Correcting Soft Errors Online in Fast Fourier Transform

Xin Liang; Jieyang Chen; Dingwen Tao; Sihuan Li; Panruo Wu; Hongbo Li; Kaiming Ouyang; Yuanlai Liu; Fengguang Song; Zizhong Chen

doi:10.1145/3126908.3126915

Correcting Soft Errors Online in Fast Fourier Transform

Xin Liang
, Jieyang Chen
, Dingwen Tao
, Sihuan Li
, Panruo Wu
, Hongbo Li
, Kaiming Ouyang
, Yuanlai Liu
, Fengguang Song
, Zizhong Chen

Visualization

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution › peer-review

15 Scopus citations

Abstract

While many algorithm-based fault tolerance (ABFT) schemes have been proposed to detect soft errors offline in the fast Fourier transform (FFT) after computation finishes, none of the existing ABFT schemes detect soft errors online before the computation finishes. This paper presents an online ABFT scheme for FFT so that soft errors can be detected online and the corrupted computation can be terminated in a much more timely manner. We also extend our scheme to tolerate both arithmetic errors and memory errors, develop strategies to reduce its fault tolerance overhead and improve its numerical stability and fault coverage, and finally incorporate it into the widely used FFTW library - one of the today's fastest FFT software implementations. Experimental results demonstrate that: (1) the proposed online ABFT scheme introduces much lower overhead than the existing offline ABFT schemes; (2) it detects errors in a much more timely manner; and (3) it also has higher numerical stability and better fault coverage.

Original language	English
Title of host publication	SC 2017 - International Conference for High Performance Computing, Networking, Storage and Analysis
Publisher	IEEE Computer Society
ISBN (Electronic)	9781450351140
DOIs	https://doi.org/10.1145/3126908.3126915
State	Published - 2017
Event	2017 International Conference for High Performance Computing, Networking, Storage and Analysis, SC 2017 - Denver, United States Duration: Nov 12 2017 → Nov 17 2017

Publication series

Name	International Conference for High Performance Computing, Networking, Storage and Analysis, SC
Volume	2017-November
ISSN (Print)	2167-4329
ISSN (Electronic)	2167-4337

Conference

Conference	2017 International Conference for High Performance Computing, Networking, Storage and Analysis, SC 2017
Country/Territory	United States
City	Denver
Period	11/12/17 → 11/17/17

Funding

Œis work is partially supported by the NSF grants OAC-1305624, CCF-1513201, the SZSTI basic research program JCYJ2015063011494-2313, and the MOST key project 2017YFB0202100.

Keywords

Algorithm-Based Fault Tolerance
DFT
FFT
FFTW
Soft Errors

Access to Document

10.1145/3126908.3126915

Cite this

Liang, X., Chen, J., Tao, D., Li, S., Wu, P., Li, H., Ouyang, K., Liu, Y., Song, F., & Chen, Z. (2017). Correcting Soft Errors Online in Fast Fourier Transform. In SC 2017 - International Conference for High Performance Computing, Networking, Storage and Analysis (International Conference for High Performance Computing, Networking, Storage and Analysis, SC; Vol. 2017-November). IEEE Computer Society. https://doi.org/10.1145/3126908.3126915

@inproceedings{9d87375d85b84fb79b10eb69e77bd820,

title = "Correcting Soft Errors Online in Fast Fourier Transform",

abstract = "While many algorithm-based fault tolerance (ABFT) schemes have been proposed to detect soft errors offline in the fast Fourier transform (FFT) after computation finishes, none of the existing ABFT schemes detect soft errors online before the computation finishes. This paper presents an online ABFT scheme for FFT so that soft errors can be detected online and the corrupted computation can be terminated in a much more timely manner. We also extend our scheme to tolerate both arithmetic errors and memory errors, develop strategies to reduce its fault tolerance overhead and improve its numerical stability and fault coverage, and finally incorporate it into the widely used FFTW library - one of the today's fastest FFT software implementations. Experimental results demonstrate that: (1) the proposed online ABFT scheme introduces much lower overhead than the existing offline ABFT schemes; (2) it detects errors in a much more timely manner; and (3) it also has higher numerical stability and better fault coverage.",

keywords = "Algorithm-Based Fault Tolerance, DFT, FFT, FFTW, Soft Errors",

author = "Xin Liang and Jieyang Chen and Dingwen Tao and Sihuan Li and Panruo Wu and Hongbo Li and Kaiming Ouyang and Yuanlai Liu and Fengguang Song and Zizhong Chen",

note = "Publisher Copyright: {\textcopyright} 2017 ACM.; 2017 International Conference for High Performance Computing, Networking, Storage and Analysis, SC 2017 ; Conference date: 12-11-2017 Through 17-11-2017",

year = "2017",

doi = "10.1145/3126908.3126915",

language = "English",

series = "International Conference for High Performance Computing, Networking, Storage and Analysis, SC",

publisher = "IEEE Computer Society",

booktitle = "SC 2017 - International Conference for High Performance Computing, Networking, Storage and Analysis",

