This article is devoted to the implementation of asynchronous parallel iterative algorithms. We will first discuss the implementation of stopping tests both from a computational approach and from a numerical analysis approach using in the latter case either the contraction property or the partial ordering convergence one and also the nested sets. After recalling some notions concerning the architecture of multiprocessor machines, the implementation principle of these asynchronous parallel iterative methods will be discussed, in particular for the sub-domain methods ; load balancing for these algorithms will be also discussed.
Read this article from a comprehensive knowledge base, updated and supplemented with articles reviewed by scientific committees.
This offer includes interactive articles. Their quizzes highlight the key information to remember and validate their acquisition: from reader to player, you can enrich your knowledge.
You can easily identify them thanks to this pictogram: