{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"indexed":{"date-parts":[[2026,8,11]],"date-time":"2026-08-11T15:46:54Z","timestamp":1786463214688,"version":"build-2736575974"},"reference-count":59,"publisher":"Association for Computing Machinery (ACM)","issue":"2","license":[{"start":{"date-parts":[[2025,2,7]],"date-time":"2025-02-07T00:00:00Z","timestamp":1738886400000},"content-version":"vor","delay-in-days":0,"URL":"https:\/\/creativecommons.org\/licenses\/by\/4.0\/"}],"content-domain":{"domain":["dl.acm.org"],"crossmark-restriction":true},"short-container-title":["ACM Trans. Des. Autom. Electron. Syst."],"published-print":{"date-parts":[[2025,3,31]]},"abstract":"<jats:p>Deep Neural Networks (DNNs) have permeated multiple applications, including cutting-edge safety-critical domains, which require relevant computational power, often provided by Graphic Processing Units (GPUs). GPUs are manufactured with advanced semiconductor technologies that can be affected by faults during the operational phase (e.g., due to wear-out, aging, or environmental harshness), whose effects possibly reach the DNN outputs, in some cases leading to catastrophic consequences. Hence, hardware-aware reliability assessments of DNNs are crucial to be considered in the context of safety-critical systems (following regulations\/standards of specific application domains). Application-level fault injection (FI) techniques (i.e., DNN parameter corruption) are often adopted for the reliability evaluation of DNNs; unfortunately, these approaches hardly represent fault effects from GPU hardware. This work proposes an FI strategy based on Hardware-Injection-Through-Program-Transformation (HITPT) to mimic the effect of permanent faults (PFs) at the GPU instruction level, enabling effective assessment of PFs on DNN\u2019s reliability. Our approach provides a good trade-off between the fault effect evaluation\u2019s accuracy and the required computational time. Using the proposed approach, for the first time, we systematically assessed the effects of PF in GPUs executing some DNN sample cases. The results indicate that the faults injected closer to the hardware, using our evaluation strategy, can produce a higher accuracy degradation than the evaluations performed by the typical application-level FI that modify only the DNN parameters. Furthermore, the proposed FI methodology provides insightful results to identify the most suitable fault-tolerance solutions (e.g., selective hardening or design diversity) for their application at thread levels inside GPU\u2019s kernels.<\/jats:p>","DOI":"10.1145\/3715327","type":"journal-article","created":{"date-parts":[[2025,1,25]],"date-time":"2025-01-25T11:44:02Z","timestamp":1737805442000},"page":"1-33","update-policy":"https:\/\/doi.org\/10.1145\/crossmark-policy","source":"Crossref","is-referenced-by-count":5,"title":["Effective Fault Effects Evaluation for Permanent Faults in GPUs executing DNNs"],"prefix":"10.1145","volume":"30","author":[{"ORCID":"https:\/\/orcid.org\/0000-0001-6852-2372","authenticated-orcid":false,"given":"Juan David","family":"Guerrero Balaguera","sequence":"first","affiliation":[{"name":"Control and Computer Engineering (DAUIN), Politecnico di Torino, Torino, Italy"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"ORCID":"https:\/\/orcid.org\/0000-0001-5957-5624","authenticated-orcid":false,"given":"Josie Esteban","family":"Rodriguez Condia","sequence":"additional","affiliation":[{"name":"Control and Computer Engineering (DAUIN), Politecnico di Torino, Torino, Italy"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"ORCID":"https:\/\/orcid.org\/0000-0003-2899-7669","authenticated-orcid":false,"given":"Matteo","family":"Sonza Reorda","sequence":"additional","affiliation":[{"name":"Control and Computer Engineering (DAUIN), Politecnico di Torino, Torino, Italy"}],"role":[{"vocabulary":"crossref","role":"author"}]}],"member":"320","published-online":{"date-parts":[[2025,2,7]]},"reference":[{"key":"e_1_3_2_2_2","doi-asserted-by":"crossref","unstructured":"M. Abdullah Hanif and M. Shafique. 2020. SalvageDNN: Salvaging deep neural network accelerators with permanent faults through saliency-driven fault-aware mapping. Philosophical Transactions of the Royal Society A 378 2164 (2020) 20190164.","DOI":"10.1098\/rsta.2019.0164"},{"key":"e_1_3_2_3_2","doi-asserted-by":"publisher","unstructured":"Mohammad Hasan Ahmadilivani Mahdi Taheri Jaan Raik Masoud Daneshtalab and Maksim Jenihhin. 2024. A systematic literature review on hardware reliability assessment methods for deep neural networks. ACM Comput. Surv. 56 6 Article 141 (Jan.2024) 39 pages. 10.1145\/3638242","DOI":"10.1145\/3638242"},{"key":"e_1_3_2_4_2","unstructured":"Advanced Micro Devices (AMD). 2024. Machine Learning Development on a Local Desktop with AMD Radeon\u2122 Graphics Cards. https:\/\/www.amd.com\/en\/developer\/resources\/ml-radeon.html. [Online; accessed 7-February-2024]."},{"key":"e_1_3_2_5_2","volume-title":"The 18th IEEE Workshop on Silicon Errors in Logic \u2013 System Effects","author":"Bacon David F.","year":"2022","unstructured":"David F. Bacon. 2022. Detection and prevention of silent data corruption in an exabyte-scale database system. In The 18th IEEE Workshop on Silicon Errors in Logic \u2013 System Effects."},{"key":"e_1_3_2_6_2","doi-asserted-by":"publisher","unstructured":"C. Bolchini L. Cassano and A. Miele. 2023. Resilience of Deep Learning applications: A systematic survey of analysis and hardening techniques. 10.48550\/arXiv.2309.16733arxiv:arXiv:2309.16733v2","DOI":"10.48550\/arXiv.2309.16733"},{"key":"e_1_3_2_7_2","doi-asserted-by":"publisher","DOI":"10.1109\/LATW.2019.8704548"},{"key":"e_1_3_2_8_2","doi-asserted-by":"publisher","DOI":"10.1007\/978-3-030-18419-3_16"},{"key":"e_1_3_2_9_2","doi-asserted-by":"publisher","unstructured":"Yiran Chen Yuan Xie Linghao Song Fan Chen and Tianqi Tang. 2020. A survey of accelerator architectures for deep neural networks. Engineering 6 3 (2020) 264\u2013274. 10.1016\/j.eng.2020.01.007","DOI":"10.1016\/j.eng.2020.01.007"},{"key":"e_1_3_2_10_2","doi-asserted-by":"publisher","DOI":"10.1109\/ISSRE5003.2020.00047"},{"key":"e_1_3_2_11_2","doi-asserted-by":"publisher","unstructured":"Josie E. Rodriguez Condia Boyang Du Matteo Sonza Reorda and Luca Sterpone. 2020. FlexGripPlus: An improved GPGPU model to support reliability analysis. Microelectronics Reliability 109 (2020) 113660. 10.1016\/j.microrel.2020.113660","DOI":"10.1016\/j.microrel.2020.113660"},{"key":"e_1_3_2_12_2","doi-asserted-by":"publisher","DOI":"10.1109\/ITC50671.2022.00036"},{"key":"e_1_3_2_13_2","unstructured":"NVIDIA Corporation. 2017. NVIDIA Tesla V100 GPU architecture. https:\/\/images.nvidia.com\/content\/volta-architecture\/pdf\/volta-architecture-whitepaper.pdf. [Online; accessed 27-August-2024]."},{"key":"e_1_3_2_14_2","unstructured":"NVIDIA Corporation. 2020. NVIDIA A100 Tensor Core GPU architecture. https:\/\/images.nvidia.com\/aem-dam\/en-zz\/Solutions\/data-center\/nvidia-ampere-architecture-whitepaper.pdf. [Online; accessed 7-August-2024]."