Solving the Maxwell-Bloch equations for resonant nonlinear optics using GPUs

Research output: Contribution to journalArticle

5 Citations (Scopus)

Abstract

We solve the Maxwell-Bloch equations of resonant nonlinear optics using GPUs and compare the computation times with traditional single- and multithreaded programs. A detailed benchmarking of programs as a function of various parameters shows how the massive parallelism built into GPUs becomes more and more advantageous as the physical problem becomes more and more demanding. For the case of multimode light propagating through an inhomogeneously broadened medium of many-level quantum systems, the program executing on GPUs can be over 20 times faster than that executing on all cores of a modern CPU. The methods presented can be applied in a wide area of atomic physics where the time evolution of atomic ensembles is to be computed.

Original languageEnglish
Pages (from-to)1203-1210
Number of pages8
JournalComputer Physics Communications
Volume184
Issue number4
DOIs
Publication statusPublished - Apr 2013

Fingerprint

Nonlinear optics
nonlinear optics
Maxwell equations
Atomic physics
atomic physics
Benchmarking
Program processors
Graphics processing unit

Keywords

  • CUDA
  • GPU
  • Maxwell-Bloch equations
  • Pulse propagation
  • Quantum ensembles
  • Resonant nonlinear optics

ASJC Scopus subject areas

  • Hardware and Architecture
  • Physics and Astronomy(all)

Cite this

Solving the Maxwell-Bloch equations for resonant nonlinear optics using GPUs. / Demeter, G.

In: Computer Physics Communications, Vol. 184, No. 4, 04.2013, p. 1203-1210.

Research output: Contribution to journalArticle

@article{ed4e59446c8a425bb9215aa25d265979,
title = "Solving the Maxwell-Bloch equations for resonant nonlinear optics using GPUs",
abstract = "We solve the Maxwell-Bloch equations of resonant nonlinear optics using GPUs and compare the computation times with traditional single- and multithreaded programs. A detailed benchmarking of programs as a function of various parameters shows how the massive parallelism built into GPUs becomes more and more advantageous as the physical problem becomes more and more demanding. For the case of multimode light propagating through an inhomogeneously broadened medium of many-level quantum systems, the program executing on GPUs can be over 20 times faster than that executing on all cores of a modern CPU. The methods presented can be applied in a wide area of atomic physics where the time evolution of atomic ensembles is to be computed.",
keywords = "CUDA, GPU, Maxwell-Bloch equations, Pulse propagation, Quantum ensembles, Resonant nonlinear optics",
author = "G. Demeter",
year = "2013",
month = "4",
doi = "10.1016/j.cpc.2012.12.019",
language = "English",
volume = "184",
pages = "1203--1210",
journal = "Computer Physics Communications",
issn = "0010-4655",
publisher = "Elsevier",
number = "4",

}

TY - JOUR

T1 - Solving the Maxwell-Bloch equations for resonant nonlinear optics using GPUs

AU - Demeter, G.

PY - 2013/4

Y1 - 2013/4

N2 - We solve the Maxwell-Bloch equations of resonant nonlinear optics using GPUs and compare the computation times with traditional single- and multithreaded programs. A detailed benchmarking of programs as a function of various parameters shows how the massive parallelism built into GPUs becomes more and more advantageous as the physical problem becomes more and more demanding. For the case of multimode light propagating through an inhomogeneously broadened medium of many-level quantum systems, the program executing on GPUs can be over 20 times faster than that executing on all cores of a modern CPU. The methods presented can be applied in a wide area of atomic physics where the time evolution of atomic ensembles is to be computed.

AB - We solve the Maxwell-Bloch equations of resonant nonlinear optics using GPUs and compare the computation times with traditional single- and multithreaded programs. A detailed benchmarking of programs as a function of various parameters shows how the massive parallelism built into GPUs becomes more and more advantageous as the physical problem becomes more and more demanding. For the case of multimode light propagating through an inhomogeneously broadened medium of many-level quantum systems, the program executing on GPUs can be over 20 times faster than that executing on all cores of a modern CPU. The methods presented can be applied in a wide area of atomic physics where the time evolution of atomic ensembles is to be computed.

KW - CUDA

KW - GPU

KW - Maxwell-Bloch equations

KW - Pulse propagation

KW - Quantum ensembles

KW - Resonant nonlinear optics

UR - http://www.scopus.com/inward/record.url?scp=84873195096&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=84873195096&partnerID=8YFLogxK

U2 - 10.1016/j.cpc.2012.12.019

DO - 10.1016/j.cpc.2012.12.019

M3 - Article

VL - 184

SP - 1203

EP - 1210

JO - Computer Physics Communications

JF - Computer Physics Communications

SN - 0010-4655

IS - 4

ER -