}

Liang, X, Chen, J, Tao, D, Li, S, Wu, P, Li, H, Ouyang, K, Liu, Y, Song, F & Chen, Z 2017, Correcting Soft Errors Online in Fast Fourier Transform. in SC 2017 - International Conference for High Performance Computing, Networking, Storage and Analysis. International Conference for High Performance Computing, Networking, Storage and Analysis, SC, vol. 2017-November, IEEE Computer Society, 2017 International Conference for High Performance Computing, Networking, Storage and Analysis, SC 2017, Denver, United States, 11/12/17. https://doi.org/10.1145/3126908.3126915

Correcting Soft Errors Online in Fast Fourier Transform. / Liang, Xin; Chen, Jieyang; Tao, Dingwen et al.
SC 2017 - International Conference for High Performance Computing, Networking, Storage and Analysis. IEEE Computer Society, 2017. (International Conference for High Performance Computing, Networking, Storage and Analysis, SC; Vol. 2017-November).

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution › peer-review

TY - GEN

T1 - Correcting Soft Errors Online in Fast Fourier Transform

AU - Liang, Xin

AU - Chen, Jieyang

AU - Tao, Dingwen

AU - Li, Sihuan

AU - Wu, Panruo

AU - Li, Hongbo

AU - Ouyang, Kaiming

AU - Liu, Yuanlai

AU - Song, Fengguang

AU - Chen, Zizhong

PY - 2017

Y1 - 2017

N2 - While many algorithm-based fault tolerance (ABFT) schemes have been proposed to detect soft errors offline in the fast Fourier transform (FFT) after computation finishes, none of the existing ABFT schemes detect soft errors online before the computation finishes. This paper presents an online ABFT scheme for FFT so that soft errors can be detected online and the corrupted computation can be terminated in a much more timely manner. We also extend our scheme to tolerate both arithmetic errors and memory errors, develop strategies to reduce its fault tolerance overhead and improve its numerical stability and fault coverage, and finally incorporate it into the widely used FFTW library - one of the today's fastest FFT software implementations. Experimental results demonstrate that: (1) the proposed online ABFT scheme introduces much lower overhead than the existing offline ABFT schemes; (2) it detects errors in a much more timely manner; and (3) it also has higher numerical stability and better fault coverage.

AB - While many algorithm-based fault tolerance (ABFT) schemes have been proposed to detect soft errors offline in the fast Fourier transform (FFT) after computation finishes, none of the existing ABFT schemes detect soft errors online before the computation finishes. This paper presents an online ABFT scheme for FFT so that soft errors can be detected online and the corrupted computation can be terminated in a much more timely manner. We also extend our scheme to tolerate both arithmetic errors and memory errors, develop strategies to reduce its fault tolerance overhead and improve its numerical stability and fault coverage, and finally incorporate it into the widely used FFTW library - one of the today's fastest FFT software implementations. Experimental results demonstrate that: (1) the proposed online ABFT scheme introduces much lower overhead than the existing offline ABFT schemes; (2) it detects errors in a much more timely manner; and (3) it also has higher numerical stability and better fault coverage.

KW - Algorithm-Based Fault Tolerance

KW - DFT

KW - FFT

KW - FFTW

KW - Soft Errors

UR - https://www.scopus.com/pages/publications/85040181846

U2 - 10.1145/3126908.3126915

DO - 10.1145/3126908.3126915

M3 - Conference contribution

AN - SCOPUS:85040181846

T3 - International Conference for High Performance Computing, Networking, Storage and Analysis, SC

BT - SC 2017 - International Conference for High Performance Computing, Networking, Storage and Analysis

PB - IEEE Computer Society

T2 - 2017 International Conference for High Performance Computing, Networking, Storage and Analysis, SC 2017

Y2 - 12 November 2017 through 17 November 2017

ER -

Correcting Soft Errors Online in Fast Fourier Transform

Abstract

Publication series

Conference

Funding

Keywords

Access to Document

Other files and links

Fingerprint

Cite this