},{"key":"e_1_3_2_15_2","unstructured":"NVIDIA Corporation. 2024. CUDA C++ Programming guide. https:\/\/docs.nvidia.com\/cuda\/cuda-c-programming-guide\/index.html. [Online; accessed 7-February-2024]."},{"key":"e_1_3_2_16_2","unstructured":"NVIDIA Corporation. 2024. Multiply-and-Accumulate instruction: MMA. https:\/\/docs.nvidia.com\/cuda\/parallel-thread-execution\/index.html#warp-level-matrix-instructions-mma. [Online; accessed 27-August-2024]."},{"key":"e_1_3_2_17_2","unstructured":"NVIDIA Corporation. 2024. NVIDIA Deep Learning TensorRT documentation. https:\/\/docs.nvidia.com\/deeplearning\/tensorrt\/. [Online; accessed 7-February-2024]."},{"key":"e_1_3_2_18_2","unstructured":"NVIDIA Corporation. 2024. Predicated esxecution. https:\/\/docs.nvidia.com\/cuda\/parallel-thread-execution\/index.html#predicated-execution. [Online; accessed 27-August-2024]."},{"key":"e_1_3_2_19_2","doi-asserted-by":"publisher","DOI":"10.1016\/B978-0-12-818130-0.00005-2"},{"key":"e_1_3_2_20_2","doi-asserted-by":"publisher","unstructured":"Corrado De Sio Sarah Azimi and Luca Sterpone. 2022. FireNN: Neural networks reliability evaluation on hybrid platforms. IEEE Transactions on Emerging Topics in Computing 10 2 (2022) 549\u2013563. 10.1109\/TETC.2022.3152668","DOI":"10.1109\/TETC.2022.3152668"},{"key":"e_1_3_2_21_2","doi-asserted-by":"publisher","unstructured":"Harish Dattatraya Dixit Sneha Pendharkar Matt Beadon Chris Mason Tejasvi Chakravarthy Bharath Muthiah and Sriram Sankar. 2021. Silent data corruptions at scale. (2021). 10.48550\/arXiv.2102.11245arxiv:arXiv:2102.11245","DOI":"10.48550\/arXiv.2102.11245"},{"key":"e_1_3_2_22_2","doi-asserted-by":"publisher","DOI":"10.1145\/3581784.3607086"},{"key":"e_1_3_2_23_2","doi-asserted-by":"publisher","DOI":"10.1109\/ITCAsia55616.2022.00027"},{"key":"e_1_3_2_24_2","doi-asserted-by":"publisher","DOI":"10.1109\/IOLTS56730.2022.9897823"},{"key":"e_1_3_2_25_2","doi-asserted-by":"publisher","DOI":"10.1109\/IRPS46558.2021.9405177"},{"key":"e_1_3_2_26_2","doi-asserted-by":"publisher","DOI":"10.1109\/ISPASS.2017.7975296"},{"key":"e_1_3_2_27_2","doi-asserted-by":"publisher","unstructured":"William Grant Hatcher and Wei Yu. 2018. A survey of deep learning: Platforms applications and emerging research trends. IEEE Access 6 (2018) 24411\u201324432. 10.1109\/ACCESS.2018.2830661","DOI":"10.1109\/ACCESS.2018.2830661"},{"key":"e_1_3_2_28_2","doi-asserted-by":"publisher","unstructured":"Younis Ibrahim Haibin Wang Man Bai Zhi Liu Jianan Wang Zhiming Yang and Zhengming Chen. 2020. Soft error resilience of deep residual networks for object recognition. IEEE Access 8 (2020) 19490\u201319503. 10.1109\/ACCESS.2020.2968129","DOI":"10.1109\/ACCESS.2020.2968129"},{"key":"e_1_3_2_29_2","volume-title":"Institute of Electrical and Electronics Engineers (IEEE)","year":"2022","unstructured":"IEEE. 2022. The International roadmap for devices and systems: 2022. In Institute of Electrical and Electronics Engineers (IEEE)."},{"key":"e_1_3_2_30_2","doi-asserted-by":"publisher","DOI":"10.1109\/VTS.2016.7477299"},{"key":"e_1_3_2_31_2","unstructured":"Lornatang. 2024. YOLOv3-PyTorch. https:\/\/github.com\/Lornatang\/YOLOv3-PyTorch. [Online; accessed 27-August-2024]."},{"key":"e_1_3_2_32_2","doi-asserted-by":"publisher","DOI":"10.1109\/ATS.2015.16"},{"key":"e_1_3_2_33_2","doi-asserted-by":"publisher","DOI":"10.1109\/DSN-W50199.2020.00014"},{"key":"e_1_3_2_34_2","doi-asserted-by":"publisher","DOI":"10.1109\/INFOCOMWKSHPS50562.2020.9162899"},{"key":"e_1_3_2_35_2","doi-asserted-by":"publisher","DOI":"10.1109\/ICOIN50884.2021.9333975"},{"key":"e_1_3_2_36_2","doi-asserted-by":"publisher","unstructured":"Anouar Nechi Lukas Groth Saleh Mulhem Farhad Merchant Rainer Buchty and Mladen Berekovic. 