Fast solution of linear systems with many right hand sides using MPI parallel skeletonization

Research output: Chapter in Book/Report/Conference proceedingConference contributionpeer-review

Abstract

It has been proved that the interpolative decomposition (ID) skeletonization can greatly accelerate the simulation where many incidents have to be repeatedly solved. However, the peak memory usage for the skeletonization may be a bottleneck. In this work, an MPI parallelization of the skeletonization is developed to alleviate the difficulty. Numerical experiments have been conducted to demonstrate the performance of the proposed parallel scheme.

Original languageEnglish
Title of host publication2017 IEEE Antennas and Propagation Society International Symposium, Proceedings
PublisherInstitute of Electrical and Electronics Engineers Inc.
Pages731-732
Number of pages2
ISBN (Electronic)9781538632840
DOIs
Publication statusPublished - 18 Oct 2017
Event2017 IEEE International Symposium on Antennas and Propagation and USNC-URSI Radio Science Meeting, APSURSI 2017 - San Diego, United States
Duration: 9 Jul 201714 Jul 2017

Publication series

Name2017 IEEE Antennas and Propagation Society International Symposium, Proceedings
Volume2017-January

Conference

Conference2017 IEEE International Symposium on Antennas and Propagation and USNC-URSI Radio Science Meeting, APSURSI 2017
Country/TerritoryUnited States
CitySan Diego
Period9/07/1714/07/17

Fingerprint

Dive into the research topics of 'Fast solution of linear systems with many right hand sides using MPI parallel skeletonization'. Together they form a unique fingerprint.

Cite this