Construction and performance characterization of parallel interior point solver on 4-way intel Itanium 2 multiprocessor system

P. Koka, Taeweon Suh, M. Smelyanskiy, R. Grzeszczuk, C. Dulong

Research output: Chapter in Book/Report/Conference proceedingConference contribution

1 Citation (Scopus)

Abstract

In recent years the interior point method (IPM) has became a dominant choice for solving large convex optimization problems for many scientific, engineering and commercial applications. Two reasons for the success of the IPM are its good scalability on existing multiprocessor systems with a small number of processors and its potential to deliver a scalable performance on systems with a large number of processors. The scalability of a parallel IPM is determined by several key issues such as exploiting parallelism due to sparsity of the problem, reducing communication overhead and proper load balancing. In this paper we present an implementation of a parallel linear programming IPM workload and characterize its scalability on a 4-way Itanium® 2 system. We show a speedup of up to 3-times for some of the datasets. We also present a detailed micro-architectural analysis of the workload using VTune performance analyzer. Our results suggest that a good IPM implementation is latency-bound. Based on these findings, we make suggestions on how to improve the performance of the IPM workload in the future.

Original languageEnglish
Title of host publicationProceedings of the 2004 7th Annual IEEE International Workshop on Workload Characterization, WWC-7
Pages73-80
Number of pages8
Publication statusPublished - 2004 Dec 1
Externally publishedYes
EventProceedings of the 2004 7th Annual IEEE International Workshop on Workload Characterization, WWC-7 - Austin, TX, United States
Duration: 2004 Oct 252004 Oct 25

Other

OtherProceedings of the 2004 7th Annual IEEE International Workshop on Workload Characterization, WWC-7
CountryUnited States
CityAustin, TX
Period04/10/2504/10/25

Fingerprint

Interiors (building)
Scalability
Convex optimization
Linear programming
Resource allocation
Communication

ASJC Scopus subject areas

  • Engineering(all)

Cite this

Koka, P., Suh, T., Smelyanskiy, M., Grzeszczuk, R., & Dulong, C. (2004). Construction and performance characterization of parallel interior point solver on 4-way intel Itanium 2 multiprocessor system. In Proceedings of the 2004 7th Annual IEEE International Workshop on Workload Characterization, WWC-7 (pp. 73-80)

Construction and performance characterization of parallel interior point solver on 4-way intel Itanium 2 multiprocessor system. / Koka, P.; Suh, Taeweon; Smelyanskiy, M.; Grzeszczuk, R.; Dulong, C.

Proceedings of the 2004 7th Annual IEEE International Workshop on Workload Characterization, WWC-7. 2004. p. 73-80.

Research output: Chapter in Book/Report/Conference proceedingConference contribution

Koka, P, Suh, T, Smelyanskiy, M, Grzeszczuk, R & Dulong, C 2004, Construction and performance characterization of parallel interior point solver on 4-way intel Itanium 2 multiprocessor system. in Proceedings of the 2004 7th Annual IEEE International Workshop on Workload Characterization, WWC-7. pp. 73-80, Proceedings of the 2004 7th Annual IEEE International Workshop on Workload Characterization, WWC-7, Austin, TX, United States, 04/10/25.
Koka P, Suh T, Smelyanskiy M, Grzeszczuk R, Dulong C. Construction and performance characterization of parallel interior point solver on 4-way intel Itanium 2 multiprocessor system. In Proceedings of the 2004 7th Annual IEEE International Workshop on Workload Characterization, WWC-7. 2004. p. 73-80
Koka, P. ; Suh, Taeweon ; Smelyanskiy, M. ; Grzeszczuk, R. ; Dulong, C. / Construction and performance characterization of parallel interior point solver on 4-way intel Itanium 2 multiprocessor system. Proceedings of the 2004 7th Annual IEEE International Workshop on Workload Characterization, WWC-7. 2004. pp. 73-80
@inproceedings{b51e23865fc5453a9b05b49c3c82c63f,
title = "Construction and performance characterization of parallel interior point solver on 4-way intel Itanium 2 multiprocessor system",
abstract = "In recent years the interior point method (IPM) has became a dominant choice for solving large convex optimization problems for many scientific, engineering and commercial applications. Two reasons for the success of the IPM are its good scalability on existing multiprocessor systems with a small number of processors and its potential to deliver a scalable performance on systems with a large number of processors. The scalability of a parallel IPM is determined by several key issues such as exploiting parallelism due to sparsity of the problem, reducing communication overhead and proper load balancing. In this paper we present an implementation of a parallel linear programming IPM workload and characterize its scalability on a 4-way Itanium{\circledR} 2 system. We show a speedup of up to 3-times for some of the datasets. We also present a detailed micro-architectural analysis of the workload using VTune ™ performance analyzer. Our results suggest that a good IPM implementation is latency-bound. Based on these findings, we make suggestions on how to improve the performance of the IPM workload in the future.",
author = "P. Koka and Taeweon Suh and M. Smelyanskiy and R. Grzeszczuk and C. Dulong",
year = "2004",
month = "12",
day = "1",
language = "English",
isbn = "0780388283",
pages = "73--80",
booktitle = "Proceedings of the 2004 7th Annual IEEE International Workshop on Workload Characterization, WWC-7",

}

TY - GEN

T1 - Construction and performance characterization of parallel interior point solver on 4-way intel Itanium 2 multiprocessor system

AU - Koka, P.

AU - Suh, Taeweon

AU - Smelyanskiy, M.

AU - Grzeszczuk, R.

AU - Dulong, C.

PY - 2004/12/1

Y1 - 2004/12/1

N2 - In recent years the interior point method (IPM) has became a dominant choice for solving large convex optimization problems for many scientific, engineering and commercial applications. Two reasons for the success of the IPM are its good scalability on existing multiprocessor systems with a small number of processors and its potential to deliver a scalable performance on systems with a large number of processors. The scalability of a parallel IPM is determined by several key issues such as exploiting parallelism due to sparsity of the problem, reducing communication overhead and proper load balancing. In this paper we present an implementation of a parallel linear programming IPM workload and characterize its scalability on a 4-way Itanium® 2 system. We show a speedup of up to 3-times for some of the datasets. We also present a detailed micro-architectural analysis of the workload using VTune ™ performance analyzer. Our results suggest that a good IPM implementation is latency-bound. Based on these findings, we make suggestions on how to improve the performance of the IPM workload in the future.

AB - In recent years the interior point method (IPM) has became a dominant choice for solving large convex optimization problems for many scientific, engineering and commercial applications. Two reasons for the success of the IPM are its good scalability on existing multiprocessor systems with a small number of processors and its potential to deliver a scalable performance on systems with a large number of processors. The scalability of a parallel IPM is determined by several key issues such as exploiting parallelism due to sparsity of the problem, reducing communication overhead and proper load balancing. In this paper we present an implementation of a parallel linear programming IPM workload and characterize its scalability on a 4-way Itanium® 2 system. We show a speedup of up to 3-times for some of the datasets. We also present a detailed micro-architectural analysis of the workload using VTune ™ performance analyzer. Our results suggest that a good IPM implementation is latency-bound. Based on these findings, we make suggestions on how to improve the performance of the IPM workload in the future.

UR - http://www.scopus.com/inward/record.url?scp=19644379079&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=19644379079&partnerID=8YFLogxK

M3 - Conference contribution

AN - SCOPUS:19644379079

SN - 0780388283

SN - 9780780388284

SP - 73

EP - 80

BT - Proceedings of the 2004 7th Annual IEEE International Workshop on Workload Characterization, WWC-7

ER -