Run-time techniques for parallelizing sparse matrix problems

M. Ujaldon; S. D. Sharma; J. Saltz; E. L. Zapata

doi:10.1007/3-540-60321-2_3

Run-time techniques for parallelizing sparse matrix problems

M. Ujaldon
, S. D. Sharma
, J. Saltz
, E. L. Zapata

Biomedical Informatics

University of Málaga
University of Maryland, College Park

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution › peer-review

7 Scopus citations

Abstract

Sparse matrix problems are difficult to parallelize efficiently on message-passing machines, since they access data through multiple levels of indirection. Inspector/executor strategies, which are typically used to parallelize such problems impose significant preprocessing overheads. This paper describes the runtime support required by new compilation techniques for sparse matrices and evaluates their performance, highlighting optimizations and improvements over previous techniques.

Original language	English
Title of host publication	Parallel Algorithms for Irregularly Structured Problems - 2nd International Workshop, IRREGULAR 1995, Proceedings
Editors	José Rolim, Afonso Ferreira
Publisher	Springer Verlag
Pages	43-57
Number of pages	15
ISBN (Print)	3540603212, 9783540603214
DOIs	https://doi.org/10.1007/3-540-60321-2_3
State	Published - 1995
Event	2nd International Workshop on Parallel Algorithms for Irregularly Structured Problems, IRREGULAR 1995 - Lyon, France Duration: Sep 4 1995 → Sep 6 1995

Publication series

Name	Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Volume	980
ISSN (Print)	0302-9743
ISSN (Electronic)	1611-3349

Conference

Conference	2nd International Workshop on Parallel Algorithms for Irregularly Structured Problems, IRREGULAR 1995
Country/Territory	France
City	Lyon
Period	09/4/95 → 09/6/95

Access to Document

10.1007/3-540-60321-2_3

Cite this

Ujaldon, M., Sharma, S. D., Saltz, J., & Zapata, E. L. (1995). Run-time techniques for parallelizing sparse matrix problems. In J. Rolim, & A. Ferreira (Eds.), Parallel Algorithms for Irregularly Structured Problems - 2nd International Workshop, IRREGULAR 1995, Proceedings (pp. 43-57). (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 980). Springer Verlag. https://doi.org/10.1007/3-540-60321-2_3

Ujaldon, M. ; Sharma, S. D. ; Saltz, J. et al. / Run-time techniques for parallelizing sparse matrix problems. Parallel Algorithms for Irregularly Structured Problems - 2nd International Workshop, IRREGULAR 1995, Proceedings. editor / José Rolim ; Afonso Ferreira. Springer Verlag, 1995. pp. 43-57 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)).

@inproceedings{dd1e145d2ab041c59630efc0efa7562d,

title = "Run-time techniques for parallelizing sparse matrix problems",

abstract = "Sparse matrix problems are difficult to parallelize efficiently on message-passing machines, since they access data through multiple levels of indirection. Inspector/executor strategies, which are typically used to parallelize such problems impose significant preprocessing overheads. This paper describes the runtime support required by new compilation techniques for sparse matrices and evaluates their performance, highlighting optimizations and improvements over previous techniques.",

author = "M. Ujaldon and Sharma, \{S. D.\} and J. Saltz and Zapata, \{E. L.\}",

note = "Publisher Copyright: {\textcopyright} Springer-Verlag Berlin Heidelberg 1995.; 2nd International Workshop on Parallel Algorithms for Irregularly Structured Problems, IRREGULAR 1995 ; Conference date: 04-09-1995 Through 06-09-1995",

year = "1995",

doi = "10.1007/3-540-60321-2\_3",

language = "English",

isbn = "3540603212",

series = "Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)",

publisher = "Springer Verlag",

pages = "43--57",

editor = "Jos{\'e} Rolim and Afonso Ferreira",

booktitle = "Parallel Algorithms for Irregularly Structured Problems - 2nd International Workshop, IRREGULAR 1995, Proceedings",

}

Ujaldon, M, Sharma, SD, Saltz, J & Zapata, EL 1995, Run-time techniques for parallelizing sparse matrix problems. in J Rolim & A Ferreira (eds), Parallel Algorithms for Irregularly Structured Problems - 2nd International Workshop, IRREGULAR 1995, Proceedings. Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics), vol. 980, Springer Verlag, pp. 43-57, 2nd International Workshop on Parallel Algorithms for Irregularly Structured Problems, IRREGULAR 1995, Lyon, France, 09/4/95. https://doi.org/10.1007/3-540-60321-2_3

Run-time techniques for parallelizing sparse matrix problems. / Ujaldon, M.; Sharma, S. D.; Saltz, J. et al.
Parallel Algorithms for Irregularly Structured Problems - 2nd International Workshop, IRREGULAR 1995, Proceedings. ed. / José Rolim; Afonso Ferreira. Springer Verlag, 1995. p. 43-57 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 980).

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution › peer-review

TY - GEN

T1 - Run-time techniques for parallelizing sparse matrix problems

AU - Ujaldon, M.

AU - Sharma, S. D.

AU - Saltz, J.

AU - Zapata, E. L.

N1 - Publisher Copyright: © Springer-Verlag Berlin Heidelberg 1995.

PY - 1995

Y1 - 1995

N2 - Sparse matrix problems are difficult to parallelize efficiently on message-passing machines, since they access data through multiple levels of indirection. Inspector/executor strategies, which are typically used to parallelize such problems impose significant preprocessing overheads. This paper describes the runtime support required by new compilation techniques for sparse matrices and evaluates their performance, highlighting optimizations and improvements over previous techniques.

AB - Sparse matrix problems are difficult to parallelize efficiently on message-passing machines, since they access data through multiple levels of indirection. Inspector/executor strategies, which are typically used to parallelize such problems impose significant preprocessing overheads. This paper describes the runtime support required by new compilation techniques for sparse matrices and evaluates their performance, highlighting optimizations and improvements over previous techniques.

UR - https://www.scopus.com/pages/publications/84957616696

U2 - 10.1007/3-540-60321-2_3

DO - 10.1007/3-540-60321-2_3

M3 - Conference contribution

AN - SCOPUS:84957616696

SN - 3540603212

SN - 9783540603214

T3 - Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)

SP - 43

EP - 57

BT - Parallel Algorithms for Irregularly Structured Problems - 2nd International Workshop, IRREGULAR 1995, Proceedings

A2 - Rolim, José

A2 - Ferreira, Afonso

PB - Springer Verlag

T2 - 2nd International Workshop on Parallel Algorithms for Irregularly Structured Problems, IRREGULAR 1995

Y2 - 4 September 1995 through 6 September 1995

ER -

Ujaldon M, Sharma SD, Saltz J, Zapata EL. Run-time techniques for parallelizing sparse matrix problems. In Rolim J, Ferreira A, editors, Parallel Algorithms for Irregularly Structured Problems - 2nd International Workshop, IRREGULAR 1995, Proceedings. Springer Verlag. 1995. p. 43-57. (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)). doi: 10.1007/3-540-60321-2_3

Run-time techniques for parallelizing sparse matrix problems

Abstract

Publication series

Conference

Access to Document

Other files and links

Fingerprint

Cite this