tilmann Hickel, Joerg Neugebauer
|On-Site Speaker (Planned)
Tremendous methodological and computational progress in atomistic simulations has substantially improved the predictive power in materials design. At the same time, the increase in complexity of such approaches often requires to combine several codes, to perform extensive and property dependent convergence checks and to handle large datasets. This makes the automated generation of input for ICME approaches challenging. We have therefore developed a Python based workbench that seamlessly connects a SQL database with hdf5-data storage and a classical file system. It allows the user to quickly find and reuse existing algorithms, gives full flexibility and simplifies the generation of automated tasks. Since all input and output is generic, a seamless transfer between DFT, MD and KMC and a reusability of scripts is achieved. To demonstrate key concepts of the approach, working examples of how to quickly setup and run automated calculations with a prescribed error convergence will be discussed.
||Planned: A print-only volume