Multilayered grouping parallel algorithm for multiple‐level multiscale analyses. (21st October 2014)

• Record Type: Journal Article
• Authors: Cho, In Ho; Porter, Keith A.
• Is Part Of: International journal for numerical methods in engineering. Volume 100:Number 12(2014)
• Page Start: 914
• Page End: 932
• Publication Date: 2014-10-21
• Subjects: Numerical analysis -- Periodicals; Engineering mathematics -- Periodicals; 620.001518
• DOI: 10.1002/nme.4791
• Languages: English
• ISSNs: 0029-5981

Abstract

<abstract abstract-type="main" id="nme4791-abs-0001"> <title>SUMMARY</title> <p id="nme4791-para-0001">Multiscale analysis technique became a successful remedy to complicated problems in which nonlinear behavior is linked with microscopic damage mechanisms. For efficient parallel multiscale analyses, hierarchical grouping algorithms (e.g., the two‐level 'coarse‐grained' method) have been suggested and proved superior over a simple parallelization. Here, we expanded the two‐level algorithm to give rise to a multilayered grouping parallel algorithm suitable for large‐scale multiple‐level multiscale analyses. With practical large‐scale applications, we demonstrated the superior performance of multilayered grouping over the coarse‐grained method. Notably, we show that the unique data transfer rates of the symmetric multiprocessor cluster system can lead to the seemingly 'super‐linear speedup' and that there appears to exist the optimal number of subgroups of three‐tiered multiscale analysis. Copyright © 2014 John Wiley &amp; Sons, Ltd.</p> </abstract>