ICL Research Profile

Overview

The Simulation and Modeling for Understanding Resilience and Faults at Scale (SMURFS) project seeks to acquire the predictive understanding of the complex interactions of a given application, a given real or hypothetical hardware and software environment, and a given fault-tolerance strategy at extreme scale.

SMURFS is characterized by two facets: (1) medium and fine-grained predictive capabilities and (2) coarse-grained fault tolerance strategy selection. Accordingly, ICL plans to design, develop, and validate new analytical and system component models that use semi-detailed software and hardware specifications to predict application performance in terms of time to solution and energy consumption. Also, based on a comprehensive set of studies using several application benchmarks, proxies, full applications, and several different fault tolerance strategies, ICL will gather valuable insights about application behavior at scale.