A job scheduling design for visualization services using GPU clusters

Wei Hsien Hsu, Chun Fu Wang, Kwan-Liu Ma, Hongfeng Yu, Jacqueline H. Chen

Research output: Contribution to conferencePaper

5 Citations (Scopus)

Abstract

Modern large-scale heterogeneous computers incorporating GPUs offer impressive processing capabilities. It is desirable to fully utilize such systems for serving multiple users concurrently to visualize large data at interactive rates. However, as the disparity between data transfer speed and compute speed continues to increase in heterogeneous systems, data locality becomes crucial for performance. We present a new job scheduling design to support multi-user exploration of large data in a heterogeneous computing environment, achieving near optimal data locality and minimizing I/O overhead. The targeted application is a parallel visualization system which allows multiple users to render large volumetric data sets in both interactive mode and batch mode. We present a cost model to assess the performance of parallel volume rendering and quantify the efficiency of job scheduling. We have tested our job scheduling scheme on two heterogeneous systems with different configurations. The largest test volume data used in our study has over two billion grid points. The timing results demonstrate that our design effectively improves data locality for complex multi-user job scheduling problems, leading to better overall performance of the service.

Original languageEnglish (US)
Pages523-533
Number of pages11
DOIs
StatePublished - Dec 12 2012
Event2012 IEEE International Conference on Cluster Computing, CLUSTER 2012 - Beijing, China
Duration: Sep 24 2012Sep 28 2012

Other

Other2012 IEEE International Conference on Cluster Computing, CLUSTER 2012
CountryChina
CityBeijing
Period9/24/129/28/12

Fingerprint

Visualization
Scheduling
Volume rendering
Data transfer
Graphics processing unit
Processing
Costs

Keywords

  • GPU clusters
  • job scheduling
  • multi-user volume rendering
  • parallel volume visualizatoin

ASJC Scopus subject areas

  • Software

Cite this

Hsu, W. H., Wang, C. F., Ma, K-L., Yu, H., & Chen, J. H. (2012). A job scheduling design for visualization services using GPU clusters. 523-533. Paper presented at 2012 IEEE International Conference on Cluster Computing, CLUSTER 2012, Beijing, China. https://doi.org/10.1109/CLUSTER.2012.63

A job scheduling design for visualization services using GPU clusters. / Hsu, Wei Hsien; Wang, Chun Fu; Ma, Kwan-Liu; Yu, Hongfeng; Chen, Jacqueline H.

2012. 523-533 Paper presented at 2012 IEEE International Conference on Cluster Computing, CLUSTER 2012, Beijing, China.

