Impact of NCFET on Neural Network Accelerators

This is the first work to investigate the impact that Negative Capacitance Field-Effect Transistor (NCFET) brings on the efficiency and accuracy of future Neural Networks (NN). NCFET is at the forefront of emerging technologies, especially after it has become compatible with the existing fabrication process of CMOS. Neural Network inference accelerators are becoming ubiquitous in modern SoCs and there is an ever-increasing demand for tighter and tighter throughput constraints and lower energy consumption. To explore the benefits that NCFET brings to NN inference regarding frequency, energy, and accuracy, we investigate different configurations of the multiply-add (MADD) circuit, which is the core computational unit in any NN accelerator. We demonstrate that, compared to the baseline 7nm FinFET technology, its negative capacitance counterpart reduces the energy by 55%, without any frequency reduction. In addition, it enables leveraging higher computational precision, which results to a considerable improvement in the inference accuracy. Importantly, the achieved accuracy improvement comes also together with a significant energy reduction and without any loss in frequency.

[1]  Forrest N. Iandola,et al.  SqueezeNet: AlexNet-level accuracy with 50x fewer parameters and <1MB model size , 2016, ArXiv.

[2]  Arun V. Thathachary,et al.  A steep-slope transistor based on abrupt electronic phase transition , 2015, Nature Communications.

[3]  Matthew Dawber,et al.  Modern Physics of Ferroelectrics: Essential Background , 2007 .

[4]  Saurabh Sinha,et al.  ASAP7: A 7-nm finFET predictive process design kit , 2016, Microelectron. J..

[5]  Bo Chen,et al.  MnasNet: Platform-Aware Neural Architecture Search for Mobile , 2018, 2019 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR).

[6]  Mark Sandler,et al.  MobileNetV2: Inverted Residuals and Linear Bottlenecks , 2018, 2018 IEEE/CVF Conference on Computer Vision and Pattern Recognition.

[7]  Om Prakash,et al.  Impact of Variability on Processor Performance in Negative Capacitance FinFET Technology , 2020, IEEE Transactions on Circuits and Systems I: Regular Papers.

[8]  Iraklis Anagnostopoulos,et al.  NPU Thermal Management , 2020, IEEE Transactions on Computer-Aided Design of Integrated Circuits and Systems.

[9]  Jörg Henkel,et al.  Unveiling the Impact of IR-Drop on Performance Gain in NCFET-Based Processors , 2019, IEEE Transactions on Electron Devices.

[10]  Xiangyu Zhang,et al.  ShuffleNet: An Extremely Efficient Convolutional Neural Network for Mobile Devices , 2017, 2018 IEEE/CVF Conference on Computer Vision and Pattern Recognition.

[11]  Christoforos E. Kozyrakis,et al.  Convolution engine: balancing efficiency & flexibility in specialized computing , 2013, ISCA.

[12]  Nitin Chawla,et al.  14.1 A 2.9TOPS/W deep convolutional neural network SoC in FD-SOI 28nm for intelligent embedded systems , 2017, 2017 IEEE International Solid-State Circuits Conference (ISSCC).

[13]  Iraklis Anagnostopoulos,et al.  Weight-Oriented Approximation for Energy-Efficient Neural Network Inference Accelerators , 2020, IEEE Transactions on Circuits and Systems I: Regular Papers.

[14]  David A. Patterson,et al.  In-datacenter performance analysis of a tensor processing unit , 2017, 2017 ACM/IEEE 44th Annual International Symposium on Computer Architecture (ISCA).

[15]  Ninghui Sun,et al.  DianNao: a small-footprint high-throughput accelerator for ubiquitous machine-learning , 2014, ASPLOS.

[16]  Justin Salamon,et al.  A Dataset and Taxonomy for Urban Sound Research , 2014, ACM Multimedia.

[17]  Chenming Hu,et al.  Analysis and Compact Modeling of Negative Capacitance Transistor with High ON-Current and Negative Output Differential Resistance—Part II: Model Validation , 2016, IEEE Transactions on Electron Devices.

[18]  Jörg Henkel,et al.  Negative Capacitance Transistor to Address the Fundamental Limitations in Technology Scaling: Processor Performance , 2018, IEEE Access.

[19]  H. Amrouch,et al.  Impact of Interface Traps on Negative Capacitance Transistor: Device and Circuit Reliability , 2020, IEEE Journal of the Electron Devices Society.

[20]  Soheil Ghiasi,et al.  Ristretto: A Framework for Empirical Study of Resource-Efficient Inference in Convolutional Neural Networks , 2018, IEEE Transactions on Neural Networks and Learning Systems.

[21]  Bo Chen,et al.  Quantization and Training of Neural Networks for Efficient Integer-Arithmetic-Only Inference , 2017, 2018 IEEE/CVF Conference on Computer Vision and Pattern Recognition.

[22]  Jörg Henkel,et al.  Impact of Radiation on Negative Capacitance FinFET , 2020, 2020 IEEE International Reliability Physics Symposium (IRPS).

[23]  Song Jin-Ook,et al.  7.1 An 11.5TOPS/W 1024-MAC Butterfly Structure Dual-Core Sparsity-Aware Neural Processing Unit in 8nm Flagship Mobile SoC , 2019 .

[24]  P. Ajayan,et al.  A subthermionic tunnel field-effect transistor with an atomically thin channel , 2015, Nature.

[25]  Houqiang Li,et al.  Quantization Networks , 2019, 2019 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR).

[26]  Li Fei-Fei,et al.  ImageNet: A large-scale hierarchical image database , 2009, CVPR.

[27]  R. Sporer,et al.  14nm Ferroelectric FinFET technology with steep subthreshold slope for ultra low power applications , 2017, 2017 IEEE International Electron Devices Meeting (IEDM).

[28]  Jian Sun,et al.  Deep Residual Learning for Image Recognition , 2015, 2016 IEEE Conference on Computer Vision and Pattern Recognition (CVPR).

[29]  R. Cavin,et al.  Nanoelectronics: negative capacitance to the rescue? , 2008, Nature nanotechnology.

[30]  S. Datta,et al.  Use of negative capacitance to provide voltage amplification for low power nanoscale devices. , 2008, Nano letters.

[31]  Thomas Mikolajick,et al.  Incipient Ferroelectricity in Al‐Doped HfO2 Thin Films , 2012 .

[32]  Jae-Gon Lee,et al.  7.1 An 11.5TOPS/W 1024-MAC Butterfly Structure Dual-Core Sparsity-Aware Neural Processing Unit in 8nm Flagship Mobile SoC , 2019, 2019 IEEE International Solid- State Circuits Conference - (ISSCC).