Abstract
We revisit the problem of evaluating matrix polynomials and introduce memory and communication efficient algorithms. Our algorithms, based on that of Patterson and Stockmeyer, are more efficient than previous ones, while being as memory-efficient as Van Loan’s variant. We supplement our theoretical analysis of the algorithms, with matching lower bounds and with experimental results showing that our algorithms outperform existing ones.
Original language | American English |
---|---|
Title of host publication | Parallel Processing and Applied Mathematics - 12th International Conference, PPAM 2017, Revised Selected Papers |
Editors | Jack Dongarra, Roman Wyrzykowski, Konrad Karczewski, Ewa Deelman |
Publisher | Springer Verlag |
Pages | 24-35 |
Number of pages | 12 |
ISBN (Print) | 9783319780238 |
DOIs | |
State | Published - 2018 |
Event | 12th International Conference on Parallel Processing and Applied Mathematics, PPAM 2017 - Czestochowa, Poland Duration: 10 Sep 2017 → 13 Sep 2017 |
Publication series
Name | Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics) |
---|---|
Volume | 10777 LNCS |
ISSN (Print) | 0302-9743 |
ISSN (Electronic) | 1611-3349 |
Conference
Conference | 12th International Conference on Parallel Processing and Applied Mathematics, PPAM 2017 |
---|---|
Country/Territory | Poland |
City | Czestochowa |
Period | 10/09/17 → 13/09/17 |
Bibliographical note
Funding Information:This research is supported by grants 1878/14, 1901/14, 965/15 and 863/15 from the Israel Science Foundation, grant 3-10891 from the Israeli Ministry of Science and Technology, by the Einstein and Minerva Foundations, by the PetaCloud consortium, by the Intel Collaborative Research Institute for Computational Intelligence, by a grant from the US-Israel Bi-national Science Foundation, and by the HUJI Cyber Security Research Center.
Publisher Copyright:
© Springer International Publishing AG, part of Springer Nature 2018.
Keywords
- Cache efficiency
- Matrix polynomials
- Polynomial evaluation