# Analyzing and Mitigating the Impact of Permanent Faults on a Systolic Array Based Neural Network Accelerator

arXiv: https://arxiv.org/abs/1802.04657 · HTML: https://mlsanity.com/paper/1802.04657/

## Key facts

- Cited by 25 papers in the map (data through 14 Oct 2026).
- First posted on arXiv 11 Feb 2018.

## Built on by (most cited first, 25 of 25)

- [Robust Machine Learning Systems: Challenges, Current Trends, Perspectives, and the Road Ahead](https://mlsanity.com/paper/2101.02559/) (2021)
- [A Systematic Literature Review on Hardware Reliability Assessment Methods for Deep Neural Networks](https://mlsanity.com/paper/2305.05750/) (2023)
- [ReSpawn: Energy-Efficient Fault-Tolerance for Spiking Neural Networks considering Unreliable Memories](https://mlsanity.com/paper/2108.10271/) (2021)
- [Towards Energy-Efficient and Secure Edge AI: A Cross-Layer Framework](https://mlsanity.com/paper/2109.09829/) (2021)
- [Understanding Silent Data Corruption in LLM Training](https://mlsanity.com/paper/2502.12340/) (2025)
- [DeepAxe: A Framework for Exploration of Approximation and Reliability Trade-offs in DNN Accelerators](https://mlsanity.com/paper/2303.08226/) (2023)
- [Exploring Fault-Energy Trade-offs in Approximate DNN Hardware Accelerators](https://mlsanity.com/paper/2101.02860/) (2021)
- [R2F: A Remote Retraining Framework for AIoT Processors with Computing Errors](https://mlsanity.com/paper/2107.03096/) (2021)
- [Special Session: Towards an Agile Design Methodology for Efficient, Reliable, and Secure ML Systems](https://mlsanity.com/paper/2204.09514/) (2022)
- [RescueSNN: Enabling Reliable Executions on Spiking Neural Network Accelerators under Permanent Faults](https://mlsanity.com/paper/2304.04041/) (2023)
- [Fault-Tolerant Deep Learning: A Hierarchical Perspective](https://mlsanity.com/paper/2204.01942/) (2022)
- [Resilience of Deep Learning applications: a systematic literature review of analysis and hardening techniques](https://mlsanity.com/paper/2309.16733/) (2023)
- [MRFI: An Open Source Multi-Resolution Fault Injection Framework for Neural Network Processing](https://mlsanity.com/paper/2306.11758/) (2023)
- [Cost-Effective Fault Tolerance for CNNs Using Parameter Vulnerability Based Hardening and Pruning](https://mlsanity.com/paper/2405.10658/) (2024)
- [Statistical Modeling of Soft Error Influence on Neural Networks](https://mlsanity.com/paper/2210.05876/) (2022)
- [FAQ: Mitigating the Impact of Faults in the Weight Memory of DNN Accelerators through Fault-Aware Quantization](https://mlsanity.com/paper/2305.12590/) (2023)
- [eFAT: Improving the Effectiveness of Fault-Aware Training for Mitigating Permanent Faults in DNN Hardware Accelerators](https://mlsanity.com/paper/2304.12949/) (2023)
- [Evaluating and Enhancing Robustness of Deep Recommendation Systems Against Hardware Errors](https://mlsanity.com/paper/2307.10244/) (2023)
- [Periodic Online Testing for Sparse Systolic Tensor Arrays](https://mlsanity.com/paper/2504.18628/) (2025)
- [FLARE: One-Shot PE-Level Fault Localization in Systolic Arrays via Algebraic Test Vectors](https://mlsanity.com/paper/2605.08594/) (2026)
- [Monitor Placement for Fault Localization in Deep Neural Network Accelerators](https://mlsanity.com/paper/2311.16594/) (2023)
- [Cross-Layer Optimization for Fault-Tolerant Deep Learning](https://mlsanity.com/paper/2312.13754/) (2023)
- [Repair Brain Damage: Real-Numbered Error Correction Code for Neural Network](https://mlsanity.com/paper/2602.00076/) (2026)
- [Long-Term and Short-Term Transistor Aging in Deep Neural Networks: Impact and Mitigation](https://mlsanity.com/paper/2606.04266/) (2026)
- [TERMon: Detecting Persistent Behavioral Threats in Edge AI via Hardware-Native Ternary Runtime Monitor](https://mlsanity.com/paper/2609.21713/) (2026)

## Abstract (arXiv)

Due to their growing popularity and computational cost, deep neural networks (DNNs) are being targeted for hardware acceleration. A popular architecture for DNN acceleration, adopted by the Google Tensor Processing Unit (TPU), utilizes a systolic array based matrix multiplication unit at its core. This paper deals with the design of fault-tolerant, systolic array based DNN accelerators for high defect rate technologies. To this end, we empirically show that the classification accuracy of a baseline TPU drops significantly even at extremely low fault rates (as low as 0.006%). We then propose two novel strategies, fault-aware pruning (FAP) and fault-aware pruning+retraining (FAP+T), that enable the TPU to operate at fault rates of up to 50%, with negligible drop in classification accuracy (as low as 0.1%) and no run-time performance overhead. The FAP+T does introduce a one-time retraining penalty per TPU chip before it is deployed, but we propose optimizations that reduce this one-time penalty to under 12 minutes. The penalty is then amortized over the entire lifetime of the TPU's operation.