Visualization for Exascale: Portable Performance is Critical

Authors

  • Kenneth Moreland Sandia National Laboratories, Albuquerque
  • Matthew Larsen University of Oregon, Eugene
  • Hank Childs University of Oregon, Eugene

DOI:

https://doi.org/10.14529/jsfi150306

Abstract

Researchers face a daunting task to provide scientific visualization capabilities for exascale computing. Of the many fundamental changes we are seeing in HPC systems, one of the most profound is a reliance on new processor types optimized for execution bandwidth over latency hiding. Multiple vendors create such accelerator processors, each with significantly different features and performance characteristics. To address these visualization needs across multiple platforms, we