Header

UZH-Logo

Maintenance Infos

NullHop: A Flexible Convolutional Neural Network Accelerator Based on Sparse Representations of Feature Maps


Aimar, Alessandro; Mostafa, Hesham; Calabrese, Enrico; Rios-Navarro, Antonio; Tapiador-Morales, Ricardo; Lungu, Iulia-Alexandra; Milde, Moritz B; Corradi, Federico; Linares-Barranco, Alejandro; Liu, Shih-Chii; Delbruck, Tobi (2017). NullHop: A Flexible Convolutional Neural Network Accelerator Based on Sparse Representations of Feature Maps. arXiv.org 1706.01406, Institute of Neuroinformatics.

Abstract

Convolutional neural networks (CNNs) have become the dominant neural network architecture for solving many state-of-the-art (SOA) visual processing tasks. Even though graphical processing units are most often used in training and deploying CNNs, their power efficiency is less than 10 GOp/s/W for single-frame runtime inference. We propose a flexible and efficient CNN accelerator architecture called NullHop that implements SOA CNNs useful for low-power and low-latency application scenarios. NullHop exploits the sparsity of neuron activations in CNNs to accelerate the computation and reduce memory requirements. The flexible architecture allows high utilization of available computing resources across kernel sizes ranging from 1×1 to 7×7 . NullHop can process up to 128 input and 128 output feature maps per layer in a single pass. We implemented the proposed architecture on a Xilinx Zynq field-programmable gate array (FPGA) platform and presented the results showing how our implementation reduces external memory transfers and compute time in five different CNNs ranging from small ones up to the widely known large VGG16 and VGG19 CNNs. Postsynthesis simulations using Mentor Modelsim in a 28-nm process with a clock frequency of 500 MHz show that the VGG19 network achieves over 450 GOp/s. By exploiting sparsity, NullHop achieves an efficiency of 368%, maintains over 98% utilization of the multiply–accumulate units, and achieves a power efficiency of over 3 TOp/s/W in a core area of 6.3 mm 2 . As further proof of NullHop’s usability, we interfaced its FPGA implementation with a neuromorphic event camera for real-time interactive demonstrations.

Abstract

Convolutional neural networks (CNNs) have become the dominant neural network architecture for solving many state-of-the-art (SOA) visual processing tasks. Even though graphical processing units are most often used in training and deploying CNNs, their power efficiency is less than 10 GOp/s/W for single-frame runtime inference. We propose a flexible and efficient CNN accelerator architecture called NullHop that implements SOA CNNs useful for low-power and low-latency application scenarios. NullHop exploits the sparsity of neuron activations in CNNs to accelerate the computation and reduce memory requirements. The flexible architecture allows high utilization of available computing resources across kernel sizes ranging from 1×1 to 7×7 . NullHop can process up to 128 input and 128 output feature maps per layer in a single pass. We implemented the proposed architecture on a Xilinx Zynq field-programmable gate array (FPGA) platform and presented the results showing how our implementation reduces external memory transfers and compute time in five different CNNs ranging from small ones up to the widely known large VGG16 and VGG19 CNNs. Postsynthesis simulations using Mentor Modelsim in a 28-nm process with a clock frequency of 500 MHz show that the VGG19 network achieves over 450 GOp/s. By exploiting sparsity, NullHop achieves an efficiency of 368%, maintains over 98% utilization of the multiply–accumulate units, and achieves a power efficiency of over 3 TOp/s/W in a core area of 6.3 mm 2 . As further proof of NullHop’s usability, we interfaced its FPGA implementation with a neuromorphic event camera for real-time interactive demonstrations.

Statistics

Downloads

46 downloads since deposited on 01 Mar 2018
21 downloads since 12 months
Detailed statistics

Additional indexing

Item Type:Working Paper
Communities & Collections:07 Faculty of Science > Institute of Neuroinformatics
Dewey Decimal Classification:570 Life sciences; biology
Language:English
Date:2017
Deposited On:01 Mar 2018 11:15
Last Modified:30 Jan 2020 09:40
Series Name:arXiv.org
ISSN:2331-8422
OA Status:Green
Free access at:Official URL. An embargo period may apply.
Official URL:https://arxiv.org/abs/1706.01406

Download

Green Open Access

Download PDF  'NullHop: A Flexible Convolutional Neural Network Accelerator Based on Sparse Representations of Feature Maps'.
Preview
Content: Published Version
Filetype: PDF
Size: 1MB