Optimizing threshold for extreme scale analysis

Robert Maynard, Kenneth Moreland, Utkarsh Atyachit, Berk Geveci, Kwan-Liu Ma

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

4 Citations (Scopus)

Abstract

As the HPC community starts focusing its efforts towards exascale, it becomes clear that we are looking at machines with a billion way concurrency. Although parallel computing has been at the core of the performance gains achieved until now, scaling over 1,000 times the current concurrency can be challenging. As discussed in this paper, even the smallest memory access and synchronization overheads can cause major bottlenecks at this scale. As we develop new software and adapt existing algorithms for exascale, we need to be cognizant of such pitfalls. In this paper, we document our experience with optimizing a fairly common and parallelizable visualization algorithm, threshold of cells based on scalar values, for such highly concurrent architectures. Our experiments help us identify design patterns that can be generalized for other visualization algorithms as well. We discuss our implementation within the Dax toolkit, which is a framework for data analysis and visualization at extreme scale. The Dax toolkit employs the patterns discussed here within the framework's scaffolding to make it easier for algorithm developers to write algorithms without having to worry about such scaling issues.

Original languageEnglish (US)
Title of host publicationProceedings of SPIE-IS and T Electronic Imaging - Visualization and Data Analysis 2013
Volume8654
DOIs
StatePublished - Apr 10 2013
Externally publishedYes
EventVisualization and Data Analysis 2013 - Burlingame, CA, United States
Duration: Feb 4 2013Feb 6 2013

Other

OtherVisualization and Data Analysis 2013
CountryUnited States
CityBurlingame, CA
Period2/4/132/6/13

Fingerprint

Extremes
thresholds
Concurrency
Visualization
Scaling
scaling
Data visualization
Data Visualization
Design Patterns
photographic developers
Parallel processing systems
Parallel Computing
Concurrent
synchronism
Data analysis
Synchronization
Scalar
scalars
computer programs
Data storage equipment

Keywords

  • Concurrent Programming
  • Programming Techniques
  • Software

ASJC Scopus subject areas

  • Electronic, Optical and Magnetic Materials
  • Condensed Matter Physics
  • Computer Science Applications
  • Applied Mathematics
  • Electrical and Electronic Engineering

Cite this

Maynard, R., Moreland, K., Atyachit, U., Geveci, B., & Ma, K-L. (2013). Optimizing threshold for extreme scale analysis. In Proceedings of SPIE-IS and T Electronic Imaging - Visualization and Data Analysis 2013 (Vol. 8654). [86540Y] https://doi.org/10.1117/12.2007320

Optimizing threshold for extreme scale analysis. / Maynard, Robert; Moreland, Kenneth; Atyachit, Utkarsh; Geveci, Berk; Ma, Kwan-Liu.

Proceedings of SPIE-IS and T Electronic Imaging - Visualization and Data Analysis 2013. Vol. 8654 2013. 86540Y.

