Evaluating Execution Models Poster March 2017

This poster describes significant outcomes of the Evaluating Execution Models project, including methodologies for empirical analysis and modeling of Execution Model impacts on large-scale workloads.  The Evaluating Execution Models (EEM) project aims to develop a methodology for the quantitative assessment of novel execution models targeting future large-scale system architectures.  Execution models are conceptual and crosscutting, impacting the manner in which algorithms are expressed, as well as defining the operations made available to runtime and application software and supported by underlying hardware.  The EEM project expands upon established Modeling and Simulation (ModSim) capabilities to gain insight into the impact of execution model design and selection in order to guide the co-design process and inform the development of runtime and application software.