GPU implementation of neural-network simulations based on adaptive-exponential models

Alexandros Neofytou; George Chatzikonstantis; Ioannis Magkanaris; George Smaragdos; Christos Strydis; Dimitrios Soudris

doi:10.1109/BIBE.2019.00067

GPU implementation of neural-network simulations based on adaptive-exponential models

Alexandros Neofytou^*, George Chatzikonstantis, Ioannis Magkanaris, George Smaragdos, Christos Strydis, Dimitrios Soudris

^*Corresponding author for this work

Neurosciences

Research output: Chapter/Conference proceeding › Conference proceeding › Academic › peer-review

1 Citation (Scopus)

Abstract

Detailed brain modeling has been presenting significant challenges to the world of high-performance computing (HPC), posing computational problems that can benefit from modern hardware-acceleration technologies. We explore the capacity of GPUs for simulating large-scale neuronal networks based on the Adaptive Exponential neuron-model, which is widely used in the neuroscientific community. Our GPU-powered simulator acts as a benchmark to evaluate the strengths and limitations of modern GPUs, as well as to explore their scaling properties when simulating large neural networks. This work presents an optimized GPU implementation that outperforms a reference multicore implementation by 50x, whereas utilizing a dual-GPU configuration can deliver a speedup of 90x for networks of 20,000 fully interconnected AdEx neurons.

Original language	English
Title of host publication	Proceedings - 2019 IEEE 19th International Conference on Bioinformatics and Bioengineering, BIBE 2019
Publisher	Institute of Electrical and Electronics Engineers Inc.
Pages	339-343
Number of pages	5
ISBN (Electronic)	9781728146171
DOIs	https://doi.org/10.1109/BIBE.2019.00067
Publication status	Published - Oct 2019
Event	19th International Conference on Bioinformatics and Bioengineering, BIBE 2019 - Athens, Greece Duration: 28 Oct 2019 → 30 Oct 2019

Publication series

Series	Proceedings - 2019 IEEE 19th International Conference on Bioinformatics and Bioengineering, BIBE 2019

Conference

Conference	19th International Conference on Bioinformatics and Bioengineering, BIBE 2019
Country/Territory	Greece
City	Athens
Period	28/10/19 → 30/10/19

Bibliographical note

Funding Information:
This research is supported by European Commission H2020 project EXA2PRO for FETHPC-02-2017 Transition to Exascale Computing (Grant agreement ID: 801015). The work was also supported by computational time granted from the Greek Research and Technology Network (GRNET) in the National HPC facility Advanced Research Information System - ARIS

Publisher Copyright: © 2019 IEEE.

Access to Document

10.1109/BIBE.2019.00067

Cite this

Neofytou, A., Chatzikonstantis, G., Magkanaris, I., Smaragdos, G., Strydis, C., & Soudris, D. (2019). GPU implementation of neural-network simulations based on adaptive-exponential models. In Proceedings - 2019 IEEE 19th International Conference on Bioinformatics and Bioengineering, BIBE 2019 (pp. 339-343). Article 8941972 Institute of Electrical and Electronics Engineers Inc.. https://doi.org/10.1109/BIBE.2019.00067

Neofytou, Alexandros ; Chatzikonstantis, George ; Magkanaris, Ioannis et al. / GPU implementation of neural-network simulations based on adaptive-exponential models. Proceedings - 2019 IEEE 19th International Conference on Bioinformatics and Bioengineering, BIBE 2019. Institute of Electrical and Electronics Engineers Inc., 2019. pp. 339-343 (Proceedings - 2019 IEEE 19th International Conference on Bioinformatics and Bioengineering, BIBE 2019).

@inproceedings{7c173ad438ab44a0a229820015155e46,

title = "GPU implementation of neural-network simulations based on adaptive-exponential models",

abstract = "Detailed brain modeling has been presenting significant challenges to the world of high-performance computing (HPC), posing computational problems that can benefit from modern hardware-acceleration technologies. We explore the capacity of GPUs for simulating large-scale neuronal networks based on the Adaptive Exponential neuron-model, which is widely used in the neuroscientific community. Our GPU-powered simulator acts as a benchmark to evaluate the strengths and limitations of modern GPUs, as well as to explore their scaling properties when simulating large neural networks. This work presents an optimized GPU implementation that outperforms a reference multicore implementation by 50x, whereas utilizing a dual-GPU configuration can deliver a speedup of 90x for networks of 20,000 fully interconnected AdEx neurons.",

author = "Alexandros Neofytou and George Chatzikonstantis and Ioannis Magkanaris and George Smaragdos and Christos Strydis and Dimitrios Soudris",

note = "Funding Information: This research is supported by European Commission H2020 project EXA2PRO for FETHPC-02-2017 Transition to Exascale Computing (Grant agreement ID: 801015). The work was also supported by computational time granted from the Greek Research and Technology Network (GRNET) in the National HPC facility Advanced Research Information System - ARIS Publisher Copyright: {\textcopyright} 2019 IEEE.; 19th International Conference on Bioinformatics and Bioengineering, BIBE 2019 ; Conference date: 28-10-2019 Through 30-10-2019",

