A methodology for assessing computation/communication overlap of MPI nonblocking collectives - Inria - Institut national de recherche en sciences et technologies du numérique Access content directly
Journal Articles Concurrency and Computation: Practice and Experience Year : 2022

A methodology for assessing computation/communication overlap of MPI nonblocking collectives

Abstract

By allowing computation/communication overlap, MPI nonblocking collectives (NBC) are supposed to improve application scalability and performance. However, it is known that to actually get overlap, the MPI library has to implement progression mechanisms in software or rely on the network hardware. These mechanisms may be present or not, adequate or perfectible, they may have an impact on communication performance or may interfere with computation by stealing CPU cycles. From a user point of view, assessing and understanding the behavior of an MPI library concerning computation/communication overlap is difficult. In this paper, we propose a methodology to assess the computation/communication overlap of NBC. We propose new metrics to measure how much communication and computation do overlap, and to evaluate how they interfere with each other. We integrate these metrics into a complete methodology. We compare our methodology with state of the art metrics and benchmarks, and show that ours provides more meaningful informations. We perform experiments on a large panel of MPI implementations and network hardware and show when and why overlap is efficient, nonexistent or even degrades performance.
Fichier principal
Vignette du fichier
article.pdf (644 Ko) Télécharger le fichier
Origin : Files produced by the author(s)

Dates and versions

hal-03922777 , version 1 (04-01-2023)
hal-03922777 , version 2 (04-01-2023)

Identifiers

Cite

Alexandre Denis, Julien Jaeger, Emmanuel Jeannot, Florian Reynier. A methodology for assessing computation/communication overlap of MPI nonblocking collectives. Concurrency and Computation: Practice and Experience, 2022, 34 (22), ⟨10.1002/cpe.7168⟩. ⟨hal-03922777v2⟩
38 View
27 Download

Altmetric

Share

Gmail Facebook Twitter LinkedIn More