StePS

A multi-GPU cosmological N-body Code for compactified simulations

G. Rácz, I. Szapudi, L. Dobos, I. Csabai, A. S. Szalay

Research output: Article

Abstract

We present the multi-GPU realization of the StePS(Stereographically Projected Cosmological Simulations) algorithm with MPI–OpenMP–CUDA hybrid parallelization and nearly ideal scale-out to multiple compute nodes. Our new zoom-in cosmological direct N-body simulation method simulates the infinite universe with unprecedented dynamic range for a given amount of memory and, in contrast to traditional periodic simulations, its fundamental geometry and topology match observations. By using a spherical geometry instead of periodic boundary conditions, and gradually decreasing the mass resolution with radius, our code is capable of running simulations with a few gigaparsecs in diameter and with a mass resolution of ∼109M in the center in four days on three compute nodes with four GTX 1080 Ti GPUs in each. The code can also be used to run extremely fast simulations with reasonable resolution for fitting cosmological parameters. These simulations are useful for prediction needs of large surveys. The StePScode is publicly available for the research community.

Original languageEnglish
Article number100303
JournalAstronomy and Computing
Volume28
DOIs
Publication statusPublished - júl. 1 2019

Fingerprint

Geometry
simulation
Topology
Boundary conditions
Data storage equipment
geometry
topology
dynamic range
Graphics processing unit
code
boundary condition
universe
boundary conditions
radii
prediction
predictions

ASJC Scopus subject areas

  • Astronomy and Astrophysics
  • Computer Science Applications
  • Space and Planetary Science

Cite this

StePS : A multi-GPU cosmological N-body Code for compactified simulations. / Rácz, G.; Szapudi, I.; Dobos, L.; Csabai, I.; Szalay, A. S.

In: Astronomy and Computing, Vol. 28, 100303, 01.07.2019.

Research output: Article

Rácz, G. ; Szapudi, I. ; Dobos, L. ; Csabai, I. ; Szalay, A. S. / StePS : A multi-GPU cosmological N-body Code for compactified simulations. In: Astronomy and Computing. 2019 ; Vol. 28.
@article{30812b7a20f14ec8b896409edbb30ea7,
title = "StePS: A multi-GPU cosmological N-body Code for compactified simulations",
abstract = "We present the multi-GPU realization of the StePS(Stereographically Projected Cosmological Simulations) algorithm with MPI–OpenMP–CUDA hybrid parallelization and nearly ideal scale-out to multiple compute nodes. Our new zoom-in cosmological direct N-body simulation method simulates the infinite universe with unprecedented dynamic range for a given amount of memory and, in contrast to traditional periodic simulations, its fundamental geometry and topology match observations. By using a spherical geometry instead of periodic boundary conditions, and gradually decreasing the mass resolution with radius, our code is capable of running simulations with a few gigaparsecs in diameter and with a mass resolution of ∼109M⊙ in the center in four days on three compute nodes with four GTX 1080 Ti GPUs in each. The code can also be used to run extremely fast simulations with reasonable resolution for fitting cosmological parameters. These simulations are useful for prediction needs of large surveys. The StePScode is publicly available for the research community.",
keywords = "Dark matter, Graphics processors, Large-scale structure of universe, Methods: N-body simulations, Methods: Numerical",
author = "G. R{\'a}cz and I. Szapudi and L. Dobos and I. Csabai and Szalay, {A. S.}",
year = "2019",
month = "7",
day = "1",
doi = "10.1016/j.ascom.2019.100303",
language = "English",
volume = "28",
journal = "Astronomy and Computing",
issn = "2213-1337",
publisher = "Elsevier BV",

}

TY - JOUR

T1 - StePS

T2 - A multi-GPU cosmological N-body Code for compactified simulations

AU - Rácz, G.

AU - Szapudi, I.

AU - Dobos, L.

AU - Csabai, I.

AU - Szalay, A. S.

PY - 2019/7/1

Y1 - 2019/7/1

N2 - We present the multi-GPU realization of the StePS(Stereographically Projected Cosmological Simulations) algorithm with MPI–OpenMP–CUDA hybrid parallelization and nearly ideal scale-out to multiple compute nodes. Our new zoom-in cosmological direct N-body simulation method simulates the infinite universe with unprecedented dynamic range for a given amount of memory and, in contrast to traditional periodic simulations, its fundamental geometry and topology match observations. By using a spherical geometry instead of periodic boundary conditions, and gradually decreasing the mass resolution with radius, our code is capable of running simulations with a few gigaparsecs in diameter and with a mass resolution of ∼109M⊙ in the center in four days on three compute nodes with four GTX 1080 Ti GPUs in each. The code can also be used to run extremely fast simulations with reasonable resolution for fitting cosmological parameters. These simulations are useful for prediction needs of large surveys. The StePScode is publicly available for the research community.

AB - We present the multi-GPU realization of the StePS(Stereographically Projected Cosmological Simulations) algorithm with MPI–OpenMP–CUDA hybrid parallelization and nearly ideal scale-out to multiple compute nodes. Our new zoom-in cosmological direct N-body simulation method simulates the infinite universe with unprecedented dynamic range for a given amount of memory and, in contrast to traditional periodic simulations, its fundamental geometry and topology match observations. By using a spherical geometry instead of periodic boundary conditions, and gradually decreasing the mass resolution with radius, our code is capable of running simulations with a few gigaparsecs in diameter and with a mass resolution of ∼109M⊙ in the center in four days on three compute nodes with four GTX 1080 Ti GPUs in each. The code can also be used to run extremely fast simulations with reasonable resolution for fitting cosmological parameters. These simulations are useful for prediction needs of large surveys. The StePScode is publicly available for the research community.

KW - Dark matter

KW - Graphics processors

KW - Large-scale structure of universe

KW - Methods: N-body simulations

KW - Methods: Numerical

UR - http://www.scopus.com/inward/record.url?scp=85068916435&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85068916435&partnerID=8YFLogxK

U2 - 10.1016/j.ascom.2019.100303

DO - 10.1016/j.ascom.2019.100303

M3 - Article

VL - 28

JO - Astronomy and Computing

JF - Astronomy and Computing

SN - 2213-1337

M1 - 100303

ER -