Research output: Contribution to conferencePaper

Hsu, WH, Wang, CF, Ma, K-L, Yu, H & Chen, JH 2012, 'A job scheduling design for visualization services using GPU clusters' Paper presented at 2012 IEEE International Conference on Cluster Computing, CLUSTER 2012, Beijing, China, 9/24/12 - 9/28/12, pp. 523-533. https://doi.org/10.1109/CLUSTER.2012.63
Hsu WH, Wang CF, Ma K-L, Yu H, Chen JH. A job scheduling design for visualization services using GPU clusters. 2012. Paper presented at 2012 IEEE International Conference on Cluster Computing, CLUSTER 2012, Beijing, China. https://doi.org/10.1109/CLUSTER.2012.63
Hsu, Wei Hsien ; Wang, Chun Fu ; Ma, Kwan-Liu ; Yu, Hongfeng ; Chen, Jacqueline H. / A job scheduling design for visualization services using GPU clusters. Paper presented at 2012 IEEE International Conference on Cluster Computing, CLUSTER 2012, Beijing, China.11 p.
@conference{36b66201331345dabaa8a64a91ca4286,
title = "A job scheduling design for visualization services using GPU clusters",
abstract = "Modern large-scale heterogeneous computers incorporating GPUs offer impressive processing capabilities. It is desirable to fully utilize such systems for serving multiple users concurrently to visualize large data at interactive rates. However, as the disparity between data transfer speed and compute speed continues to increase in heterogeneous systems, data locality becomes crucial for performance. We present a new job scheduling design to support multi-user exploration of large data in a heterogeneous computing environment, achieving near optimal data locality and minimizing I/O overhead. The targeted application is a parallel visualization system which allows multiple users to render large volumetric data sets in both interactive mode and batch mode. We present a cost model to assess the performance of parallel volume rendering and quantify the efficiency of job scheduling. We have tested our job scheduling scheme on two heterogeneous systems with different configurations. The largest test volume data used in our study has over two billion grid points. The timing results demonstrate that our design effectively improves data locality for complex multi-user job scheduling problems, leading to better overall performance of the service.",
keywords = "GPU clusters, job scheduling, multi-user volume rendering, parallel volume visualizatoin",
author = "Hsu, {Wei Hsien} and Wang, {Chun Fu} and Kwan-Liu Ma and Hongfeng Yu and Chen, {Jacqueline H.}",
year = "2012",
month = "12",
day = "12",
doi = "10.1109/CLUSTER.2012.63",
language = "English (US)",
pages = "523--533",
note = "2012 IEEE International Conference on Cluster Computing, CLUSTER 2012 ; Conference date: 24-09-2012 Through 28-09-2012",

}

TY - CONF

T1 - A job scheduling design for visualization services using GPU clusters

AU - Hsu, Wei Hsien

AU - Wang, Chun Fu

AU - Ma, Kwan-Liu

AU - Yu, Hongfeng

AU - Chen, Jacqueline H.

PY - 2012/12/12

Y1 - 2012/12/12

N2 - Modern large-scale heterogeneous computers incorporating GPUs offer impressive processing capabilities. It is desirable to fully utilize such systems for serving multiple users concurrently to visualize large data at interactive rates. However, as the disparity between data transfer speed and compute speed continues to increase in heterogeneous systems, data locality becomes crucial for performance. We present a new job scheduling design to support multi-user exploration of large data in a heterogeneous computing environment, achieving near optimal data locality and minimizing I/O overhead. The targeted application is a parallel visualization system which allows multiple users to render large volumetric data sets in both interactive mode and batch mode. We present a cost model to assess the performance of parallel volume rendering and quantify the efficiency of job scheduling. We have tested our job scheduling scheme on two heterogeneous systems with different configurations. The largest test volume data used in our study has over two billion grid points. The timing results demonstrate that our design effectively improves data locality for complex multi-user job scheduling problems, leading to better overall performance of the service.

AB - Modern large-scale heterogeneous computers incorporating GPUs offer impressive processing capabilities. It is desirable to fully utilize such systems for serving multiple users concurrently to visualize large data at interactive rates. However, as the disparity between data transfer speed and compute speed continues to increase in heterogeneous systems, data locality becomes crucial for performance. We present a new job scheduling design to support multi-user exploration of large data in a heterogeneous computing environment, achieving near optimal data locality and minimizing I/O overhead. The targeted application is a parallel visualization system which allows multiple users to render large volumetric data sets in both interactive mode and batch mode. We present a cost model to assess the performance of parallel volume rendering and quantify the efficiency of job scheduling. We have tested our job scheduling scheme on two heterogeneous systems with different configurations. The largest test volume data used in our study has over two billion grid points. The timing results demonstrate that our design effectively improves data locality for complex multi-user job scheduling problems, leading to better overall performance of the service.

KW - GPU clusters

KW - job scheduling

KW - multi-user volume rendering

KW - parallel volume visualizatoin

UR - http://www.scopus.com/inward/record.url?scp=84870697204&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=84870697204&partnerID=8YFLogxK

U2 - 10.1109/CLUSTER.2012.63

DO - 10.1109/CLUSTER.2012.63

M3 - Paper

SP - 523

EP - 533

ER -