(683e) Saffire: Enabling Large Scale Simulations of Rare Events

We have developed software platform SAFFIRE to perform large scale forward flux sampling (FFS) simulations. FFS is an established scientific method for sampling rare events in simulations. However, as the difficulty of the scientific problem increases, the amount of data and the number of tasks required for FFS is challenging to manage with traditional scripting tools and languages for high performance computing. SAFFIRE software framework has been developed to address these challenges. SAFFIRE utilizes Hadoop to manage a large number of tasks and data for large-scale FFS simulations. The framework is shown to be highly scalable and is able to the support large-scale FFS simulations necessary to extend the use of FFS to complex molecular systems on commodity cluster computing systems. In the talk we will discuss the journey of developing the software highlighting effective and ineffective strategies, as well as discuss plans for sustaining the software and training students and other groups to use this platform.