Speeding up and parallelizing the GARFIELD+

A. Sheharyar, Othmane Bouhali, A. Castaneda

Research output: Contribution to journalConference article

Abstract

Garfield++ is a toolkit for the simulation of particle detectors that use gas and semi-conductors as sensitive medium. It takes enormous amount of time to complete the simulation of complex scenarios such as those involving high detector voltages, gases with large gains, or electric field meshes with large number of elements. We observed that most of the simulation time is being consumed in finding the correct element in the electric field mesh. We optimized the element search operation and achieved significant boost in the speed up. In addition. We added the parallel computing support in the toolkit to simulate multiple events simultaneously over multiple machines. In this paper, we present our approach of speeding up the computations and benchmark results.

Original languageEnglish
Article number06004
JournalEPJ Web of Conferences
Volume174
DOIs
Publication statusPublished - 21 Feb 2018

Fingerprint

mesh
electric fields
simulation
radiation counters
acceleration (physics)
gases
conductors
detectors
electric potential

ASJC Scopus subject areas

  • Physics and Astronomy(all)

Cite this

Speeding up and parallelizing the GARFIELD+. / Sheharyar, A.; Bouhali, Othmane; Castaneda, A.

In: EPJ Web of Conferences, Vol. 174, 06004, 21.02.2018.

Research output: Contribution to journalConference article

Sheharyar, A. ; Bouhali, Othmane ; Castaneda, A. / Speeding up and parallelizing the GARFIELD+. In: EPJ Web of Conferences. 2018 ; Vol. 174.
@article{a4a5412e0d4d4eb486c7487354db2458,
title = "Speeding up and parallelizing the GARFIELD+",
abstract = "Garfield++ is a toolkit for the simulation of particle detectors that use gas and semi-conductors as sensitive medium. It takes enormous amount of time to complete the simulation of complex scenarios such as those involving high detector voltages, gases with large gains, or electric field meshes with large number of elements. We observed that most of the simulation time is being consumed in finding the correct element in the electric field mesh. We optimized the element search operation and achieved significant boost in the speed up. In addition. We added the parallel computing support in the toolkit to simulate multiple events simultaneously over multiple machines. In this paper, we present our approach of speeding up the computations and benchmark results.",
author = "A. Sheharyar and Othmane Bouhali and A. Castaneda",
year = "2018",
month = "2",
day = "21",
doi = "10.1051/epjconf/201817406004",
language = "English",
volume = "174",
journal = "EPJ Web of Conferences",
issn = "2101-6275",
publisher = "EDP Sciences",

}

TY - JOUR

T1 - Speeding up and parallelizing the GARFIELD+

AU - Sheharyar, A.

AU - Bouhali, Othmane

AU - Castaneda, A.

PY - 2018/2/21

Y1 - 2018/2/21

N2 - Garfield++ is a toolkit for the simulation of particle detectors that use gas and semi-conductors as sensitive medium. It takes enormous amount of time to complete the simulation of complex scenarios such as those involving high detector voltages, gases with large gains, or electric field meshes with large number of elements. We observed that most of the simulation time is being consumed in finding the correct element in the electric field mesh. We optimized the element search operation and achieved significant boost in the speed up. In addition. We added the parallel computing support in the toolkit to simulate multiple events simultaneously over multiple machines. In this paper, we present our approach of speeding up the computations and benchmark results.

AB - Garfield++ is a toolkit for the simulation of particle detectors that use gas and semi-conductors as sensitive medium. It takes enormous amount of time to complete the simulation of complex scenarios such as those involving high detector voltages, gases with large gains, or electric field meshes with large number of elements. We observed that most of the simulation time is being consumed in finding the correct element in the electric field mesh. We optimized the element search operation and achieved significant boost in the speed up. In addition. We added the parallel computing support in the toolkit to simulate multiple events simultaneously over multiple machines. In this paper, we present our approach of speeding up the computations and benchmark results.

UR - http://www.scopus.com/inward/record.url?scp=85043243584&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85043243584&partnerID=8YFLogxK

U2 - 10.1051/epjconf/201817406004

DO - 10.1051/epjconf/201817406004

M3 - Conference article

AN - SCOPUS:85043243584

VL - 174

JO - EPJ Web of Conferences

JF - EPJ Web of Conferences

SN - 2101-6275

M1 - 06004

ER -