Orthogonal recursive bisection as data decomposition strategy for massively parallel cardiac simulations

Matthias Reumann, Blake G. Fitch, Aleksandr Rayshubskiy, Michael C. Pitman, John J. Rice

Research output: Contribution to journalArticle

Abstract

We present the orthogonal recursive bisection algorithm that hierarchically segments the anatomical model structure into subvolumes that are distributed to cores. The anatomy is derived from the Visible Human Project, with electrophysiology based on the FitzHugh-Nagumo (FHN) and ten Tusscher (TT04) models with monodomain diffusion. Benchmark simulations with up to 16,384 and 32,768 cores on IBM Blue Gene/P and L supercomputers for both FHN and TT04 results show good load balancing with almost perfect speedup factors that are close to linear with the number of cores. Hence, strong scaling is demonstrated. With 32,768 cores, a 1000 ms simulation of full heart beat requires about 6.5 min of wall clock time for a simulation of the FHN model. For the largest machine partitions, the simulations execute at a rate of 0.548 s (BG/P) and 0.394 s (BG/L) of wall clock time per 1 ms of simulation time. To our knowledge, these simulations show strong scaling to substantially higher numbers of cores than reported previously for organ-level simulation of the heart, thus significantly reducing run times. The ability to reduce runtimes could play a critical role in enabling wider use of cardiac models in research and clinical applications.

Original languageEnglish (US)
Pages (from-to)129-145
Number of pages17
JournalBiomedizinische Technik
Volume56
Issue number3
DOIs
StatePublished - Jun 1 2011
Externally publishedYes

Fingerprint

Decomposition
Clocks
Electrophysiology
Visible Human Projects
Supercomputers
Model structures
Anatomic Models
Benchmarking
Resource allocation
Genes
Anatomy
Research

Keywords

  • cardiac models
  • high performance computing
  • load balancing
  • massively parallel distributed memory supercomputer
  • performance scaling

ASJC Scopus subject areas

  • Biomedical Engineering
  • Medicine(all)

Cite this

Orthogonal recursive bisection as data decomposition strategy for massively parallel cardiac simulations. / Reumann, Matthias; Fitch, Blake G.; Rayshubskiy, Aleksandr; Pitman, Michael C.; Rice, John J.

In: Biomedizinische Technik, Vol. 56, No. 3, 01.06.2011, p. 129-145.

Research output: Contribution to journalArticle

Reumann, Matthias ; Fitch, Blake G. ; Rayshubskiy, Aleksandr ; Pitman, Michael C. ; Rice, John J. / Orthogonal recursive bisection as data decomposition strategy for massively parallel cardiac simulations. In: Biomedizinische Technik. 2011 ; Vol. 56, No. 3. pp. 129-145.
@article{05b03f67c0fd4fd0a6aa26b5e7e4f1a8,
title = "Orthogonal recursive bisection as data decomposition strategy for massively parallel cardiac simulations",
abstract = "We present the orthogonal recursive bisection algorithm that hierarchically segments the anatomical model structure into subvolumes that are distributed to cores. The anatomy is derived from the Visible Human Project, with electrophysiology based on the FitzHugh-Nagumo (FHN) and ten Tusscher (TT04) models with monodomain diffusion. Benchmark simulations with up to 16,384 and 32,768 cores on IBM Blue Gene/P and L supercomputers for both FHN and TT04 results show good load balancing with almost perfect speedup factors that are close to linear with the number of cores. Hence, strong scaling is demonstrated. With 32,768 cores, a 1000 ms simulation of full heart beat requires about 6.5 min of wall clock time for a simulation of the FHN model. For the largest machine partitions, the simulations execute at a rate of 0.548 s (BG/P) and 0.394 s (BG/L) of wall clock time per 1 ms of simulation time. To our knowledge, these simulations show strong scaling to substantially higher numbers of cores than reported previously for organ-level simulation of the heart, thus significantly reducing run times. The ability to reduce runtimes could play a critical role in enabling wider use of cardiac models in research and clinical applications.",
keywords = "cardiac models, high performance computing, load balancing, massively parallel distributed memory supercomputer, performance scaling",
author = "Matthias Reumann and Fitch, {Blake G.} and Aleksandr Rayshubskiy and Pitman, {Michael C.} and Rice, {John J.}",
year = "2011",
month = "6",
day = "1",
doi = "10.1515/BMT.2011.100",
language = "English (US)",
volume = "56",
pages = "129--145",
journal = "Biomedizinische Technik. Biomedical engineering",
issn = "0013-5585",
publisher = "Walter de Gruyter GmbH & Co. KG",
number = "3",

