PPAM'07 - SEVENTH INTERNATIONAL CONFERENCE ON Parallel Processing and Applied Mathematics

Sparse Matrix-Vector Multiplication - Final Solution?
Authors: Ivan Šimeček and Pavel Tvrdík
Keywords

Diagonal register blocking, variable-size blocking format, sparse matrix-vector multiplication, dynamic programming.


Abstract

Algorithms for the sparse matrix-vector multiplication (shortly spMV) are important building blocks in solvers of sparse systems of linear equations. Due to matrix sparsity, the memory access patterns are irregular and the utilization of a cache suffers from low spatial and temporal locality. To reduce this effect, the register blocking formats were designed. This paper introduces a new combined format, for storing sparse matrices that extends possibilities of the variable-sized register blocking format.



BibTex entry:
@inproceedings{JA_PPAM07,
author = {{\v S}ime{\v c}ek, I. and Tvrd{\'\i}k, P.},
title = {Sparse Matrix-Vector Multiplication - Final Solution?},
booktitle = {Parallel Processing and Applied Mathematics},
publisher = {Springer},
address = {Berlin},
year = {2008},
volume = {4967},
pages = {156--165},
ISBN = {978-3-540-68105-2},
ISSN = {0302-9743},
language = {English},
doi = {10.1007/978-3-540-68111-3_17},
URL = {http://www.springerlink.com/content/48x1345471067304/}
}


BACK