Gated-CNN: Combating NBTI and HCI aging effects in on-chip activation memories of Convolutional Neural Network accelerators

Muñoz, Nicolás Landeros; Valero, Alejandro; Tejero, Rubén Gran; Zoni, Davide

doi:10.1016/j.sysarc.2022.102553

Cited by 4 publications

(3 citation statements)

References 45 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…The highlighted boxes are used to mitigate NBTI. Also, because of the vulnerability and continuous degrading of SRAM storage due to NBTI wear-out effects, the state-ofthe-art research of [155] is concentrated on aging mitigation of on-chip SRAM activation buffers of DNN accelerator mitigating the stress. It is aimed at minimizing both NBTI and HCI aging effects.…”

Section: ) Thermal Managementmentioning

confidence: 99%

See 1 more Smart Citation

Dependable DNN Accelerator for Safety-Critical Systems: A Review on the Aging Perspective

Moghaddasi,

Gorgin,

Lee

2023

IEEE Access

View full text Add to dashboard Cite

Nowadays, artificial intelligence (AI) and deep learning (DL) progressively adapt to various spheres of our lives. These disciplines contain safety-critical applications such as autonomous driving with a high risk of human injury in the case of malfunction, requiring a high promise of dependability. Even the dependability becomes more crucial as shrinking CMOS technology feature size worsens the resilience concerns due to factors like aging. This paper addresses the overarching dependability issue of advanced deep neural networks (DNN) accelerators from the aging perspective. Especially, a comprehensive survey and taxonomy of techniques used to evaluate and mitigate aging effects are introduced. We cover different aging effects like permanent faults, timing errors, and lifetime issues. We review research by the layer-wise approach and categorize several resilience classes to bring out major features. The concluding part of this review highlights the questions answered and several future research directions. This study is expected to benefit researchers in different areas of DNN deployment, especially the dependability of this emergent paradigm.

show abstract

Section: ) Thermal Managementmentioning

confidence: 99%

“…FIGURE 29. Gated-CNN main modules (modified from [155]). This article has been accepted for publication in IEEE Access.…”

Section: ) Voltage Adaptationmentioning

confidence: 99%

Dependable DNN Accelerator for Safety-Critical Systems: A Review on the Aging Perspective

Moghaddasi,

Gorgin,

Lee

2023

IEEE Access

View full text Add to dashboard Cite

show abstract

“…In this work, we extend the TensorFlow 2.5.0 simulation framework to model a CNN accelerator consisting of an inference processing matrix and on-chip storage. The on-chip storage consists of two input/output buffers for the storage of activations, and another buffer for weights, each of them with a storage capacity of 2 MiB [LVTZ22]. In order to study the impact of voltage underscaling in these buffers, we have assumed the faulty maps of two real FPGA platforms [ SSUCK18].…”

Section: Framework Overviewmentioning

confidence: 99%