PiC-BNN: A 128-kbit 65 nm Processing-in-CAM-Based End-to-End Binary Neural Network Accelerator

TL;DR

PiC-BNN is an end-to-end binary neural network accelerator built in 65nm technology, utilizing Hamming distance tolerance in content addressable memory to achieve high accuracy, throughput, and power efficiency without full precision operations.

Contribution

It introduces PiC-BNN, a novel end-to-end binary neural network accelerator that leverages Hamming distance tolerance in CAM to eliminate the need for full precision layers.

Findings

Achieves 95.2% accuracy on MNIST

Delivers 560K inferences/sec throughput

Provides 703M inferences/sec/W power efficiency

Abstract

Binary Neural Networks (BNNs), where weights and activations are constrained to binary values (+1, -1), are a highly efficient alternative to traditional neural networks. Unfortunately, typical BNNs, while binarizing linear layers (matrix-vector multiplication), still implement other network layers (batch normalization, softmax, output layer, and sometimes the input layer of a convolutional neural network) in full precision. This limits the area and energy benefits and requires architectural support for full precision operations. We propose PiC-BNN, a true end-to-end binary in-approximate search (Hamming distance tolerant) Content Addressable Memory based BNN accelerator. PiC-BNN is designed and manufactured in a commercial 65nm process. PiC-BNN uses Hamming distance tolerance to apply the law of large numbers to enable accurate classification without implementing full precision operations. PiC-BNN achieves baseline software accuracy (95.2%) on the MNIST dataset and 93.5% on the Hand Gesture (HG) dataset, a throughput of 560K inferences/s, and presents a power efficiency of 703M inferences/s/W when implementing a binary MLP model for MNIST/HG dataset classification.