Impact of the implementation of MPI point-to-point communications on the performance of two general sparse solvers

被引：8

作者：

Amestoy, PR

Duff, IS

L'Excellent, JY

Li, XYS

机构：

[1] Ecole Normale Super Lyon, INRIA Rhone Alpes, CNRS, URM 5668, F-69364 Lyon 07, France

[2] Ecole Normale Super Lyon, CNRS, Lab Informat Parallelisme, UMR 5668, F-69364 Lyon, France

[3] IRIT, ENSEEIHT, F-31071 Toulouse, France

[4] Lawrence Berkeley Natl Lab, NERSC, Berkeley, CA 94720 USA

[5] Rutherford Appleton Lab, Didcot OX11 0QX, Oxon, England

来源：

PARALLEL COMPUTING | 2003年 / 29卷 / 07期

关键词：

MPI; message passing; immediate communication primitives; direct solvers; sparse matrices;

D O I：

10.1016/S0167-8191(03)00067-X

中图分类号：

TP301 [理论、方法];

学科分类号：

081202 ;

摘要：

We examine the send and receive mechanisms of MPI and show how to implement message passing robustly so that performance is not significantly affected by changes to the MPI system. We discuss this within the context of two different parallel algorithms for sparse Gaussian elimination: a multifrontal solver (MUMPS), and a supernodal one (SuperLU). The performance of our initial strategies based on simple MPI point-to-point communication primitives is very sensitive to the MPI system, particularly the way MPI buffers are used. Using nonblocking communication primitives improves the performance and robustness, but at the cost of increased code complexity. (C) 2003 Elsevier Science B.V. All rights reserved.

引用

页码：833 / 849

页数：17

共 50 条

[1] Open MPI: A High Performance, Flexible Implementation of MPI Point-to-Point Communications
Graham, Richard L.
Barrett, Brain W.
Shipman, Galen M.
Woodall, Timothy S.
Bosilca, George
[J]. PARALLEL PROCESSING LETTERS, 2007, 17 (01) : 79 - 88
[2] On the Effects of CPU Caches on MPI Point-to-Point Communications
Pellegrini, Simone
Hoefler, Torsten
Fahringer, Thomas
[J]. 2012 IEEE INTERNATIONAL CONFERENCE ON CLUSTER COMPUTING (CLUSTER), 2012, : 495 - 503
[3] Regularizing Irregularly Sparse Point-to-point Communications
Selvitopi, Oguz
Aykanat, Cevdet
[J]. PROCEEDINGS OF SC19: THE INTERNATIONAL CONFERENCE FOR HIGH PERFORMANCE COMPUTING, NETWORKING, STORAGE AND ANALYSIS, 2019,
[4] Modeling and predicting point-to-point communications of MPI parallel programs in NOW environments
de Oliveira, HM
Laine, JM
Midorikawa, ET
Sato, LM
Ching, LK
Gaudiot, JL
[J]. PDPTA'03: PROCEEDINGS OF THE INTERNATIONAL CONFERENCE ON PARALLEL AND DISTRIBUTED PROCESSING TECHNIQUES AND APPLICATIONS, VOLS 1-4, 2003, : 1656 - 1662
[5] Open MPI's TEG point-to-point communications methodology: Comparison to existing implementations
Woodall, TS
Graham, RL
Castain, RH
Daniel, DJ
Sukalski, MW
Fagg, GE
Gabriel, E
Bosilca, G
Angskun, T
Dongarra, JJ
Squyres, JM
Sahay, V
Kambadur, R
Barrett, B
Lumsdaine, A
[J]. RECENT ADVANCES IN PARALLEL VIRTUAL MACHINE AND MESSAGE PASSING INTERFACE, PROCEEDINGS, 2004, 3241 : 105 - 111
[6] Debugging point-to-point communication in MPI and PVM.
Kranzlmüller, D
Volkert, J
[J]. RECENT ADVANCES IN PARALLEL VIRTUAL MACHINE AND MESSAGE PASSING INTERFACE, 1998, 1497 : 265 - 272
[7] Impact and Limitations of Point-to-point Performance on Collective Algorithms
Jha, Shweta
Gabriel, Edgar
[J]. 2016 16TH IEEE/ACM INTERNATIONAL SYMPOSIUM ON CLUSTER, CLOUD AND GRID COMPUTING (CCGRID), 2016, : 261 - 266
[8] Delay performance of cognitive radio networks for point-to-point and point-to-multipoint communications
Hung Tran
Trung Q Duong
Hans-Jürgen Zepernick
[J]. EURASIP Journal on Wireless Communications and Networking, 2012
[9] Delay performance of cognitive radio networks for point-to-point and point-to-multipoint communications
Tran, Hung
Duong, Trung Q.
Zepernick, Hans-Jurgen
[J]. EURASIP JOURNAL ON WIRELESS COMMUNICATIONS AND NETWORKING, 2012,
[10] Characterization of Wideband Antennas for Point-to-Point Communications
Almoteriy, Mohammed A.
Sobhy, Mohamed I.
Batchelor, John C.
[J]. IEEE TRANSACTIONS ON ANTENNAS AND PROPAGATION, 2018, 66 (09) : 4466 - 4473

← 1 2 3 4 5 →