Large-scale algorithm design for parallel FFT-based simulations on GPUs

Anuva Kulkarni, Franz Franchetti, Jelena Kovacevic

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

Abstract

We describe and analyze a co-design of algorithm and software for high-performance simulation of a partial differential equation (PDE) numerical solver for large-scale datasets. Large-scale scientific simulations involving parallel Fast Fourier Transforms (FFTs) have extreme memory requirements and high communication cost. This hampers high resolution analysis with fine grids. Moreover, it is difficult to accelerate legacy Fortran scientific codes with modern hardware such as GPUs because of memory constraints of GPUs. Our proposed solution uses signal processing techniques such as lossy compression and domain-local FFTs to lower iteration cost without adversely impacting accuracy of the result. In this work, we discuss proof-of-concept results for various aspects of algorithm development.

Original languageEnglish (US)
Title of host publication2018 IEEE Global Conference on Signal and Information Processing, GlobalSIP 2018 - Proceedings
PublisherInstitute of Electrical and Electronics Engineers Inc.
Pages301-305
Number of pages5
ISBN (Electronic)9781728112954
DOIs
StatePublished - Feb 20 2019
Event2018 IEEE Global Conference on Signal and Information Processing, GlobalSIP 2018 - Anaheim, United States
Duration: Nov 26 2018Nov 29 2018

Publication series

Name2018 IEEE Global Conference on Signal and Information Processing, GlobalSIP 2018 - Proceedings

Conference

Conference2018 IEEE Global Conference on Signal and Information Processing, GlobalSIP 2018
CountryUnited States
CityAnaheim
Period11/26/1811/29/18

Fingerprint

Fast Fourier transforms
Data storage equipment
Computer hardware
Partial differential equations
Costs
Signal processing
Compaction
Communication
Graphics processing unit

Keywords

  • Algorithm design
  • GPU
  • Irregular domain decomposition
  • Lossy compression

ASJC Scopus subject areas

  • Information Systems
  • Signal Processing

Cite this

Kulkarni, A., Franchetti, F., & Kovacevic, J. (2019). Large-scale algorithm design for parallel FFT-based simulations on GPUs. In 2018 IEEE Global Conference on Signal and Information Processing, GlobalSIP 2018 - Proceedings (pp. 301-305). [8646675] (2018 IEEE Global Conference on Signal and Information Processing, GlobalSIP 2018 - Proceedings). Institute of Electrical and Electronics Engineers Inc.. https://doi.org/10.1109/GlobalSIP.2018.8646675

Large-scale algorithm design for parallel FFT-based simulations on GPUs. / Kulkarni, Anuva; Franchetti, Franz; Kovacevic, Jelena.

2018 IEEE Global Conference on Signal and Information Processing, GlobalSIP 2018 - Proceedings. Institute of Electrical and Electronics Engineers Inc., 2019. p. 301-305 8646675 (2018 IEEE Global Conference on Signal and Information Processing, GlobalSIP 2018 - Proceedings).

