One-step algorithm for mixed data and task parallel scheduling without data replication

Vincent Boudet; Frédéric Desprez; Frédéric Suter

doi:10.1109/IPDPS.2003.1213127

One-step algorithm for mixed data and task parallel scheduling without data replication

Vincent Boudet, Frédéric Desprez, Frédéric Suter

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

20 Scopus citations

Abstract

In this paper we propose an original algorithm for mixed data and task parallel scheduling. The main specificities of this algorithm are to simultaneously perform the allocation and scheduling processes, and avoid data replication. The idea is to base the scheduling on an accurate evaluation of each task of the application depending on the processor grid. Then no assumption is made with regard to the homogeneity of the execution platform. The complexity of our algorithm is given. Performance achieved by our schedules both in homogeneous and heterogeneous worlds, are compared to data-parallel executions for two applications: the complex matrix multiplication and the Strassen decomposition.

Original language	English
Title of host publication	Proceedings - International Parallel and Distributed Processing Symposium, IPDPS 2003
Publisher	Institute of Electrical and Electronics Engineers Inc.
ISBN (Electronic)	0769519261, 9780769519265
DOIs	https://doi.org/10.1109/IPDPS.2003.1213127
State	Published - 2003
Externally published	Yes
Event	International Parallel and Distributed Processing Symposium, IPDPS 2003 - Nice, France Duration: Apr 22 2003 → Apr 26 2003

Publication series

Name	Proceedings - International Parallel and Distributed Processing Symposium, IPDPS 2003

Conference

Conference	International Parallel and Distributed Processing Symposium, IPDPS 2003
Country/Territory	France
City	Nice
Period	04/22/03 → 04/26/03

Keywords

Mixed-parallelism
Ressource allocation
Scheduling

Access to Document

10.1109/IPDPS.2003.1213127

Cite this

Boudet, V., Desprez, F., & Suter, F. (2003). One-step algorithm for mixed data and task parallel scheduling without data replication. In Proceedings - International Parallel and Distributed Processing Symposium, IPDPS 2003 Article 1213127 (Proceedings - International Parallel and Distributed Processing Symposium, IPDPS 2003). Institute of Electrical and Electronics Engineers Inc.. https://doi.org/10.1109/IPDPS.2003.1213127

Boudet, Vincent ; Desprez, Frédéric ; Suter, Frédéric. / One-step algorithm for mixed data and task parallel scheduling without data replication. Proceedings - International Parallel and Distributed Processing Symposium, IPDPS 2003. Institute of Electrical and Electronics Engineers Inc., 2003. (Proceedings - International Parallel and Distributed Processing Symposium, IPDPS 2003).

@inproceedings{757edb52b231433189a7ff858307258f,

title = "One-step algorithm for mixed data and task parallel scheduling without data replication",

abstract = "In this paper we propose an original algorithm for mixed data and task parallel scheduling. The main specificities of this algorithm are to simultaneously perform the allocation and scheduling processes, and avoid data replication. The idea is to base the scheduling on an accurate evaluation of each task of the application depending on the processor grid. Then no assumption is made with regard to the homogeneity of the execution platform. The complexity of our algorithm is given. Performance achieved by our schedules both in homogeneous and heterogeneous worlds, are compared to data-parallel executions for two applications: the complex matrix multiplication and the Strassen decomposition.",

keywords = "Mixed-parallelism, Ressource allocation, Scheduling",

author = "Vincent Boudet and Fr{\'e}d{\'e}ric Desprez and Fr{\'e}d{\'e}ric Suter",

note = "Publisher Copyright: {\textcopyright} 2003 IEEE.; International Parallel and Distributed Processing Symposium, IPDPS 2003 ; Conference date: 22-04-2003 Through 26-04-2003",

year = "2003",

doi = "10.1109/IPDPS.2003.1213127",

language = "English",

series = "Proceedings - International Parallel and Distributed Processing Symposium, IPDPS 2003",

publisher = "Institute of Electrical and Electronics Engineers Inc.",

booktitle = "Proceedings - International Parallel and Distributed Processing Symposium, IPDPS 2003",

}

Boudet, V, Desprez, F & Suter, F 2003, One-step algorithm for mixed data and task parallel scheduling without data replication. in Proceedings - International Parallel and Distributed Processing Symposium, IPDPS 2003., 1213127, Proceedings - International Parallel and Distributed Processing Symposium, IPDPS 2003, Institute of Electrical and Electronics Engineers Inc., International Parallel and Distributed Processing Symposium, IPDPS 2003, Nice, France, 04/22/03. https://doi.org/10.1109/IPDPS.2003.1213127

One-step algorithm for mixed data and task parallel scheduling without data replication. / Boudet, Vincent; Desprez, Frédéric; Suter, Frédéric.
Proceedings - International Parallel and Distributed Processing Symposium, IPDPS 2003. Institute of Electrical and Electronics Engineers Inc., 2003. 1213127 (Proceedings - International Parallel and Distributed Processing Symposium, IPDPS 2003).

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

TY - GEN

T1 - One-step algorithm for mixed data and task parallel scheduling without data replication

AU - Boudet, Vincent

AU - Desprez, Frédéric

AU - Suter, Frédéric

PY - 2003

Y1 - 2003

N2 - In this paper we propose an original algorithm for mixed data and task parallel scheduling. The main specificities of this algorithm are to simultaneously perform the allocation and scheduling processes, and avoid data replication. The idea is to base the scheduling on an accurate evaluation of each task of the application depending on the processor grid. Then no assumption is made with regard to the homogeneity of the execution platform. The complexity of our algorithm is given. Performance achieved by our schedules both in homogeneous and heterogeneous worlds, are compared to data-parallel executions for two applications: the complex matrix multiplication and the Strassen decomposition.

AB - In this paper we propose an original algorithm for mixed data and task parallel scheduling. The main specificities of this algorithm are to simultaneously perform the allocation and scheduling processes, and avoid data replication. The idea is to base the scheduling on an accurate evaluation of each task of the application depending on the processor grid. Then no assumption is made with regard to the homogeneity of the execution platform. The complexity of our algorithm is given. Performance achieved by our schedules both in homogeneous and heterogeneous worlds, are compared to data-parallel executions for two applications: the complex matrix multiplication and the Strassen decomposition.

KW - Mixed-parallelism

KW - Ressource allocation

KW - Scheduling

UR - http://www.scopus.com/inward/record.url?scp=84947212025&partnerID=8YFLogxK

U2 - 10.1109/IPDPS.2003.1213127

DO - 10.1109/IPDPS.2003.1213127

M3 - Conference contribution

AN - SCOPUS:84947212025

T3 - Proceedings - International Parallel and Distributed Processing Symposium, IPDPS 2003

BT - Proceedings - International Parallel and Distributed Processing Symposium, IPDPS 2003

PB - Institute of Electrical and Electronics Engineers Inc.

T2 - International Parallel and Distributed Processing Symposium, IPDPS 2003

Y2 - 22 April 2003 through 26 April 2003

ER -

Boudet V, Desprez F, Suter F. One-step algorithm for mixed data and task parallel scheduling without data replication. In Proceedings - International Parallel and Distributed Processing Symposium, IPDPS 2003. Institute of Electrical and Electronics Engineers Inc. 2003. 1213127. (Proceedings - International Parallel and Distributed Processing Symposium, IPDPS 2003). doi: 10.1109/IPDPS.2003.1213127

One-step algorithm for mixed data and task parallel scheduling without data replication

Abstract

Publication series

Conference

Keywords

Access to Document

Other files and links

Fingerprint

Cite this