Title	Lossy compression of FE coefficients for reducing communication in time-parallel simulations
Creator	Weiser, Martin
Publisher	Banff International Research Station for Mathematical Innovation and Discovery
Date Issued	2016-12-01T14:41
Description	Communication bandwidth between nodes of parallel machines grows slower than the nodes' computational performance and increasingly becomes a bottleneck of PDE solvers. We present lossy compression techniques based on multilevel prediction in hierarchical grids and entropy coding as a means to reduce the amount of data to be communicated between nodes in SDC-based time-parallel simulations. Compression rate and impact on convergence depending on the quantization error are investigated both theoretically and at some numerical examples realized in the finite element code Kaskade 7, which leads to an adaptive choice of the quantization threshold.
Extent	24 minutes
Subject	Mathematics; Numerical analysis; Partial differential equations
Type	Moving Image
File Format	video/mp4
Language	eng
Notes	Author affiliation: Zuse Institute Berlin
Series	BIRS Workshop Lecture Videos (Banff, Alta)
Date Available	2017-05-15
Provider	Vancouver : University of British Columbia Library
Rights	Attribution-NonCommercial-NoDerivatives 4.0 International
DOI	10.14288/1.0347498
URI	http://hdl.handle.net/2429/61667
Affiliation	Non UBC
Peer Review Status	Unreviewed
Scholarly Level	Other
Rights URI	http://creativecommons.org/licenses/by-nc-nd/4.0/
Aggregated Source Repository	DSpace

Open Collections

BIRS Workshop Lecture Videos