2023. FPGA-based deep learning inference accelerators: Where are we standing? ACM Trans. Reconfigurable Technol. Syst. 16 4 Article 60 (Oct.2023) 32 pages. 10.1145\/3613963","DOI":"10.1145\/3613963"},{"key":"e_1_3_2_37_2","doi-asserted-by":"publisher","unstructured":"Biagio Peccerillo Mirco Mannino Andrea Mondelli and Sandro Bartolini. 2022. A survey on hardware accelerators: Taxonomy trends challenges and perspectives. Journal of Systems Architecture 129 (2022) 102561. 10.1016\/j.sysarc.2022.102561","DOI":"10.1016\/j.sysarc.2022.102561"},{"key":"e_1_3_2_38_2","doi-asserted-by":"publisher","unstructured":"Ratheesh Ravindran Michael J. Santora and Mohsin M. Jamali. 2021. Multi-object detection and tracking based on DNN for autonomous vehicles: A review. IEEE Sensors Journal 21 5 (2021) 5668\u20135677. 10.1109\/JSEN.2020.3041615","DOI":"10.1109\/JSEN.2020.3041615"},{"key":"e_1_3_2_39_2","doi-asserted-by":"publisher","DOI":"10.1109\/DAC.2018.8465834"},{"key":"e_1_3_2_40_2","unstructured":"Joseph Redmon. 2013\u20132016. Darknet: Open Source Neural Networks in C. http:\/\/pjreddie.com\/darknet\/"},{"key":"e_1_3_2_41_2","doi-asserted-by":"publisher","DOI":"10.1109\/DFT50435.2020.9250866"},{"key":"e_1_3_2_42_2","doi-asserted-by":"publisher","DOI":"10.1109\/DSD51259.2020.00109"},{"key":"e_1_3_2_43_2","doi-asserted-by":"publisher","DOI":"10.23919\/DATE56975.2023.10136998"},{"key":"e_1_3_2_44_2","doi-asserted-by":"publisher","unstructured":"Annachiara Ruospo and Ernesto Sanchez. 2021. On the reliability assessment of artificial neural networks running on AI-oriented MPSoCs. Applied Sciences 11 14 (2021). 10.3390\/app11146455","DOI":"10.3390\/app11146455"},{"key":"e_1_3_2_45_2","doi-asserted-by":"publisher","unstructured":"Annachiara Ruospo Ernesto Sanchez Lucas Matana Luza Luigi Dilillo Marcello Traiola and Alberto Bosio. 2023. A survey on deep learning resilience assessment methodologies. Computer 56 2 (2023) 57\u201366. 10.1109\/MC.2022.3217841","DOI":"10.1109\/MC.2022.3217841"},{"key":"e_1_3_2_46_2","doi-asserted-by":"publisher","unstructured":"Annachiara Ruospo Ernesto Sanchez Marcello Traiola Ian O\u2019Connor and Alberto Bosio. 2021. Investigating data representation for efficient and reliable convolutional neural networks. Microprocessors and Microsystems 86 (2021) 104318. 10.1016\/j.micpro.2021.104318","DOI":"10.1016\/j.micpro.2021.104318"},{"key":"e_1_3_2_47_2","doi-asserted-by":"publisher","unstructured":"Mehdi Sadi and Ujjwal Guin. 2022. Test and yield loss reduction of AI and deep learning accelerators. IEEE Transactions on Computer-Aided Design of Integrated Circuits and Systems 41 1 (2022) 104\u2013115. 10.1109\/TCAD.2021.3051841","DOI":"10.1109\/TCAD.2021.3051841"},{"key":"e_1_3_2_48_2","doi-asserted-by":"publisher","DOI":"10.1109\/IPDPS49936.2021.00037"},{"key":"e_1_3_2_49_2","doi-asserted-by":"publisher","unstructured":"Fernando Fernandes dos Santos Angeliki Kritikakou Josie E. Rodriguez Condia Juan-David Guerrero-Balaguera Matteo Sonza Reorda Olivier Sentieys and Paolo Rech. 2023. Characterizing a neutron-induced fault model for deep neural networks. IEEE Transactions on Nuclear Science 70 4 (2023) 370\u2013380. 10.1109\/TNS.2022.3224538","DOI":"10.1109\/TNS.2022.3224538"},{"key":"e_1_3_2_50_2","doi-asserted-by":"publisher","unstructured":"Fernando Fernandes dos Santos Pedro Foletto Pimenta Caio Lunardi Lucas Draghetti Luigi Carro David Kaeli and Paolo Rech. 2019. Analyzing and increasing the reliability of convolutional neural networks on GPUs. IEEE Transactions on Reliability 68 2 (2019) 663\u2013677. 