An Evaluation of Two Implementation Strategies for Optimizing One-Sided Atomic Reduction
Jarek Nieplocha, V. Tipparaju, Edoardo Aprà · 2005
One-sided atomic reduction, also known as the accumulate operation, combines atomically a content of the local buffer with data at remote memory location. This operation has been included in the MPI-2 standard a, MPI/spl I.bar/Accumulate. The current paper discusses two strategies for implementing one-sided atomic reduction called owner-computes and caller-computes. Performance of these two schemes has been investigated on the HP AlphaServer SC45 and HP zx-2600 clusters both equipped with the Quadrics Elan-3 network.