On improving the performance of sparse matrix-vector multiplication

James B. White; P. Sadayappan

On improving the performance of sparse matrix-vector multiplication

Research output: Contribution to conference › Paper › peer-review

47 Scopus citations

Abstract

We analyze single-node performance of sparse matrix-vector multiplication by investigating issues of data locality and fine-grained parallelism. We examine the data-locality characteristics of the compressed-sparse-row representation and consider improvements in locality through matrix permutation. Motivated by potential improvements in fine-grained parallelism, we evaluate modified sparse-matrix representations. The results lead to general conclusions about improving single-node performance of sparse matrix-vector multiplication in parallel libraries of sparse iterative solvers.

Original language	English
Pages	66-71
Number of pages	6
State	Published - 1997
Externally published	Yes
Event	Proceedings of the 1997 4th International Conference on High Performance Computing, HiPC - Bangalore, India Duration: Dec 18 1997 → Dec 21 1997

Conference

Conference	Proceedings of the 1997 4th International Conference on High Performance Computing, HiPC
City	Bangalore, India
Period	12/18/97 → 12/21/97

Cite this

@conference{5a0a77409cba45e19cc27d8f2777c94a,

title = "On improving the performance of sparse matrix-vector multiplication",

abstract = "We analyze single-node performance of sparse matrix-vector multiplication by investigating issues of data locality and fine-grained parallelism. We examine the data-locality characteristics of the compressed-sparse-row representation and consider improvements in locality through matrix permutation. Motivated by potential improvements in fine-grained parallelism, we evaluate modified sparse-matrix representations. The results lead to general conclusions about improving single-node performance of sparse matrix-vector multiplication in parallel libraries of sparse iterative solvers.",

author = "White, \{James B.\} and P. Sadayappan",

year = "1997",

language = "English",

pages = "66--71",

note = "Proceedings of the 1997 4th International Conference on High Performance Computing, HiPC ; Conference date: 18-12-1997 Through 21-12-1997",

}

TY - CONF

T1 - On improving the performance of sparse matrix-vector multiplication

AU - White, James B.

AU - Sadayappan, P.

PY - 1997

Y1 - 1997

N2 - We analyze single-node performance of sparse matrix-vector multiplication by investigating issues of data locality and fine-grained parallelism. We examine the data-locality characteristics of the compressed-sparse-row representation and consider improvements in locality through matrix permutation. Motivated by potential improvements in fine-grained parallelism, we evaluate modified sparse-matrix representations. The results lead to general conclusions about improving single-node performance of sparse matrix-vector multiplication in parallel libraries of sparse iterative solvers.

AB - We analyze single-node performance of sparse matrix-vector multiplication by investigating issues of data locality and fine-grained parallelism. We examine the data-locality characteristics of the compressed-sparse-row representation and consider improvements in locality through matrix permutation. Motivated by potential improvements in fine-grained parallelism, we evaluate modified sparse-matrix representations. The results lead to general conclusions about improving single-node performance of sparse matrix-vector multiplication in parallel libraries of sparse iterative solvers.

UR - https://www.scopus.com/pages/publications/0031364322

M3 - Paper

AN - SCOPUS:0031364322

SP - 66

EP - 71

T2 - Proceedings of the 1997 4th International Conference on High Performance Computing, HiPC

Y2 - 18 December 1997 through 21 December 1997

ER -

On improving the performance of sparse matrix-vector multiplication

Abstract

Conference

Other files and links

Fingerprint

Cite this