A Block SPP Algorithm for Multidimensional Tridiagonal Equations with Optimal Message Vector Length

A Block SPP Algorithm for Multidimensional Tridiagonal Equations with Optimal Message Vector Length

A parallel strategy for solving multidimensional tridiagonal equations is investigated in this paper. We present in detail an improved version of single parallel partition (SPP) algorithm in conjunction with message vectorization, which aggregates several communication messages into one to reduce th...

Full description

Bibliographic Details
Main Authors: Hong Guo, Zhao-Hua Yin, Li Yuan
Format: Article
Language:English
Published: SAGE Publishing 2009-06-01
Series:Journal of Algorithms & Computational Technology
Online Access:https://doi.org/10.1260/174830109787913967
id doaj-a2f63dae237b4c9e9c15b944afd876b5
record_format Article
spelling doaj-a2f63dae237b4c9e9c15b944afd876b52020-11-25T03:06:42ZengSAGE PublishingJournal of Algorithms & Computational Technology1748-30181748-30262009-06-01310.1260/174830109787913967A Block SPP Algorithm for Multidimensional Tridiagonal Equations with Optimal Message Vector LengthHong Guo0Zhao-Hua Yin1Li Yuan2 LSEC and Institute of Computational Mathematics and Scientific/Engineering Computing, Academy of Mathematics & Systems Science, Chinese Academy of Sciences, Beijing 100190, P.R. China National Microgravity Laboratory, Institute of Mechanics, Chinese Academy of Sciences, Beijing 100190, P.R. China LSEC and Institute of Computational Mathematics and Scientific/Engineering Computing, Academy of Mathematics & Systems Science, Chinese Academy of Sciences, Beijing 100190, P.R. ChinaA parallel strategy for solving multidimensional tridiagonal equations is investigated in this paper. We present in detail an improved version of single parallel partition (SPP) algorithm in conjunction with message vectorization, which aggregates several communication messages into one to reduce the communication cost. We show the resulting block SPP can achieve good speedup for a wide range of message vector length (MVL), especially when the number of grid points in the divided direction is large. Instead of only using the largest possible MVL, we adopt numerical tests and modeling analysis to determine an optimal MVL so that significant improvement in speedup can be obtained.https://doi.org/10.1260/174830109787913967
collection DOAJ
language English
format Article
sources DOAJ
author Hong Guo
Zhao-Hua Yin
Li Yuan
spellingShingle Hong Guo
Zhao-Hua Yin
Li Yuan
A Block SPP Algorithm for Multidimensional Tridiagonal Equations with Optimal Message Vector Length
Journal of Algorithms & Computational Technology
author_facet Hong Guo
Zhao-Hua Yin
Li Yuan
author_sort Hong Guo
title A Block SPP Algorithm for Multidimensional Tridiagonal Equations with Optimal Message Vector Length
title_short A Block SPP Algorithm for Multidimensional Tridiagonal Equations with Optimal Message Vector Length
title_full A Block SPP Algorithm for Multidimensional Tridiagonal Equations with Optimal Message Vector Length
title_fullStr A Block SPP Algorithm for Multidimensional Tridiagonal Equations with Optimal Message Vector Length
title_full_unstemmed A Block SPP Algorithm for Multidimensional Tridiagonal Equations with Optimal Message Vector Length
title_sort block spp algorithm for multidimensional tridiagonal equations with optimal message vector length
publisher SAGE Publishing
series Journal of Algorithms & Computational Technology
issn 1748-3018
1748-3026
publishDate 2009-06-01
description A parallel strategy for solving multidimensional tridiagonal equations is investigated in this paper. We present in detail an improved version of single parallel partition (SPP) algorithm in conjunction with message vectorization, which aggregates several communication messages into one to reduce the communication cost. We show the resulting block SPP can achieve good speedup for a wide range of message vector length (MVL), especially when the number of grid points in the divided direction is large. Instead of only using the largest possible MVL, we adopt numerical tests and modeling analysis to determine an optimal MVL so that significant improvement in speedup can be obtained.
url https://doi.org/10.1260/174830109787913967
work_keys_str_mv AT hongguo ablocksppalgorithmformultidimensionaltridiagonalequationswithoptimalmessagevectorlength
AT zhaohuayin ablocksppalgorithmformultidimensionaltridiagonalequationswithoptimalmessagevectorlength
AT liyuan ablocksppalgorithmformultidimensionaltridiagonalequationswithoptimalmessagevectorlength
AT hongguo blocksppalgorithmformultidimensionaltridiagonalequationswithoptimalmessagevectorlength
AT zhaohuayin blocksppalgorithmformultidimensionaltridiagonalequationswithoptimalmessagevectorlength
AT liyuan blocksppalgorithmformultidimensionaltridiagonalequationswithoptimalmessagevectorlength
_version_ 1724673006634008576

Similar Items