A Visual Analytics System for Optimizing Communications in Massively Parallel Applications

Takanori Fujiwara, Preeti Malakar, Khairi Reda, Venkatram Vishwanath, Michael E. Papka, Kwan-Liu Ma

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

4 Citations (Scopus)

Abstract

Current and future supercomputers have tens of thousands of compute nodes interconnected with high-dimensional networks and complex network topologies for improved performance. Application developers are required to write scalable parallel programs in order to achieve high throughput on these machines. Application performance is largely determined by efficient inter-process communication. A common way to analyze and optimize performance is through profiling parallel codes to identify communication bottlenecks. However, understanding gigabytes of profiled at a is not a trivial task. In this paper, we present a visual analytics system for identifying the scalability bottlenecks and improving the communication efficiency of massively parallel applications. Visualization methods used in this system are designed to comprehend large-scale and varied communication patterns on thousands of nodes in complex networks such as the 5D torus and the dragonfly. We also present efficient rerouting and remapping algorithms that can be coupled with our interactive visual analytics design for performance optimization. We demonstrate the utility of our system with several case studies using three benchmark applications on two leading supercomputers. The mapping suggestion from our system led to 38% improvement in hop-bytes for Mini AMR application on 4,096 MPI processes.

Original languageEnglish (US)
Title of host publication2017 IEEE Conference on Visual Analytics Science and Technology, VAST 2017 - Proceedings
EditorsTobias Schreck, Brian Fisher, Shixia Liu
PublisherInstitute of Electrical and Electronics Engineers Inc.
Pages59-70
Number of pages12
ISBN (Electronic)9781538631638
DOIs
StatePublished - Dec 21 2018
Event2017 IEEE Conference on Visual Analytics Science and Technology, VAST 2017 - Phoenix, United States
Duration: Oct 1 2017Oct 6 2017

Other

Other2017 IEEE Conference on Visual Analytics Science and Technology, VAST 2017
CountryUnited States
CityPhoenix
Period10/1/1710/6/17

Fingerprint

Communication
Supercomputers
Complex networks
Scalability
Visualization
Throughput
Topology

Keywords

  • Communication visualization
  • Parallel communications
  • Performance analysis
  • Supercomputing
  • Visual analytics

ASJC Scopus subject areas

  • Computer Graphics and Computer-Aided Design
  • Human-Computer Interaction
  • Information Systems
  • Media Technology

Cite this

Fujiwara, T., Malakar, P., Reda, K., Vishwanath, V., Papka, M. E., & Ma, K-L. (2018). A Visual Analytics System for Optimizing Communications in Massively Parallel Applications. In T. Schreck, B. Fisher, & S. Liu (Eds.), 2017 IEEE Conference on Visual Analytics Science and Technology, VAST 2017 - Proceedings (pp. 59-70). [8585646] Institute of Electrical and Electronics Engineers Inc.. https://doi.org/10.1109/VAST.2017.8585646

A Visual Analytics System for Optimizing Communications in Massively Parallel Applications. / Fujiwara, Takanori; Malakar, Preeti; Reda, Khairi; Vishwanath, Venkatram; Papka, Michael E.; Ma, Kwan-Liu.

2017 IEEE Conference on Visual Analytics Science and Technology, VAST 2017 - Proceedings. ed. / Tobias Schreck; Brian Fisher; Shixia Liu. Institute of Electrical and Electronics Engineers Inc., 2018. p. 59-70 8585646.