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

Kulkarni, A, Franchetti, F & Kovacevic, J 2019, Large-scale algorithm design for parallel FFT-based simulations on GPUs. in 2018 IEEE Global Conference on Signal and Information Processing, GlobalSIP 2018 - Proceedings., 8646675, 2018 IEEE Global Conference on Signal and Information Processing, GlobalSIP 2018 - Proceedings, Institute of Electrical and Electronics Engineers Inc., pp. 301-305, 2018 IEEE Global Conference on Signal and Information Processing, GlobalSIP 2018, Anaheim, United States, 11/26/18. https://doi.org/10.1109/GlobalSIP.2018.8646675
Kulkarni A, Franchetti F, Kovacevic J. Large-scale algorithm design for parallel FFT-based simulations on GPUs. In 2018 IEEE Global Conference on Signal and Information Processing, GlobalSIP 2018 - Proceedings. Institute of Electrical and Electronics Engineers Inc. 2019. p. 301-305. 8646675. (2018 IEEE Global Conference on Signal and Information Processing, GlobalSIP 2018 - Proceedings). https://doi.org/10.1109/GlobalSIP.2018.8646675
Kulkarni, Anuva ; Franchetti, Franz ; Kovacevic, Jelena. / Large-scale algorithm design for parallel FFT-based simulations on GPUs. 2018 IEEE Global Conference on Signal and Information Processing, GlobalSIP 2018 - Proceedings. Institute of Electrical and Electronics Engineers Inc., 2019. pp. 301-305 (2018 IEEE Global Conference on Signal and Information Processing, GlobalSIP 2018 - Proceedings).
@inproceedings{55b2d139171d40b4bdd84563c2fbe163,
title = "Large-scale algorithm design for parallel FFT-based simulations on GPUs",
abstract = "We describe and analyze a co-design of algorithm and software for high-performance simulation of a partial differential equation (PDE) numerical solver for large-scale datasets. Large-scale scientific simulations involving parallel Fast Fourier Transforms (FFTs) have extreme memory requirements and high communication cost. This hampers high resolution analysis with fine grids. Moreover, it is difficult to accelerate legacy Fortran scientific codes with modern hardware such as GPUs because of memory constraints of GPUs. Our proposed solution uses signal processing techniques such as lossy compression and domain-local FFTs to lower iteration cost without adversely impacting accuracy of the result. In this work, we discuss proof-of-concept results for various aspects of algorithm development.",
keywords = "Algorithm design, GPU, Irregular domain decomposition, Lossy compression",
author = "Anuva Kulkarni and Franz Franchetti and Jelena Kovacevic",
year = "2019",
month = "2",
day = "20",
doi = "10.1109/GlobalSIP.2018.8646675",
language = "English (US)",
series = "2018 IEEE Global Conference on Signal and Information Processing, GlobalSIP 2018 - Proceedings",
publisher = "Institute of Electrical and Electronics Engineers Inc.",
pages = "301--305",
booktitle = "2018 IEEE Global Conference on Signal and Information Processing, GlobalSIP 2018 - Proceedings",

}

TY - GEN

T1 - Large-scale algorithm design for parallel FFT-based simulations on GPUs

AU - Kulkarni, Anuva

AU - Franchetti, Franz

AU - Kovacevic, Jelena

PY - 2019/2/20

Y1 - 2019/2/20

N2 - We describe and analyze a co-design of algorithm and software for high-performance simulation of a partial differential equation (PDE) numerical solver for large-scale datasets. Large-scale scientific simulations involving parallel Fast Fourier Transforms (FFTs) have extreme memory requirements and high communication cost. This hampers high resolution analysis with fine grids. Moreover, it is difficult to accelerate legacy Fortran scientific codes with modern hardware such as GPUs because of memory constraints of GPUs. Our proposed solution uses signal processing techniques such as lossy compression and domain-local FFTs to lower iteration cost without adversely impacting accuracy of the result. In this work, we discuss proof-of-concept results for various aspects of algorithm development.

AB - We describe and analyze a co-design of algorithm and software for high-performance simulation of a partial differential equation (PDE) numerical solver for large-scale datasets. Large-scale scientific simulations involving parallel Fast Fourier Transforms (FFTs) have extreme memory requirements and high communication cost. This hampers high resolution analysis with fine grids. Moreover, it is difficult to accelerate legacy Fortran scientific codes with modern hardware such as GPUs because of memory constraints of GPUs. Our proposed solution uses signal processing techniques such as lossy compression and domain-local FFTs to lower iteration cost without adversely impacting accuracy of the result. In this work, we discuss proof-of-concept results for various aspects of algorithm development.

KW - Algorithm design

KW - GPU

KW - Irregular domain decomposition

KW - Lossy compression

UR - http://www.scopus.com/inward/record.url?scp=85063099856&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85063099856&partnerID=8YFLogxK

U2 - 10.1109/GlobalSIP.2018.8646675

DO - 10.1109/GlobalSIP.2018.8646675

M3 - Conference contribution

AN - SCOPUS:85063099856

T3 - 2018 IEEE Global Conference on Signal and Information Processing, GlobalSIP 2018 - Proceedings

SP - 301

EP - 305

BT - 2018 IEEE Global Conference on Signal and Information Processing, GlobalSIP 2018 - Proceedings

PB - Institute of Electrical and Electronics Engineers Inc.

ER -