English

A Study on the Influence of Caching: Sequences of Dense Linear Algebra Kernels

Mathematical Software 2014-02-25 v1 Numerical Analysis Performance

Abstract

It is universally known that caching is critical to attain high- performance implementations: In many situations, data locality (in space and time) plays a bigger role than optimizing the (number of) arithmetic floating point operations. In this paper, we show evidence that at least for linear algebra algorithms, caching is also a crucial factor for accurate performance modeling and performance prediction.

Keywords

Cite

@article{arxiv.1402.5897,
  title  = {A Study on the Influence of Caching: Sequences of Dense Linear Algebra Kernels},
  author = {Elmar Peise and Paolo Bientinesi},
  journal= {arXiv preprint arXiv:1402.5897},
  year   = {2014}
}

Comments

Submitted to the Ninth International Workshop on Automatic Performance Tuning (iWAPT2014)