10.1109\/TR.2018.2878387","DOI":"10.1109\/TR.2018.2878387"},{"key":"e_1_3_2_51_2","doi-asserted-by":"publisher","DOI":"10.1109\/ISPASS55109.2022.00004"},{"key":"e_1_3_2_52_2","doi-asserted-by":"publisher","DOI":"10.23919\/DATE.2018.8342151"},{"key":"e_1_3_2_53_2","doi-asserted-by":"publisher","unstructured":"C. Silvano D. Ielmini F. Ferrandi L. Fiorin S. Curzel L. Benini F. Conti A. Garofalo C. Zambelli E. Caloreet al.2023. A survey on deep learning hardware accelerators for heterogeneous HPC platforms. 10.48550\/arXiv.2306.15552arxiv:arXiv:2306.15552","DOI":"10.48550\/arXiv.2306.15552"},{"key":"e_1_3_2_54_2","doi-asserted-by":"publisher","DOI":"10.1109\/VTS56346.2023.10139970"},{"key":"e_1_3_2_55_2","doi-asserted-by":"publisher","unstructured":"Fei Su Chunsheng Liu and Haralampos-G. Stratigopoulos. 2023. Testability and dependability of AI hardware: Survey trends challenges and perspectives. IEEE Design & Test 40 2 (2023) 8\u201358. 10.1109\/MDAT.2023.3241116","DOI":"10.1109\/MDAT.2023.3241116"},{"key":"e_1_3_2_56_2","doi-asserted-by":"publisher","unstructured":"Cesar Torres-Huitzil and Bernard Girau. 2017. Fault and error tolerance in neural networks: A review. IEEE Access 5 (2017) 17322\u201317341. 10.1109\/ACCESS.2017.2742698","DOI":"10.1109\/ACCESS.2017.2742698"},{"key":"e_1_3_2_57_2","doi-asserted-by":"publisher","DOI":"10.1109\/DSN48987.2021.00041"},{"key":"e_1_3_2_58_2","doi-asserted-by":"publisher","DOI":"10.1109\/IOLTS.2013.6604089"},{"key":"e_1_3_2_59_2","doi-asserted-by":"publisher","DOI":"10.1109\/VTS.2018.8368656"},{"key":"e_1_3_2_60_2","doi-asserted-by":"publisher","unstructured":"Rui Zhao Ruqiang Yan Zhenghua Chen Kezhi Mao Peng Wang and Robert X. Gao. 2019. Deep learning and its applications to machine health monitoring. Mechanical Systems and Signal Processing 115 (2019) 213\u2013237. 10.1016\/j.ymssp.2018.05.050","DOI":"10.1016\/j.ymssp.2018.05.050"}],"container-title":["ACM Transactions on Design Automation of Electronic Systems"],"original-title":[],"language":"en","link":[{"URL":"https:\/\/dl.acm.org\/doi\/10.1145\/3715327","content-type":"unspecified","content-version":"vor","intended-application":"text-mining"},{"URL":"https:\/\/dl.acm.org\/doi\/pdf\/10.1145\/3715327","content-type":"unspecified","content-version":"vor","intended-application":"similarity-checking"}],"deposited":{"date-parts":[[2025,6,19]],"date-time":"2025-06-19T01:18:18Z","timestamp":1750295898000},"score":1,"resource":{"primary":{"URL":"https:\/\/dl.acm.org\/doi\/10.1145\/3715327"}},"subtitle":[],"short-title":[],"issued":{"date-parts":[[2025,2,7]]},"references-count":59,"journal-issue":{"issue":"2","published-print":{"date-parts":[[2025,3,31]]}},"alternative-id":["10.1145\/3715327"],"URL":"https:\/\/doi.org\/10.1145\/3715327","relation":{},"ISSN":["1084-4309","1557-7309"],"issn-type":[{"value":"1084-4309","type":"print"},{"value":"1557-7309","type":"electronic"}],"subject":[],"published":{"date-parts":[[2025,2,7]]},"assertion":[{"value":"2024-09-19","order":0,"name":"received","label":"Received","group":{"name":"publication_history","label":"Publication History"}},{"value":"2025-01-16","order":2,"name":"accepted","label":"Accepted","group":{"name":"publication_history","label":"Publication History"}},{"value":"2025-02-07","order":3,"name":"published","label":"Published","group":{"name":"publication_history","label":"Publication History"}}]}}