}

TY - JOUR

T1 - Orthogonal recursive bisection as data decomposition strategy for massively parallel cardiac simulations

AU - Reumann, Matthias

AU - Fitch, Blake G.

AU - Rayshubskiy, Aleksandr

AU - Pitman, Michael C.

AU - Rice, John J.

PY - 2011/6/1

Y1 - 2011/6/1

N2 - We present the orthogonal recursive bisection algorithm that hierarchically segments the anatomical model structure into subvolumes that are distributed to cores. The anatomy is derived from the Visible Human Project, with electrophysiology based on the FitzHugh-Nagumo (FHN) and ten Tusscher (TT04) models with monodomain diffusion. Benchmark simulations with up to 16,384 and 32,768 cores on IBM Blue Gene/P and L supercomputers for both FHN and TT04 results show good load balancing with almost perfect speedup factors that are close to linear with the number of cores. Hence, strong scaling is demonstrated. With 32,768 cores, a 1000 ms simulation of full heart beat requires about 6.5 min of wall clock time for a simulation of the FHN model. For the largest machine partitions, the simulations execute at a rate of 0.548 s (BG/P) and 0.394 s (BG/L) of wall clock time per 1 ms of simulation time. To our knowledge, these simulations show strong scaling to substantially higher numbers of cores than reported previously for organ-level simulation of the heart, thus significantly reducing run times. The ability to reduce runtimes could play a critical role in enabling wider use of cardiac models in research and clinical applications.

AB - We present the orthogonal recursive bisection algorithm that hierarchically segments the anatomical model structure into subvolumes that are distributed to cores. The anatomy is derived from the Visible Human Project, with electrophysiology based on the FitzHugh-Nagumo (FHN) and ten Tusscher (TT04) models with monodomain diffusion. Benchmark simulations with up to 16,384 and 32,768 cores on IBM Blue Gene/P and L supercomputers for both FHN and TT04 results show good load balancing with almost perfect speedup factors that are close to linear with the number of cores. Hence, strong scaling is demonstrated. With 32,768 cores, a 1000 ms simulation of full heart beat requires about 6.5 min of wall clock time for a simulation of the FHN model. For the largest machine partitions, the simulations execute at a rate of 0.548 s (BG/P) and 0.394 s (BG/L) of wall clock time per 1 ms of simulation time. To our knowledge, these simulations show strong scaling to substantially higher numbers of cores than reported previously for organ-level simulation of the heart, thus significantly reducing run times. The ability to reduce runtimes could play a critical role in enabling wider use of cardiac models in research and clinical applications.

KW - cardiac models

KW - high performance computing

KW - load balancing

KW - massively parallel distributed memory supercomputer

KW - performance scaling

UR - http://www.scopus.com/inward/record.url?scp=79959439631&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=79959439631&partnerID=8YFLogxK

U2 - 10.1515/BMT.2011.100

DO - 10.1515/BMT.2011.100

M3 - Article

C2 - 21657987

AN - SCOPUS:79959439631

VL - 56

SP - 129

EP - 145

JO - Biomedizinische Technik. Biomedical engineering

JF - Biomedizinische Technik. Biomedical engineering

SN - 0013-5585

IS - 3

ER -