UCNN: Exploiting Computational Reuse in Deep Neural Networks via Weight Repetition

by   Kartik Hegde, et al.

Convolutional Neural Networks (CNNs) have begun to permeate all corners of electronic society (from voice recognition to scene generation) due to their high accuracy and machine efficiency per operation. At their core, CNN computations are made up of multi-dimensional dot products between weight and input vectors. This paper studies how weight repetition ---when the same weight occurs multiple times in or across weight vectors--- can be exploited to save energy and improve performance during CNN inference. This generalizes a popular line of work to improve efficiency from CNN weight sparsity, as reducing computation due to repeated zero weights is a special case of reducing computation due to repeated weights. To exploit weight repetition, this paper proposes a new CNN accelerator called the Unique Weight CNN Accelerator (UCNN). UCNN uses weight repetition to reuse CNN sub-computations (e.g., dot products) and to reduce CNN model size when stored in off-chip DRAM ---both of which save energy. UCNN further improves performance by exploiting sparsity in weights. We evaluate UCNN with an accelerator-level cycle and energy model and with an RTL implementation of the UCNN processing element. On three contemporary CNNs, UCNN improves throughput-normalized energy consumption by 1.2x - 4x, relative to a similarly provisioned baseline accelerator that uses Eyeriss-style sparsity optimizations. At the same time, the UCNN processing element adds only 17-24 area overhead relative to the same baseline.


page 1

page 11


CREW: Computation Reuse and Efficient Weight Storage for Hardware-accelerated MLPs and RNNs

Deep Neural Networks (DNNs) have achieved tremendous success for cogniti...

SCNN: An Accelerator for Compressed-sparse Convolutional Neural Networks

Convolutional Neural Networks (CNNs) have emerged as a fundamental techn...

Sense: Model Hardware Co-design for Accelerating Sparse Neural Networks

Sparsity is an intrinsic property of neural network(NN). Many software r...

SIMCNN – Exploiting Computational Similarity to Accelerate CNN Training in Hardware

Convolution neural networks (CNN) are computation intensive to train. It...

Digital Neuron: A Hardware Inference Accelerator for Convolutional Deep Neural Networks

We propose a Digital Neuron, a hardware inference accelerator for convol...

IKW: Inter-Kernel Weights for Power Efficient Edge Computing

Deep Convolutional Neural Networks (CNN) have achieved state-of-the-art ...

CoDR: Computation and Data Reuse Aware CNN Accelerator

Computation and Data Reuse is critical for the resource-limited Convolut...

Please sign up or login with your details

Forgot password? Click here to reset