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

Maynard, R, Moreland, K, Atyachit, U, Geveci, B & Ma, K-L 2013, Optimizing threshold for extreme scale analysis. in Proceedings of SPIE-IS and T Electronic Imaging - Visualization and Data Analysis 2013. vol. 8654, 86540Y, Visualization and Data Analysis 2013, Burlingame, CA, United States, 2/4/13. https://doi.org/10.1117/12.2007320
Maynard R, Moreland K, Atyachit U, Geveci B, Ma K-L. Optimizing threshold for extreme scale analysis. In Proceedings of SPIE-IS and T Electronic Imaging - Visualization and Data Analysis 2013. Vol. 8654. 2013. 86540Y https://doi.org/10.1117/12.2007320
Maynard, Robert ; Moreland, Kenneth ; Atyachit, Utkarsh ; Geveci, Berk ; Ma, Kwan-Liu. / Optimizing threshold for extreme scale analysis. Proceedings of SPIE-IS and T Electronic Imaging - Visualization and Data Analysis 2013. Vol. 8654 2013.
@inproceedings{cbe1c626069545919b2bb76c17dc883a,
title = "Optimizing threshold for extreme scale analysis",
abstract = "As the HPC community starts focusing its efforts towards exascale, it becomes clear that we are looking at machines with a billion way concurrency. Although parallel computing has been at the core of the performance gains achieved until now, scaling over 1,000 times the current concurrency can be challenging. As discussed in this paper, even the smallest memory access and synchronization overheads can cause major bottlenecks at this scale. As we develop new software and adapt existing algorithms for exascale, we need to be cognizant of such pitfalls. In this paper, we document our experience with optimizing a fairly common and parallelizable visualization algorithm, threshold of cells based on scalar values, for such highly concurrent architectures. Our experiments help us identify design patterns that can be generalized for other visualization algorithms as well. We discuss our implementation within the Dax toolkit, which is a framework for data analysis and visualization at extreme scale. The Dax toolkit employs the patterns discussed here within the framework's scaffolding to make it easier for algorithm developers to write algorithms without having to worry about such scaling issues.",
keywords = "Concurrent Programming, Programming Techniques, Software",
author = "Robert Maynard and Kenneth Moreland and Utkarsh Atyachit and Berk Geveci and Kwan-Liu Ma",
year = "2013",
month = "4",
day = "10",
doi = "10.1117/12.2007320",
language = "English (US)",
isbn = "9780819494276",
volume = "8654",
booktitle = "Proceedings of SPIE-IS and T Electronic Imaging - Visualization and Data Analysis 2013",

}

TY - GEN

T1 - Optimizing threshold for extreme scale analysis

AU - Maynard, Robert

AU - Moreland, Kenneth

AU - Atyachit, Utkarsh

AU - Geveci, Berk

AU - Ma, Kwan-Liu

PY - 2013/4/10

Y1 - 2013/4/10

N2 - As the HPC community starts focusing its efforts towards exascale, it becomes clear that we are looking at machines with a billion way concurrency. Although parallel computing has been at the core of the performance gains achieved until now, scaling over 1,000 times the current concurrency can be challenging. As discussed in this paper, even the smallest memory access and synchronization overheads can cause major bottlenecks at this scale. As we develop new software and adapt existing algorithms for exascale, we need to be cognizant of such pitfalls. In this paper, we document our experience with optimizing a fairly common and parallelizable visualization algorithm, threshold of cells based on scalar values, for such highly concurrent architectures. Our experiments help us identify design patterns that can be generalized for other visualization algorithms as well. We discuss our implementation within the Dax toolkit, which is a framework for data analysis and visualization at extreme scale. The Dax toolkit employs the patterns discussed here within the framework's scaffolding to make it easier for algorithm developers to write algorithms without having to worry about such scaling issues.

AB - As the HPC community starts focusing its efforts towards exascale, it becomes clear that we are looking at machines with a billion way concurrency. Although parallel computing has been at the core of the performance gains achieved until now, scaling over 1,000 times the current concurrency can be challenging. As discussed in this paper, even the smallest memory access and synchronization overheads can cause major bottlenecks at this scale. As we develop new software and adapt existing algorithms for exascale, we need to be cognizant of such pitfalls. In this paper, we document our experience with optimizing a fairly common and parallelizable visualization algorithm, threshold of cells based on scalar values, for such highly concurrent architectures. Our experiments help us identify design patterns that can be generalized for other visualization algorithms as well. We discuss our implementation within the Dax toolkit, which is a framework for data analysis and visualization at extreme scale. The Dax toolkit employs the patterns discussed here within the framework's scaffolding to make it easier for algorithm developers to write algorithms without having to worry about such scaling issues.

KW - Concurrent Programming

KW - Programming Techniques

KW - Software

UR - http://www.scopus.com/inward/record.url?scp=84875823449&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=84875823449&partnerID=8YFLogxK

U2 - 10.1117/12.2007320

DO - 10.1117/12.2007320

M3 - Conference contribution

SN - 9780819494276

VL - 8654

BT - Proceedings of SPIE-IS and T Electronic Imaging - Visualization and Data Analysis 2013

ER -