Project: ACA

Participation in the Advanced Computing Architectures Project (completed)

Project: ACA

Goals

We contributed to the ACA project by covering five tasks that dealt with the communication between computational units, the communication to a host environment, and the creation of a system model based on a requirement database derived from representative science and test cases of neuroscientific neural network models.

Duration

Nov. 2018 – Oct. 2022

Participants

  • Forschungszentrum Jülich GmbH
  • RWTH Aachen University
  • Heidelberg University
  • University of Manchester

Further Info
ACA Website

Contacts

Dr.-Ing. Markus RobensResearcherBuilding 02.5 / Room 217+49 2461/61-3023
Dr.-Ing. Gudrun WagenknechtSenior Scientist Neuromorphic Computing - Principal InvestigatorBuilding 02.5 / Room 119+49 2461/61-3184

Project Description

The pilot project ACA (Advanced Computing Architectures) was carried out from November 2018 to October 2022. It was funded by the Helmholtz Initiative and Networking Fund as well as Forschungszentrum Jülich GmbH and combined expertise at Forschungszentrum Jülich GmbH, the RWTH Aachen University, Heidelberg University, and the University of Manchester. It aimed at providing a specification of a future neuromorphic computing architecture, efficient neuromorphic computing concepts, as well as workflows for a systematic validation and benchmarking of neuromorphic architectures. Research initiated inside the project is continued in the frame of the Jülich Neuromorphic Computing Alliance (JUNCA).

Find out more about ACA.

We contributed to this project by dealing with different tasks including the intra- and inter-chip communication, the external communication and requirements derived from representative science and test cases of neuroscientific neural network models, which – in combination with additional tasks – led to the definition of a system model.

Achievements

  • NeuCoNS: Python-based network simulator for the examination of communication within large scale heterogeneous spiking neural networks [1, 2]
  • NENoCSi and McAERsim: SystemC-based, cycle accurate network simulators for the examination of communication within heterogeneous spiking neural networks [3]
  • Neuron to computational node mapping based on the Hilbert space-filling curve [4]
  • Network concepts based on the random addition of connections [5] or hierarchical organization [6]
  • Requirement database derived from representative science and test cases of neuroscientific neural network models [7]
  • System model implemented in SysML [8]
  • Concept for a hybrid FPGA based accelerator system [8]

References

References

[1] R. Kleijnen, M. Robens, M. Schiek, and S. van Waasen, “A Network Simulator for the Estimation of Bandwidth Load and Latency Created by Heterogeneous Spiking Neural Networks on Neuromorphic Computing Communication Networks,” Journal of Low Power Electronics and Applications, Vol. 12, No. 2, Apr. 2022, doi: 10.3390/jlpea12020023

[2] R. Kleijnen, M. Robens, M. Schiek, and S. van Waasen, “Verification of a Neuromorphic Computing Network Simulator Using Experimental Traffic Data,” Frontiers in Neuroscience, Vol. 16, Aug. 2022, doi: 10.3389/fnins.2022.958343

[3] M. Robens, R. Kleijnen, M. Schiek, and S. van Waasen, “NoC Simulation Steered by NEST: McAERsim and a Noxim Patch,” Frontiers in Neuroscience, Vol. 18, Jun. 2024, doi: 10.3389/fnins.2024.1371103

[4] R. Kleijnen, “NeuCoNS and Stacked-Net: Facilitating the Communication for Accelerated Neuroscientific Simulations,” Ph.D. dissertation, ZEA-2, Forschungszentrum Jülich GmbH, Germany, 2024, ch. 3, doi: 10.34734/FZJ-2024-06126

[5] R. Kleijnen and P. Ebrahimzadeh, “Modified Communication Networks for the Simulation of Neuromorphic Systems,” Networks 2021, virtual, USA, 5-10 Jul. 2021 [Slides]

[6] R. Kleijnen, “NeuCoNS and Stacked-Net: Facilitating the Communication for Accelerated Neuroscientific Simulations,” Ph.D. dissertation, ZEA-2, Forschungszentrum Jülich GmbH, Germany, 2024, ch. 5, doi: 10.34734/FZJ-2024-06126

[7] G. Wagenknecht, “Table and Summarizing Description of Important Science & Test Cases Parameters Required for Future SW & HW Developments,” JuGit, Feb. 2023 [Online, restricted]

[8] D. Amarasingha, G. Wagenknecht, S. Suslov, R. Kleijnen, P. Ebrahimzadeh, M. Robens, and M. Schiek, “Demonstrator Architecture derived from Science Cases with Focus on Communication,” ACA Farewell Meeting, Niederzier-Hambach, Germany, 19-20 Sep. 2022

Copyright ACA-Logo and Icons: Forschungszentrum Jülich GmbH / SBC Lehmann

Last Modified: 07.03.2025