Skip to Main content Skip to Navigation
Conference papers

Non-clairvoyant reduction algorithms for heterogeneous platforms

Abstract : We revisit the classical problem of the reduction collective operation in a heterogeneous environment. We discuss and evaluate four algorithms that are non-clairvoyant, i.e., they do not know in advance the computation and communication costs. On the one hand, \bins and \fibo are static algorithms that decide in advance which operations will be reduced, without adapting to the environment; they were originally defined for homogeneous settings. On the other hand, \dyn and \dynnc are fully dynamic algorithms, for commutative or non-commutative reductions. With identical computation costs, we show that these algorithms are approximation algorithms. When costs are exponentially distributed, we perform an analysis of \dyn based on Markov chains. Finally, we assess the relative performance of all four non-clairvoyant algorithms with heterogeneous costs though a set of simulations.
Complete list of metadatas

Cited literature [15 references]  Display  Hide  Download

https://hal.inria.fr/hal-00926093
Contributor : Equipe Roma <>
Submitted on : Thursday, January 9, 2014 - 10:07:56 AM
Last modification on : Wednesday, September 16, 2020 - 10:42:26 AM
Long-term archiving on: : Thursday, April 10, 2014 - 2:40:37 PM

File

heteropar.pdf
Files produced by the author(s)

Identifiers

  • HAL Id : hal-00926093, version 1

Citation

Anne Benoit, Louis-Claude Canon, Loris Marchal. Non-clairvoyant reduction algorithms for heterogeneous platforms. HeteroPar'2013, in conjunction with Euro-Par 2013, Aug 2013, Aachen, Germany. ⟨hal-00926093⟩

Share

Metrics

Record views

373

Files downloads

449