Multiscale Co-Design Analysis of Energy, Latency, Area, and Accuracy of a ReRAM Analog Neural Training Accelerator

Abstract

Neural networks are an increasingly attractive algorithm for natural language processing and pattern recognition. Deep networks with >50M parameters are made possible by modern GPU clusters operating at <50 pJ per op and more recently, production accelerators capable of <5pJ per operation at the board level. However, with the slowing of CMOS scaling, new paradigms will be required to achieve the next several orders of magnitude in performance per watt gains. Using an analog resistive memory (ReRAM) crossbar to perform key matrix operations in an accelerator is an attractive option. This work presents a detailed design using a state of the art 14/16 nm PDK for of an analog crossbar circuit block designed to process three key kernels required in training and inference of neural networks. A detailed circuit and device-level analysis of energy, latency, area, and accuracy are given and compared to relevant designs using standard digital ReRAM and SRAM operations. It is shown that the analog accelerator has a 270x energy and 540x latency advantage over a similar block utilizing only digital ReRAM and takes only 11 fJ per multiply and accumulate (MAC). Compared to an SRAM based accelerator, the energy is 430X better and latency is 34X better. Although training accuracy is degraded in the analog accelerator, several options to improve this are presented. The possible gains over a similar digital-only version of this accelerator block suggest that continued optimization of analog resistive memories is valuable. This detailed circuit and device analysis of a training accelerator may serve as a foundation for further architecture-level studies.

Cited by in corpus (18)

• Brain-inspired computing: We need a master plan
• Prospects and applications of photonic neural networks
• Hafnia-based Double Layer Ferroelectric Tunnel Junctions as Artificial Synapses for Neuromorphic Computing
• Mixed-precision deep learning based on computational memory
• On the Accuracy of Analog Neural Network Inference Accelerators
• RAELLA: Reforming the Arithmetic for Efficient, Low-Resolution, and Low-Loss Analog PIM: No Retraining Required!
• Nonvolatile Electrochemical Random-Access Memory Under Short Circuit
• Harnessing Intrinsic Noise in Memristor Hopfield Neural Networks for Combinatorial Optimization
• Solving Boolean satisfiability problems with resistive content addressable memories
• Examining the Role and Limits of Batchnorm Optimization to Mitigate Diverse Hardware-noise in In-memory Computing
• FORMS: Fine-grained Polarized ReRAM-based In-situ Computation for Mixed-signal DNN Accelerator
• Quantized Non-Volatile Nanomagnetic Synapse based Autoencoder for Efficient Unsupervised Network Anomaly Detection
• Device Modeling Bias in ReRAM-based Neural Network Simulations
• Amplifier-Enhanced Memristive Massive MIMO Linear Detector Circuit: An Ultra-Energy-Efficient and Robust-to-Conductance-Error Design
• Benchmarking Inference Performance of Deep Learning Models on Analog Devices
• In-Memory Massive MIMO Linear Detector Circuit with Extremely High Energy Efficiency and Strong Memristive Conductance Deviation Robustness
• Benchmarking Physical Performance of Neural Inference Circuits
• DetectX -- Adversarial Input Detection using Current Signatures in Memristive XBar Arrays