skip to main content
10.1145/3194554.3194627acmconferencesArticle/Chapter ViewAbstractPublication PagesglsvlsiConference Proceedingsconference-collections
research-article

Dataflow-Based Mapping of Spiking Neural Networks on Neuromorphic Hardware

Published: 30 May 2018 Publication History

Abstract

Spiking Neural Networks (SNNs) are powerful computation engines for pattern recognition and image classification applications. Apart from application performance such as recognition and classification accuracy, system performance such as throughput becomes important when executing these applications on a hardware. We propose a systematic design-flow to map SNN-based applications on a crossbar-based neuromorphic hardware, guaranteeing application as well as system performance. Synchronous Dataflow Graphs (SDFGs) are used to model these applications with extended semantics to represent neural network topologies. Self-timed scheduling is then used to analyze throughput, incorporating hardware constraints such as synaptic memory, communication and I/O bandwidth of crossbars. Our design-flow integrates CARLsim, a GPU-accelerated application-level SNN simulator with SDF3, a tool for mapping SDFG on hardware. We conducted experiments with realistic and synthetic SNNs on representative neuromorphic hardware, demonstrating throughput-resource trade-offs for a given application performance. For throughput-constrained applications, we show average 20% reduction of hardware usage with 19% reduction in energy consumption. For throughput-scalable applications, we show an average 53% higher throughput compared to a state-of-the-art approach.

References

[1]
Marco Bacis et al. 2017. A pipelined and scalable dataflow implementation of convolutional neural networks on FPGA International Parallel and Distributed Processing Symposium Workshops (IPDPSW).
[2]
Michael Beyeler et al. 2015. CARLsim 3: A user-friendly and highly optimized library for the creation of neurobiologically detailed spiking neural networks International Joint Conference on Neural Networks (IJCNN).
[3]
Yu-Hsin Chen, Joel Emer, and Vivienne Sze. 2017. Using Dataflow to Optimize Energy Efficiency of Deep Neural Network Accelerators. IEEE Micro, Vol. 37, 3 (2017), 12--21.
[4]
Anup Das et al. 2018. Mapping Of Local and Global Synapses On Spiking Neuromorphic Hardware Proceedings of the Conference on Design, Automation and Test (DATE).
[5]
Anup Das et al. 2018. Unsupervised heart-rate estimation in wearables with Liquid states and a probabilistic readout. Neural Networks (2018).
[6]
Peter U Diehl and Matthew Cook. 2015. Unsupervised learning of digit recognition using spike-timing-dependent plasticity. Frontiers in computational neuroscience Vol. 9 (2015).
[7]
Hyesun Hong, Hyunok Oh, and Soonhoi Ha. 2017. Hierarchical Dataflow Modeling of Iterative Applications Proceedings of the Annual Design Automation Conference (DAC).
[8]
Giacomo Indiveri, Federico Corradi, and Ning Qiao. 2015. Neuromorphic architectures for spiking deep neural networks IEDM.
[9]
Yu Ji et al. 2016. NEUTRAMS: Neural network transformation and co-design under neuromorphic hardware constraints. In International Symposium on Microarchitecture (MICRO).
[10]
Edward A Lee and David G Messerschmitt. 1987. Synchronous data flow. Proc. IEEE Vol. 75, 9 (1987), 1235--1245.
[11]
Wenyan Lu et al. 2017. FlexFlow: A Flexible Dataflow Accelerator Architecture for Convolutional Neural Networks HPCA.
[12]
Wolfgang Maass. 1997. Networks of spiking neurons: the third generation of neural network models. Neural networks, Vol. 10, 9 (1997), 1659--1671.
[13]
Sundararajan Sriram and Shuvra Bhattacharyya. 2009. Embedded multiprocessors: Scheduling and synchronization. CRC press.
[14]
Evangelos Stromatias et al. 2015. Scalable energy-efficient, low-latency implementations of trained spiking deep belief networks on SpiNNaker. In International Joint Conference on Neural Networks (IJCNN).
[15]
Sander Stuijk, Twan Basten, MCW Geilen, and Henk Corporaal. 2007. Multiprocessor resource allocation for throughput-constrained synchronous dataflow graphs. In Proceedings of the Annual Design Automation Conference (DAC).
[16]
Sander Stuijk, Marc Geilen, and Twan Basten. 2006. SDF3: SDF for free. In ACSD.
[17]
Amirhossein Tavanaei, Timothée Masquelier, and Anthony S Maida. 2016. Acquisition of visual features through probabilistic spike-timing-dependent plasticity. International Joint Conference on Neural Networks (IJCNN).
[18]
Shouyi Yin et al. 2017. Learning Convolutional Neural Networks for Data-Flow Graph Mapping on Spatial Programmable Architectures. In FPGA.

Cited By

View all
  • (2023)Preserving Privacy of Neuromorphic Hardware From PCIe Congestion Side-Channel Attack2023 IEEE 47th Annual Computers, Software, and Applications Conference (COMPSAC)10.1109/COMPSAC57700.2023.00094(689-698)Online publication date: Jun-2023
  • (2023)Platform-Based Design of Embedded Neuromorphic SystemsEmbedded Machine Learning for Cyber-Physical, IoT, and Edge Computing10.1007/978-3-031-19568-6_12(337-358)Online publication date: 1-Oct-2023
  • (2022)Energy-Efficient Respiratory Anomaly Detection in Premature Newborn InfantsElectronics10.3390/electronics1105068211:5(682)Online publication date: 23-Feb-2022
  • Show More Cited By

Recommendations

Comments

Information & Contributors

Information

Published In

cover image ACM Conferences
GLSVLSI '18: Proceedings of the 2018 Great Lakes Symposium on VLSI
May 2018
533 pages
ISBN:9781450357241
DOI:10.1145/3194554
Permission to make digital or hard copies of all or part of this work for personal or classroom use is granted without fee provided that copies are not made or distributed for profit or commercial advantage and that copies bear this notice and the full citation on the first page. Copyrights for components of this work owned by others than ACM must be honored. Abstracting with credit is permitted. To copy otherwise, or republish, to post on servers or to redistribute to lists, requires prior specific permission and/or a fee. Request permissions from [email protected]

Sponsors

Publisher

Association for Computing Machinery

New York, NY, United States

Publication History

Published: 30 May 2018

Permissions

Request permissions for this article.

Check for updates

Author Tags

  1. neuromorphic computing
  2. spiking neural networks
  3. synchronous dataflow graph (SDFG)

Qualifiers

  • Research-article

Conference

GLSVLSI '18
Sponsor:
GLSVLSI '18: Great Lakes Symposium on VLSI 2018
May 23 - 25, 2018
IL, Chicago, USA

Acceptance Rates

GLSVLSI '18 Paper Acceptance Rate 48 of 197 submissions, 24%;
Overall Acceptance Rate 312 of 1,156 submissions, 27%

Upcoming Conference

GLSVLSI '25
Great Lakes Symposium on VLSI 2025
June 30 - July 2, 2025
New Orleans , LA , USA

Contributors

Other Metrics

Bibliometrics & Citations

Bibliometrics

Article Metrics

  • Downloads (Last 12 months)28
  • Downloads (Last 6 weeks)4
Reflects downloads up to 16 Feb 2025

Other Metrics

Citations

Cited By

View all
  • (2023)Preserving Privacy of Neuromorphic Hardware From PCIe Congestion Side-Channel Attack2023 IEEE 47th Annual Computers, Software, and Applications Conference (COMPSAC)10.1109/COMPSAC57700.2023.00094(689-698)Online publication date: Jun-2023
  • (2023)Platform-Based Design of Embedded Neuromorphic SystemsEmbedded Machine Learning for Cyber-Physical, IoT, and Edge Computing10.1007/978-3-031-19568-6_12(337-358)Online publication date: 1-Oct-2023
  • (2022)Energy-Efficient Respiratory Anomaly Detection in Premature Newborn InfantsElectronics10.3390/electronics1105068211:5(682)Online publication date: 23-Feb-2022
  • (2022)Design-Technology Co-Optimization for NVM-Based Neuromorphic Processing ElementsACM Transactions on Embedded Computing Systems10.1145/352406821:6(1-27)Online publication date: 12-Dec-2022
  • (2022)DFSynthesizer: Dataflow-based Synthesis of Spiking Neural Networks to Neuromorphic HardwareACM Transactions on Embedded Computing Systems10.1145/347915621:3(1-35)Online publication date: 28-May-2022
  • (2022)Special Session: Towards an Agile Design Methodology for Efficient, Reliable, and Secure ML Systems2022 IEEE 40th VLSI Test Symposium (VTS)10.1109/VTS52500.2021.9794253(1-14)Online publication date: 25-Apr-2022
  • (2022)Real-Time Scheduling of Machine Learning Operations on Heterogeneous Neuromorphic SoC2022 20th ACM-IEEE International Conference on Formal Methods and Models for System Design (MEMOCODE)10.1109/MEMOCODE57689.2022.9954596(1-12)Online publication date: 13-Oct-2022
  • (2021)NeuroXplorer 1.0: An Extensible Framework for Architectural Exploration with Spiking Neural NetworksInternational Conference on Neuromorphic Systems 202110.1145/3477145.3477156(1-9)Online publication date: 27-Jul-2021
  • (2021)Dynamic Reliability Management in Neuromorphic ComputingACM Journal on Emerging Technologies in Computing Systems10.1145/346233017:4(1-27)Online publication date: 21-Jul-2021
  • (2021)Special Session: Reliability Analysis for AI/ML Hardware2021 IEEE 39th VLSI Test Symposium (VTS)10.1109/VTS50974.2021.9441050(1-10)Online publication date: 25-Apr-2021
  • Show More Cited By

View Options

Login options

View options

PDF

View or Download as a PDF file.

PDF

eReader

View online with eReader.

eReader

Figures

Tables

Media

Share

Share

Share this Publication link

Share on social media