skip to main content
DOE PAGES title logo U.S. Department of Energy
Office of Scientific and Technical Information

Title: A 3D Parallel Algorithm for QR Decomposition

Abstract

Interprocessor communication often dominates the runtime of large matrix computations. Here, we present a parallel algorithm for computing QR decompositions whose bandwidth cost (communication volume) can be decreased at the cost of increasing its latency cost (number of messages). By varying a parameter to navigate the bandwidth/latency tradeoff, we can tune this algorithm for machines with different communication costs.

Authors:
 [1];  [2];  [3];  [4];  [5]
  1. Wake Forest Univ., Winston Salem, NC (United States)
  2. Univ. of California, Berkeley, CA (United States)
  3. INRIA Paris-Rocquencourt, Paris (France)
  4. Lawrence Berkeley National Lab. (LBNL), Berkeley, CA (United States)
  5. New York Univ. (NYU), NY (United States)
Publication Date:
Research Org.:
Lawrence Berkeley National Lab. (LBNL), Berkeley, CA (United States)
Sponsoring Org.:
USDOE Office of Science (SC); National Science Foundation (NSF); European Union (EU)
OSTI Identifier:
1525281
Grant/Contract Number:  
AC02-05CH11231; ACI-1642385; 671633
Resource Type:
Accepted Manuscript
Journal Name:
Proceedings of the 30th on Symposium on Parallelism in Algorithms and Architectures - SPAA '18
Additional Journal Information:
Conference: Proceedings of the 30th on Symposium on Parallelism in Algorithms and Architectures - SPAA '18, Vienna (Austria), 16-18 July 2018
Country of Publication:
United States
Language:
English
Subject:
97 MATHEMATICS AND COMPUTING

Citation Formats

Ballard, Grey, Demmel, James, Grigori, Laura, Jacquelin, Mathias, and Knight, Nicholas. A 3D Parallel Algorithm for QR Decomposition. United States: N. p., 2018. Web. doi:10.1145/3210377.3210415.
Ballard, Grey, Demmel, James, Grigori, Laura, Jacquelin, Mathias, & Knight, Nicholas. A 3D Parallel Algorithm for QR Decomposition. United States. doi:10.1145/3210377.3210415.
Ballard, Grey, Demmel, James, Grigori, Laura, Jacquelin, Mathias, and Knight, Nicholas. Wed . "A 3D Parallel Algorithm for QR Decomposition". United States. doi:10.1145/3210377.3210415. https://www.osti.gov/servlets/purl/1525281.
@article{osti_1525281,
title = {A 3D Parallel Algorithm for QR Decomposition},
author = {Ballard, Grey and Demmel, James and Grigori, Laura and Jacquelin, Mathias and Knight, Nicholas},
abstractNote = {Interprocessor communication often dominates the runtime of large matrix computations. Here, we present a parallel algorithm for computing QR decompositions whose bandwidth cost (communication volume) can be decreased at the cost of increasing its latency cost (number of messages). By varying a parameter to navigate the bandwidth/latency tradeoff, we can tune this algorithm for machines with different communication costs.},
doi = {10.1145/3210377.3210415},
journal = {Proceedings of the 30th on Symposium on Parallelism in Algorithms and Architectures - SPAA '18},
number = ,
volume = ,
place = {United States},
year = {2018},
month = {7}
}

Journal Article:
Free Publicly Available Full Text
Publisher's Version of Record

Save / Share: