NeuroVP: A System-Level Virtual Platform for Integration of Neuromorphic Accelerators

Melvin Galicia; Ali Banagozar; Karl Sturm; Felix Staudigl; Sander Stuijk; Henk Corporaal; Rainer Leupers

doi:10.1109/SOCC52499.2021.9739585

NeuroVP: A System-Level Virtual Platform for Integration of Neuromorphic Accelerators

Melvin Galicia, Ali Banagozar, Karl Sturm, Felix Staudigl, Sander Stuijk, Henk Corporaal, Rainer Leupers

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution › Academic › peer-review

7 Citations (Scopus)

1 Downloads (Pure)

Abstract

Executing neural network (NN) applications on general-purpose processors result in a large power and performance overhead, due to the high cost of data movement between the processor and the main memory. Neuromorphic computing systems based on memristor crossbars, perform the NN main operation i.e., vector-matrix multiplications (VMM) in an efficient way in the analog domain. Thus, they circumvent the costly energy overhead of its digital counterpart. It can be expected that neuromorphic systems will be used initially as complements to current high-performance systems rather than as a replacement. This paper presents NeuroVP, a virtual platform integrating a neuromorphic accelerator, developed in SystemC that can model functionality, timing, and power consumption of the components integrating the system. Using NeuroVP to evaluate performance and power consumption at the electronic system level (ESL), it is corroborated that the execution of NN applications with a neuromorphic accelerator yields of up to 46x higher power efficiency and 26x speedup relative to a general-purpose computing system.

Original language	English
Title of host publication	Proceedings 34th IEEE International System-on-Chip Conference (SOCC)
Editors	Gang Qu, Jinjun Xiong, Danella Zhao, Venki Muthukumar, Md Farhadur Reza, Ramalingam Sridhar
Publisher	Institute of Electrical and Electronics Engineers
Pages	236-241
Number of pages	6
ISBN (Electronic)	978-1-6654-2931-3
DOIs	https://doi.org/10.1109/SOCC52499.2021.9739585
Publication status	Published - 24 Mar 2022
Event	34th IEEE International System on Chip Conference, SOCC 2021 - Virtual, Online, Las Vegas, United States Duration: 14 Sept 2021 → 17 Sept 2021 Conference number: 34

Conference

Conference	34th IEEE International System on Chip Conference, SOCC 2021
Abbreviated title	SOCC
Country/Territory	United States
City	Las Vegas
Period	14/09/21 → 17/09/21

Funding

This work was supported by the Federal Ministry of Education and Research (BMBF, Germany) within the NEUROTEC project (No. 16ES1134 and 16ES1133K).

Keywords

ESL Power Estimation
Memristor
Neuromorphic Accelerator
RISC-V
SystemC
Virtual Platform

Access to Document

10.1109/SOCC52499.2021.9739585

Cite this

Galicia, M., Banagozar, A., Sturm, K., Staudigl, F., Stuijk, S., Corporaal, H., & Leupers, R. (2022). NeuroVP: A System-Level Virtual Platform for Integration of Neuromorphic Accelerators. In G. Qu, J. Xiong, D. Zhao, V. Muthukumar, M. F. Reza, & R. Sridhar (Eds.), Proceedings 34th IEEE International System-on-Chip Conference (SOCC) (pp. 236-241). Article 9739585 Institute of Electrical and Electronics Engineers. https://doi.org/10.1109/SOCC52499.2021.9739585

Galicia, Melvin ; Banagozar, Ali ; Sturm, Karl et al. / NeuroVP : A System-Level Virtual Platform for Integration of Neuromorphic Accelerators. Proceedings 34th IEEE International System-on-Chip Conference (SOCC). editor / Gang Qu ; Jinjun Xiong ; Danella Zhao ; Venki Muthukumar ; Md Farhadur Reza ; Ramalingam Sridhar. Institute of Electrical and Electronics Engineers, 2022. pp. 236-241

@inproceedings{8b5b8f0a1b2245898e754327320c3f9b,

title = "NeuroVP: A System-Level Virtual Platform for Integration of Neuromorphic Accelerators",

abstract = "Executing neural network (NN) applications on general-purpose processors result in a large power and performance overhead, due to the high cost of data movement between the processor and the main memory. Neuromorphic computing systems based on memristor crossbars, perform the NN main operation i.e., vector-matrix multiplications (VMM) in an efficient way in the analog domain. Thus, they circumvent the costly energy overhead of its digital counterpart. It can be expected that neuromorphic systems will be used initially as complements to current high-performance systems rather than as a replacement. This paper presents NeuroVP, a virtual platform integrating a neuromorphic accelerator, developed in SystemC that can model functionality, timing, and power consumption of the components integrating the system. Using NeuroVP to evaluate performance and power consumption at the electronic system level (ESL), it is corroborated that the execution of NN applications with a neuromorphic accelerator yields of up to 46x higher power efficiency and 26x speedup relative to a general-purpose computing system. ",

keywords = "ESL Power Estimation, Memristor, Neuromorphic Accelerator, RISC-V, SystemC, Virtual Platform",

author = "Melvin Galicia and Ali Banagozar and Karl Sturm and Felix Staudigl and Sander Stuijk and Henk Corporaal and Rainer Leupers",

year = "2022",

month = mar,

day = "24",

doi = "10.1109/SOCC52499.2021.9739585",

language = "English",

pages = "236--241",

editor = "Gang Qu and Jinjun Xiong and Danella Zhao and Venki Muthukumar and Reza, {Md Farhadur} and Ramalingam Sridhar",

booktitle = "Proceedings 34th IEEE International System-on-Chip Conference (SOCC)",

publisher = "Institute of Electrical and Electronics Engineers",

address = "United States",

note = "34th IEEE International System on Chip Conference, SOCC 2021, SOCC ; Conference date: 14-09-2021 Through 17-09-2021",

}

Galicia, M, Banagozar, A, Sturm, K, Staudigl, F, Stuijk, S , Corporaal, H & Leupers, R 2022, NeuroVP: A System-Level Virtual Platform for Integration of Neuromorphic Accelerators. in G Qu, J Xiong, D Zhao, V Muthukumar, MF Reza & R Sridhar (eds), Proceedings 34th IEEE International System-on-Chip Conference (SOCC)., 9739585, Institute of Electrical and Electronics Engineers, pp. 236-241, 34th IEEE International System on Chip Conference, SOCC 2021, Las Vegas, Nevada, United States, 14/09/21. https://doi.org/10.1109/SOCC52499.2021.9739585

NeuroVP: A System-Level Virtual Platform for Integration of Neuromorphic Accelerators. / Galicia, Melvin; Banagozar, Ali; Sturm, Karl et al.
Proceedings 34th IEEE International System-on-Chip Conference (SOCC). ed. / Gang Qu; Jinjun Xiong; Danella Zhao; Venki Muthukumar; Md Farhadur Reza; Ramalingam Sridhar. Institute of Electrical and Electronics Engineers, 2022. p. 236-241 9739585.

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution › Academic › peer-review

TY - GEN

T1 - NeuroVP

T2 - 34th IEEE International System on Chip Conference, SOCC 2021

AU - Galicia, Melvin

AU - Banagozar, Ali

AU - Sturm, Karl

AU - Staudigl, Felix

AU - Stuijk, Sander

AU - Corporaal, Henk

AU - Leupers, Rainer

N1 - Conference code: 34

PY - 2022/3/24

Y1 - 2022/3/24

N2 - Executing neural network (NN) applications on general-purpose processors result in a large power and performance overhead, due to the high cost of data movement between the processor and the main memory. Neuromorphic computing systems based on memristor crossbars, perform the NN main operation i.e., vector-matrix multiplications (VMM) in an efficient way in the analog domain. Thus, they circumvent the costly energy overhead of its digital counterpart. It can be expected that neuromorphic systems will be used initially as complements to current high-performance systems rather than as a replacement. This paper presents NeuroVP, a virtual platform integrating a neuromorphic accelerator, developed in SystemC that can model functionality, timing, and power consumption of the components integrating the system. Using NeuroVP to evaluate performance and power consumption at the electronic system level (ESL), it is corroborated that the execution of NN applications with a neuromorphic accelerator yields of up to 46x higher power efficiency and 26x speedup relative to a general-purpose computing system.

AB - Executing neural network (NN) applications on general-purpose processors result in a large power and performance overhead, due to the high cost of data movement between the processor and the main memory. Neuromorphic computing systems based on memristor crossbars, perform the NN main operation i.e., vector-matrix multiplications (VMM) in an efficient way in the analog domain. Thus, they circumvent the costly energy overhead of its digital counterpart. It can be expected that neuromorphic systems will be used initially as complements to current high-performance systems rather than as a replacement. This paper presents NeuroVP, a virtual platform integrating a neuromorphic accelerator, developed in SystemC that can model functionality, timing, and power consumption of the components integrating the system. Using NeuroVP to evaluate performance and power consumption at the electronic system level (ESL), it is corroborated that the execution of NN applications with a neuromorphic accelerator yields of up to 46x higher power efficiency and 26x speedup relative to a general-purpose computing system.

KW - ESL Power Estimation

KW - Memristor

KW - Neuromorphic Accelerator

KW - RISC-V

KW - SystemC

KW - Virtual Platform

UR - http://www.scopus.com/inward/record.url?scp=85122120919&partnerID=8YFLogxK

U2 - 10.1109/SOCC52499.2021.9739585

DO - 10.1109/SOCC52499.2021.9739585

M3 - Conference contribution

AN - SCOPUS:85122120919

SP - 236

EP - 241

BT - Proceedings 34th IEEE International System-on-Chip Conference (SOCC)

A2 - Qu, Gang

A2 - Xiong, Jinjun

A2 - Zhao, Danella

A2 - Muthukumar, Venki

A2 - Reza, Md Farhadur

A2 - Sridhar, Ramalingam

PB - Institute of Electrical and Electronics Engineers

Y2 - 14 September 2021 through 17 September 2021

ER -

Galicia M, Banagozar A, Sturm K, Staudigl F, Stuijk S , Corporaal H et al. NeuroVP: A System-Level Virtual Platform for Integration of Neuromorphic Accelerators. In Qu G, Xiong J, Zhao D, Muthukumar V, Reza MF, Sridhar R, editors, Proceedings 34th IEEE International System-on-Chip Conference (SOCC). Institute of Electrical and Electronics Engineers. 2022. p. 236-241. 9739585 doi: 10.1109/SOCC52499.2021.9739585

NeuroVP: A System-Level Virtual Platform for Integration of Neuromorphic Accelerators

Abstract

Conference

Funding

Keywords

Access to Document

Other files and links

Fingerprint

Cite this