FLoPAD-GRU: A Flexible, Low Power, Accelerated DSP for Gated Recurrent Unit Neural Network

Research output: Chapter in Book/Report/Conference proceedingPaper in conference proceedingpeer-review

Abstract

Recurrent neural networks (RNNs) are efficient for classification of sequential data such as speech and audio due to their high precision on tasks. However, power efficiency, the required memory capacity and bandwidth requirements make them less suitable for battery powered devices. In this work, we introduce FLoPAD-GRU: a system on a chip (SoC) for efficient processing of gated recurrent unit (GRU) networks, that consists of a digital signal processor (DSP), supplemented with an optimized hardware accelerator, which reduces memory accesses and cost. The system is programmable and scalable, which allows for execution of different network sizes. Synthesized in 28 nm CMOS technology, real-time classification is achieved at 4 MHz, with an energy dissipation of 4.1 pJ/classification, an improvement of 15 × compared to a pure DSP realization. The memory requirements are reduced by 75 %, which results in a silicon area of 0.7 mm2for the entire SoC.

Original languageEnglish
Title of host publicationProceedings - 34th SBC/SBMicro/IEEE/ACM Symposium on Integrated Circuits and Systems Design, SBCCI 2021
PublisherIEEE - Institute of Electrical and Electronics Engineers Inc.
ISBN (Electronic)9781665421706
DOIs
Publication statusPublished - 2021 Aug 23
Event34th SBC/SBMicro/IEEE/ACM Symposium on Integrated Circuits and Systems Design, SBCCI 2021 - Campinas, Brazil
Duration: 2021 Aug 232021 Aug 27

Publication series

NameProceedings - 34th SBC/SBMicro/IEEE/ACM Symposium on Integrated Circuits and Systems Design, SBCCI 2021

Conference

Conference34th SBC/SBMicro/IEEE/ACM Symposium on Integrated Circuits and Systems Design, SBCCI 2021
Country/TerritoryBrazil
CityCampinas
Period2021/08/232021/08/27

Subject classification (UKÄ)

  • Computer Science

Free keywords

  • Deep Learning
  • Digital Signal Processor
  • GRU
  • Hardware Accelerator
  • RNN
  • SoC
  • Speech Recognition

Fingerprint

Dive into the research topics of 'FLoPAD-GRU: A Flexible, Low Power, Accelerated DSP for Gated Recurrent Unit Neural Network'. Together they form a unique fingerprint.

Cite this