year = "2019",

month = oct,

doi = "10.1109/BIBE.2019.00067",

language = "English",

series = "Proceedings - 2019 IEEE 19th International Conference on Bioinformatics and Bioengineering, BIBE 2019",

pages = "339--343",

booktitle = "Proceedings - 2019 IEEE 19th International Conference on Bioinformatics and Bioengineering, BIBE 2019",

publisher = "Institute of Electrical and Electronics Engineers Inc.",

address = "United States",

}

Neofytou, A, Chatzikonstantis, G, Magkanaris, I, Smaragdos, G , Strydis, C & Soudris, D 2019, GPU implementation of neural-network simulations based on adaptive-exponential models. in Proceedings - 2019 IEEE 19th International Conference on Bioinformatics and Bioengineering, BIBE 2019., 8941972, Institute of Electrical and Electronics Engineers Inc., Proceedings - 2019 IEEE 19th International Conference on Bioinformatics and Bioengineering, BIBE 2019, pp. 339-343, 19th International Conference on Bioinformatics and Bioengineering, BIBE 2019, Athens, Greece, 28/10/19. https://doi.org/10.1109/BIBE.2019.00067

GPU implementation of neural-network simulations based on adaptive-exponential models. / Neofytou, Alexandros; Chatzikonstantis, George; Magkanaris, Ioannis et al.
Proceedings - 2019 IEEE 19th International Conference on Bioinformatics and Bioengineering, BIBE 2019. Institute of Electrical and Electronics Engineers Inc., 2019. p. 339-343 8941972 (Proceedings - 2019 IEEE 19th International Conference on Bioinformatics and Bioengineering, BIBE 2019).

Research output: Chapter/Conference proceeding › Conference proceeding › Academic › peer-review

TY - GEN

T1 - GPU implementation of neural-network simulations based on adaptive-exponential models

AU - Neofytou, Alexandros

AU - Chatzikonstantis, George

AU - Magkanaris, Ioannis

AU - Smaragdos, George

AU - Strydis, Christos

AU - Soudris, Dimitrios

N1 - Funding Information: This research is supported by European Commission H2020 project EXA2PRO for FETHPC-02-2017 Transition to Exascale Computing (Grant agreement ID: 801015). The work was also supported by computational time granted from the Greek Research and Technology Network (GRNET) in the National HPC facility Advanced Research Information System - ARIS Publisher Copyright: © 2019 IEEE.

PY - 2019/10

Y1 - 2019/10

N2 - Detailed brain modeling has been presenting significant challenges to the world of high-performance computing (HPC), posing computational problems that can benefit from modern hardware-acceleration technologies. We explore the capacity of GPUs for simulating large-scale neuronal networks based on the Adaptive Exponential neuron-model, which is widely used in the neuroscientific community. Our GPU-powered simulator acts as a benchmark to evaluate the strengths and limitations of modern GPUs, as well as to explore their scaling properties when simulating large neural networks. This work presents an optimized GPU implementation that outperforms a reference multicore implementation by 50x, whereas utilizing a dual-GPU configuration can deliver a speedup of 90x for networks of 20,000 fully interconnected AdEx neurons.

AB - Detailed brain modeling has been presenting significant challenges to the world of high-performance computing (HPC), posing computational problems that can benefit from modern hardware-acceleration technologies. We explore the capacity of GPUs for simulating large-scale neuronal networks based on the Adaptive Exponential neuron-model, which is widely used in the neuroscientific community. Our GPU-powered simulator acts as a benchmark to evaluate the strengths and limitations of modern GPUs, as well as to explore their scaling properties when simulating large neural networks. This work presents an optimized GPU implementation that outperforms a reference multicore implementation by 50x, whereas utilizing a dual-GPU configuration can deliver a speedup of 90x for networks of 20,000 fully interconnected AdEx neurons.

UR - http://www.scopus.com/inward/record.url?scp=85077976064&partnerID=8YFLogxK

U2 - 10.1109/BIBE.2019.00067

DO - 10.1109/BIBE.2019.00067

M3 - Conference proceeding

AN - SCOPUS:85077976064

T3 - Proceedings - 2019 IEEE 19th International Conference on Bioinformatics and Bioengineering, BIBE 2019

SP - 339

EP - 343

BT - Proceedings - 2019 IEEE 19th International Conference on Bioinformatics and Bioengineering, BIBE 2019

PB - Institute of Electrical and Electronics Engineers Inc.

T2 - 19th International Conference on Bioinformatics and Bioengineering, BIBE 2019

Y2 - 28 October 2019 through 30 October 2019

ER -

Neofytou A, Chatzikonstantis G, Magkanaris I, Smaragdos G , Strydis C, Soudris D. GPU implementation of neural-network simulations based on adaptive-exponential models. In Proceedings - 2019 IEEE 19th International Conference on Bioinformatics and Bioengineering, BIBE 2019. Institute of Electrical and Electronics Engineers Inc. 2019. p. 339-343. 8941972. (Proceedings - 2019 IEEE 19th International Conference on Bioinformatics and Bioengineering, BIBE 2019). doi: 10.1109/BIBE.2019.00067

GPU implementation of neural-network simulations based on adaptive-exponential models

Abstract

Publication series

Conference

Bibliographical note

Access to Document

Other files and links

Fingerprint

Cite this