Joint work with Johnathan Ebbers, Stefan P. Feer, Vitus J. Leung,
Nickolas W. Price, Zachary D. Rhodes, and Matthew Swank.
Poster presented at
SC 2013.
Abstract:
We examine task mapping algorithms for systems that allocate jobs non-contiguously, such as the Cray X series. Several studies have shown that task placement affects job running time for both contiguously and non-contiguously allocated jobs. Our focus is on jobs with a stencil communication pattern. We use experiments on a Cray XE to evaluate novel task mapping algorithms as well as some adapted to this setting. This is done with the miniGhost miniApp which mimics the performance of CTH, a shock physics application. Our strategies improve average running time by as much as 28% over a baseline strategy.