API-Based Hardware Fault Simulation for DNN Accelerators

Patrik Omland; Yang Peng; Michael Paulitsch; Jorge Parra; Gustavo Espinosa; Abishai Daniel; Gereon Hinz; Alois Knoll

doi:10.1109/MDAT.2022.3180977

API-Based Hardware Fault Simulation for DNN Accelerators

Patrik Omland, Yang Peng, Michael Paulitsch, Jorge Parra, Gustavo Espinosa, Abishai Daniel, Gereon Hinz, Alois Knoll

nformatics 6 - Chair of IRobotics, Artificial Intelligence and Real-time Systems

Research output: Contribution to journal › Article › peer-review

2 Scopus citations

Abstract

Editor's notes: This article presents an application program interface (API)-based hardware fault simulation method to investigate the effect of hardware faults on the failure probability of deep neural network (DNN) accelerators.

Original language	English
Pages (from-to)	75-81
Number of pages	7
Journal	IEEE Design and Test
Volume	40
Issue number	2
DOIs	https://doi.org/10.1109/MDAT.2022.3180977
State	Published - 1 Apr 2023

Keywords

deep neural networks
fault injection
fault model
fault tolerance
hardware faults
hardware reliability
program vulnerability

Access to Document

10.1109/MDAT.2022.3180977

Cite this

@article{a1e5cf8d1bc34cea888e53068dfaff2b,

title = "API-Based Hardware Fault Simulation for DNN Accelerators",

abstract = "Editor's notes: This article presents an application program interface (API)-based hardware fault simulation method to investigate the effect of hardware faults on the failure probability of deep neural network (DNN) accelerators.",

keywords = "deep neural networks, fault injection, fault model, fault tolerance, hardware faults, hardware reliability, program vulnerability",

author = "Patrik Omland and Yang Peng and Michael Paulitsch and Jorge Parra and Gustavo Espinosa and Abishai Daniel and Gereon Hinz and Alois Knoll",

note = "Publisher Copyright: {\textcopyright} 2013 IEEE.",

year = "2023",

month = apr,

day = "1",

doi = "10.1109/MDAT.2022.3180977",

language = "English",

volume = "40",

pages = "75--81",

journal = "IEEE Design and Test",

issn = "2168-2356",

publisher = "IEEE Computer Society",

number = "2",

}

TY - JOUR

T1 - API-Based Hardware Fault Simulation for DNN Accelerators

AU - Omland, Patrik

AU - Peng, Yang

AU - Paulitsch, Michael

AU - Parra, Jorge

AU - Espinosa, Gustavo

AU - Daniel, Abishai

AU - Hinz, Gereon

AU - Knoll, Alois

PY - 2023/4/1

Y1 - 2023/4/1

N2 - Editor's notes: This article presents an application program interface (API)-based hardware fault simulation method to investigate the effect of hardware faults on the failure probability of deep neural network (DNN) accelerators.

AB - Editor's notes: This article presents an application program interface (API)-based hardware fault simulation method to investigate the effect of hardware faults on the failure probability of deep neural network (DNN) accelerators.

KW - deep neural networks

KW - fault injection

KW - fault model

KW - fault tolerance

KW - hardware faults

KW - hardware reliability

KW - program vulnerability

UR - http://www.scopus.com/inward/record.url?scp=85131788932&partnerID=8YFLogxK

U2 - 10.1109/MDAT.2022.3180977

DO - 10.1109/MDAT.2022.3180977

M3 - Article

AN - SCOPUS:85131788932

SN - 2168-2356

VL - 40

SP - 75

EP - 81

JO - IEEE Design and Test

JF - IEEE Design and Test

IS - 2

ER -

API-Based Hardware Fault Simulation for DNN Accelerators

Abstract

Keywords

Access to Document

Other files and links

Fingerprint

Cite this