Lupulus: A Flexible Hardware Accelerator for Neural Networks

Neural networks have become indispensable for a wide range of applications, but they suffer from high computational- and memory-requirements, requiring optimizations from the algorithmic description of the network to the hardware implementation. Moreover, the high rate of innovation in machine learning makes it important that hardware implementations provide a high level of programmability to support current and future requirements of neural networks. In this work, we present a flexible hardware accelerator for neural networks, called Lupulus, supporting various methods for scheduling and mapping of operations onto the accelerator. Lupulus was implemented in a 28nm FD-SOI technology and demonstrates a peak performance of 380 GOPS/GHz with latencies of 21.4ms and 183.6ms for the convolutional layers of AlexNet and VGG-16, respectively.

READ FULL TEXT

page 1

page 2

page 3

page 4

research
06/23/2020

hxtorch: PyTorch for ANNs on BrainScaleS-2

We present software facilitating the usage of the BrainScaleS-2 analog n...
research
12/16/2019

A flexible FPGA accelerator for convolutional neural networks

Though CNNs are highly parallel workloads, in the absence of efficient o...
research
09/11/2019

QuTiBench: Benchmarking Neural Networks on Heterogeneous Hardware

Neural Networks have become one of the most successful universal machine...
research
07/09/2018

XNOR Neural Engine: a Hardware Accelerator IP for 21.6 fJ/op Binary Neural Network Inference

Binary Neural Networks (BNNs) are promising to deliver accuracy comparab...
research
12/05/2021

Using Convolutional Neural Networks for fault analysis and alleviation in accelerator systems

Today, Neural Networks are the basis of breakthroughs in virtually every...
research
08/09/2021

Efficient Majority Voting in Digital Hardware

In recent years, machine learning methods became increasingly important ...
research
06/24/2021

A Construction Kit for Efficient Low Power Neural Network Accelerator Designs

Implementing embedded neural network processing at the edge requires eff...

Please sign up or login with your details

Forgot password? Click here to reset