Using Computing Containers and Continuous Integration to Improve Numerical Research Reproducibility

Cloud computing has opened new options of collaboration between research teams in the field of high performance computing and numerical research. Running computational workloads in virtual machines became common in recent years. However, the use of computing containers provides many additional advantages besides just proving new possible runtime choice. One of the most important (and often underappreciated) is an option to improve the reproducibility of research results based on complex mathematical modeling. This paper provides an overview of architecture based on computing containers and continuous integration tools we used to achieve reproducible numerical result