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

Fujiwara, T, Malakar, P, Reda, K, Vishwanath, V, Papka, ME & Ma, K-L 2018, A Visual Analytics System for Optimizing Communications in Massively Parallel Applications. in T Schreck, B Fisher & S Liu (eds), 2017 IEEE Conference on Visual Analytics Science and Technology, VAST 2017 - Proceedings., 8585646, Institute of Electrical and Electronics Engineers Inc., pp. 59-70, 2017 IEEE Conference on Visual Analytics Science and Technology, VAST 2017, Phoenix, United States, 10/1/17. https://doi.org/10.1109/VAST.2017.8585646
Fujiwara T, Malakar P, Reda K, Vishwanath V, Papka ME, Ma K-L. A Visual Analytics System for Optimizing Communications in Massively Parallel Applications. In Schreck T, Fisher B, Liu S, editors, 2017 IEEE Conference on Visual Analytics Science and Technology, VAST 2017 - Proceedings. Institute of Electrical and Electronics Engineers Inc. 2018. p. 59-70. 8585646 https://doi.org/10.1109/VAST.2017.8585646
Fujiwara, Takanori ; Malakar, Preeti ; Reda, Khairi ; Vishwanath, Venkatram ; Papka, Michael E. ; Ma, Kwan-Liu. / A Visual Analytics System for Optimizing Communications in Massively Parallel Applications. 2017 IEEE Conference on Visual Analytics Science and Technology, VAST 2017 - Proceedings. editor / Tobias Schreck ; Brian Fisher ; Shixia Liu. Institute of Electrical and Electronics Engineers Inc., 2018. pp. 59-70
@inproceedings{00a50a1c00ba477087029faa7887f504,
title = "A Visual Analytics System for Optimizing Communications in Massively Parallel Applications",
abstract = "Current and future supercomputers have tens of thousands of compute nodes interconnected with high-dimensional networks and complex network topologies for improved performance. Application developers are required to write scalable parallel programs in order to achieve high throughput on these machines. Application performance is largely determined by efficient inter-process communication. A common way to analyze and optimize performance is through profiling parallel codes to identify communication bottlenecks. However, understanding gigabytes of profiled at a is not a trivial task. In this paper, we present a visual analytics system for identifying the scalability bottlenecks and improving the communication efficiency of massively parallel applications. Visualization methods used in this system are designed to comprehend large-scale and varied communication patterns on thousands of nodes in complex networks such as the 5D torus and the dragonfly. We also present efficient rerouting and remapping algorithms that can be coupled with our interactive visual analytics design for performance optimization. We demonstrate the utility of our system with several case studies using three benchmark applications on two leading supercomputers. The mapping suggestion from our system led to 38{\%} improvement in hop-bytes for Mini AMR application on 4,096 MPI processes.",
keywords = "Communication visualization, Parallel communications, Performance analysis, Supercomputing, Visual analytics",
author = "Takanori Fujiwara and Preeti Malakar and Khairi Reda and Venkatram Vishwanath and Papka, {Michael E.} and Kwan-Liu Ma",
year = "2018",
month = "12",
day = "21",
doi = "10.1109/VAST.2017.8585646",
language = "English (US)",
pages = "59--70",
editor = "Tobias Schreck and Brian Fisher and Shixia Liu",
booktitle = "2017 IEEE Conference on Visual Analytics Science and Technology, VAST 2017 - Proceedings",
publisher = "Institute of Electrical and Electronics Engineers Inc.",
address = "United States",

}

TY - GEN

T1 - A Visual Analytics System for Optimizing Communications in Massively Parallel Applications

AU - Fujiwara, Takanori

AU - Malakar, Preeti

AU - Reda, Khairi

AU - Vishwanath, Venkatram

AU - Papka, Michael E.

AU - Ma, Kwan-Liu

PY - 2018/12/21

Y1 - 2018/12/21

N2 - Current and future supercomputers have tens of thousands of compute nodes interconnected with high-dimensional networks and complex network topologies for improved performance. Application developers are required to write scalable parallel programs in order to achieve high throughput on these machines. Application performance is largely determined by efficient inter-process communication. A common way to analyze and optimize performance is through profiling parallel codes to identify communication bottlenecks. However, understanding gigabytes of profiled at a is not a trivial task. In this paper, we present a visual analytics system for identifying the scalability bottlenecks and improving the communication efficiency of massively parallel applications. Visualization methods used in this system are designed to comprehend large-scale and varied communication patterns on thousands of nodes in complex networks such as the 5D torus and the dragonfly. We also present efficient rerouting and remapping algorithms that can be coupled with our interactive visual analytics design for performance optimization. We demonstrate the utility of our system with several case studies using three benchmark applications on two leading supercomputers. The mapping suggestion from our system led to 38% improvement in hop-bytes for Mini AMR application on 4,096 MPI processes.

AB - Current and future supercomputers have tens of thousands of compute nodes interconnected with high-dimensional networks and complex network topologies for improved performance. Application developers are required to write scalable parallel programs in order to achieve high throughput on these machines. Application performance is largely determined by efficient inter-process communication. A common way to analyze and optimize performance is through profiling parallel codes to identify communication bottlenecks. However, understanding gigabytes of profiled at a is not a trivial task. In this paper, we present a visual analytics system for identifying the scalability bottlenecks and improving the communication efficiency of massively parallel applications. Visualization methods used in this system are designed to comprehend large-scale and varied communication patterns on thousands of nodes in complex networks such as the 5D torus and the dragonfly. We also present efficient rerouting and remapping algorithms that can be coupled with our interactive visual analytics design for performance optimization. We demonstrate the utility of our system with several case studies using three benchmark applications on two leading supercomputers. The mapping suggestion from our system led to 38% improvement in hop-bytes for Mini AMR application on 4,096 MPI processes.

KW - Communication visualization

KW - Parallel communications

KW - Performance analysis

KW - Supercomputing

KW - Visual analytics

UR - http://www.scopus.com/inward/record.url?scp=85058637865&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85058637865&partnerID=8YFLogxK

U2 - 10.1109/VAST.2017.8585646

DO - 10.1109/VAST.2017.8585646

M3 - Conference contribution

AN - SCOPUS:85058637865

SP - 59

EP - 70

BT - 2017 IEEE Conference on Visual Analytics Science and Technology, VAST 2017 - Proceedings

A2 - Schreck, Tobias

A2 - Fisher, Brian

A2 - Liu, Shixia

PB - Institute of Electrical and Electronics Engineers Inc.

ER -