Optimizing parallel performance of unstructured volume rendering for the earth simulator

Li Chen, Issei Fujishiro, Kengo Nakajima

Research output: Contribution to journalArticle

18 Citations (Scopus)

Abstract

A scalable and high-performance parallel visualization subsystem has been developed in GeoFEM for the Earth Simulator (ES). As part of the ES project in Japan, the proposed subsystem is effective for the visualization of huge-scale unstructured datasets, and can be concurrent with computation on the same high-performance parallel computer. Moreover, some parallel visualization modules have obtained a good parallel performance, covering scalar, vector and tensor fields. This paper will take volume rendering method as an example to describe a number of efficient parallel performance optimization strategies we adopted for large-scale unstructured data visualization on SMP cluster machines, including suitable design of visualization method, the three-level hybrid parallelization which means message passing for inter-SMP node communication, loop directives for intra-SMP node parallelization, and vectorization for each processing element, plus dynamic load balancing. Good visualization images and high parallel performance have been achieved on the ES, thus demonstrating the feasibility and effectiveness of the proposed method.

Original languageEnglish
Pages (from-to)355-371
Number of pages17
JournalParallel Computing
Volume29
Issue number3
DOIs
Publication statusPublished - 2003 Mar
Externally publishedYes

Fingerprint

Volume Rendering
Volume rendering
Simulator
Visualization
Simulators
Earth (planet)
Parallelization
Subsystem
High Performance
Parallel Optimization
Vectorization
Dynamic Load Balancing
Data visualization
Performance Optimization
Data Visualization
Message passing
Parallel Computers
Message Passing
Dynamic loads
Vertex of a graph

Keywords

  • Earth Simulator
  • Large-scale parallel visualization
  • Parallel volume rendering
  • SMP cluster architecture
  • Unstructured datasets

ASJC Scopus subject areas

  • Computer Science Applications
  • Hardware and Architecture
  • Control and Systems Engineering

Cite this

Optimizing parallel performance of unstructured volume rendering for the earth simulator. / Chen, Li; Fujishiro, Issei; Nakajima, Kengo.

In: Parallel Computing, Vol. 29, No. 3, 03.2003, p. 355-371.

Research output: Contribution to journalArticle

@article{f68408204169448bb112ae1b1a4ff1ff,
title = "Optimizing parallel performance of unstructured volume rendering for the earth simulator",
abstract = "A scalable and high-performance parallel visualization subsystem has been developed in GeoFEM for the Earth Simulator (ES). As part of the ES project in Japan, the proposed subsystem is effective for the visualization of huge-scale unstructured datasets, and can be concurrent with computation on the same high-performance parallel computer. Moreover, some parallel visualization modules have obtained a good parallel performance, covering scalar, vector and tensor fields. This paper will take volume rendering method as an example to describe a number of efficient parallel performance optimization strategies we adopted for large-scale unstructured data visualization on SMP cluster machines, including suitable design of visualization method, the three-level hybrid parallelization which means message passing for inter-SMP node communication, loop directives for intra-SMP node parallelization, and vectorization for each processing element, plus dynamic load balancing. Good visualization images and high parallel performance have been achieved on the ES, thus demonstrating the feasibility and effectiveness of the proposed method.",
keywords = "Earth Simulator, Large-scale parallel visualization, Parallel volume rendering, SMP cluster architecture, Unstructured datasets",
author = "Li Chen and Issei Fujishiro and Kengo Nakajima",
year = "2003",
month = "3",
doi = "10.1016/S0167-8191(02)00251-X",
language = "English",
volume = "29",
pages = "355--371",
journal = "Parallel Computing",
issn = "0167-8191",
publisher = "Elsevier",
number = "3",

}

TY - JOUR

T1 - Optimizing parallel performance of unstructured volume rendering for the earth simulator

AU - Chen, Li

AU - Fujishiro, Issei

AU - Nakajima, Kengo

PY - 2003/3

Y1 - 2003/3

N2 - A scalable and high-performance parallel visualization subsystem has been developed in GeoFEM for the Earth Simulator (ES). As part of the ES project in Japan, the proposed subsystem is effective for the visualization of huge-scale unstructured datasets, and can be concurrent with computation on the same high-performance parallel computer. Moreover, some parallel visualization modules have obtained a good parallel performance, covering scalar, vector and tensor fields. This paper will take volume rendering method as an example to describe a number of efficient parallel performance optimization strategies we adopted for large-scale unstructured data visualization on SMP cluster machines, including suitable design of visualization method, the three-level hybrid parallelization which means message passing for inter-SMP node communication, loop directives for intra-SMP node parallelization, and vectorization for each processing element, plus dynamic load balancing. Good visualization images and high parallel performance have been achieved on the ES, thus demonstrating the feasibility and effectiveness of the proposed method.

AB - A scalable and high-performance parallel visualization subsystem has been developed in GeoFEM for the Earth Simulator (ES). As part of the ES project in Japan, the proposed subsystem is effective for the visualization of huge-scale unstructured datasets, and can be concurrent with computation on the same high-performance parallel computer. Moreover, some parallel visualization modules have obtained a good parallel performance, covering scalar, vector and tensor fields. This paper will take volume rendering method as an example to describe a number of efficient parallel performance optimization strategies we adopted for large-scale unstructured data visualization on SMP cluster machines, including suitable design of visualization method, the three-level hybrid parallelization which means message passing for inter-SMP node communication, loop directives for intra-SMP node parallelization, and vectorization for each processing element, plus dynamic load balancing. Good visualization images and high parallel performance have been achieved on the ES, thus demonstrating the feasibility and effectiveness of the proposed method.

KW - Earth Simulator

KW - Large-scale parallel visualization

KW - Parallel volume rendering

KW - SMP cluster architecture

KW - Unstructured datasets

UR - http://www.scopus.com/inward/record.url?scp=0037362982&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=0037362982&partnerID=8YFLogxK

U2 - 10.1016/S0167-8191(02)00251-X

DO - 10.1016/S0167-8191(02)00251-X

M3 - Article

AN - SCOPUS:0037362982

VL - 29

SP - 355

EP - 371

JO - Parallel Computing

JF - Parallel Computing

SN - 0167-8191

IS - 3

ER -