Convolutional neural networks on dataflow engines

Conference paper (2017)

Authors

Nils Voss Imperial College London, Maxeler Technologies

Marco Bacis Politecnico di Milano, Maxeler Technologies

Oskar Mencer Maxeler Technologies

G. Gaydadjiev Imperial College London, Maxeler Technologies

Wayne Luk Imperial College London

Affiliation

External organisation

Deep Learning CNN Inference FPGA DFE DSE

To reference this document use:

http://resolver.tudelft.nl/uuid:10f15013-5c51-4720-9849-f0e578e6afb5

More Info

expand_more

Published Date

22-11-2017

Language

English

Reuse Rights

Other than for strictly personal use, it is not permitted to download, forward or distribute the text or part of it, without the consent of the author(s) and/or copyright holder(s), unless the work is under an open content license such as Creative Commons.

Affiliation

External organisation

Abstract

In this paper we discuss a high performance implementation for Convolutional Neural Networks (CNNs) inference on the latest generation of Dataflow Engines (DFEs). We discuss the architectural choices made during the design phase taking into account the DFE chip properties. We then perform design space exploration, considering the memory bandwidth and resources utilisation constraints derived from the used DFE and the chosen architecture. Finally, we discuss the high performance implementation and compare the obtained performance against other implementations, showing that our proposed design reaches 2,450 GOPS when running VGG16 as a test case.