Challenges and Recommendations for Preparing HPC Applications for Exascale

In this paper we propose an approach for preparing HPC applications for Exascale. A high-level application model capturing nonfunctional properties combined with an abstract machine model will enable application improvements, and will serve as input to runtime systems to handle performance and energy optimizations, and self-aware fault management. A tunable abstract machine model includes current assumptions for future Exascale systems and enables a priori application improvement before the execution platform is specified. At runtime, the model is a posteriori tuned to support iterative code improvement and dynamic optimization. We outline our approach and propose concrete steps for implementing it in form of a conceptual porting framework. Furthermore, we identify challenges that need to be addressed and recommend solutions in key areas of interest for our approach, including formal modeling, static analysis and optimization, runtime analysis and optimization, and autonomic computing.