An improved reduction algorithm with deeply pipelined operators

Abstract

Many scientific applications involve reduction or accumulation operations on sequential data streams. Examples such as matrix-vector multiplication include multiple inner product operations on different data sets. If the core operator of the reduction is deeply pipelined, which is usually the case, dependencies between the input data cause data hazards in… (More)
DOI: 10.1109/ICSMC.2009.5345939

5 Figures and Tables

Topics

  • Presentations referencing similar topics