Assessing a mini-application as a performance proxy for a finite element method engineering application

摘要

The performance of a large-scale, production-quality science and engineering application (‘app’) is oftenrndominated by a small subset of the code. Even within that subset, computational and data access patterns arernoften repeated, so that an even smaller portion can represent the performance-impacting features. If applicationrndevelopers, parallel computing experts, and computer architects can together identify this representativernsubset and then develop a small mini-application (‘miniapp’) that can capture these primary performancerncharacteristics, then this miniapp can be used to both improve the performance of the app as well as providerna tool for co-design for the high-performance computing community. However, a critical question is whetherrna miniapp can effectively capture key performance behavior of an app. This study provides a comparison ofrnan implicit finite element semiconductor device modeling app on unstructured meshes with an implicit finiternelement miniapp on unstructured meshes. The goal is to assess whether the miniapp is predictive of the performancernof the app. Single compute node performance will be compared, as well as scaling up to 16,000rncores. Results indicate that the miniapp can be reasonably predictive of the performance characteristics ofrnthe app for a single iteration of the solver on a single compute node. Published 2015. This article is a U.S.rnGovernment work and is in the public domain in the USA.