Open Access
Science, volume 382, issue 6668, pages 329-335
Neural inference at the frontier of energy, space, and time
Dharmendra S. Modha
1
,
Filipp Akopyan
1
,
Alexander Andreopoulos
1
,
Rathinakumar Appuswamy
1
,
John V Arthur
1
,
Andrew Cassidy
1
,
Pallab Datta
1
,
Michael V Debole
1
,
Steven K. Esser
1
,
Carlos Ortega Otero
1
,
Jun Sawada
1
,
Brian Taba
1
,
Arnon Amir
1
,
Deepika Bablani
1
,
Peter J Carlson
1
,
Myron D Flickner
1
,
Rajamohan Gandhasri
1
,
Guillaume J. Garreau
1
,
Megumi Ito
1
,
Jennifer L. Klamo
1
,
Jeffrey A. Kusnitz
1
,
Nathaniel J Mcclatchey
1
,
Jeffrey L. McKinstry
1
,
Yutaka Nakamura
1
,
TAPAN K. NAYAK
1
,
William P. Risk
1
,
Kai Schleupen
1
,
Ben Shaw
1
,
Jay Sivagnaname
1
,
Daniel F Smith
1
,
Ignacio Terrizzano
1
,
Takanori Ueda
1
1
IBM Research, San Jose, CA, USA.
|
Publication type: Journal Article
Publication date: 2023-10-20
Multidisciplinary
Abstract
Computing, since its inception, has been processor-centric, with memory separated from compute. Inspired by the organic brain and optimized for inorganic silicon, NorthPole is a neural inference architecture that blurs this boundary by eliminating off-chip memory, intertwining compute with memory on-chip, and appearing externally as an active memory chip. NorthPole is a low-precision, massively parallel, densely interconnected, energy-efficient, and spatial computing architecture with a co-optimized, high-utilization programming model. On the ResNet50 benchmark image classification network, relative to a graphics processing unit (GPU) that uses a comparable 12-nanometer technology process, NorthPole achieves a 25 times higher energy metric of frames per second (FPS) per watt, a 5 times higher space metric of FPS per transistor, and a 22 times lower time metric of latency. Similar results are reported for the Yolo-v4 detection network. NorthPole outperforms all prevalent architectures, even those that use more-advanced technology processes.
Top-30
Journals
|
1
2
|
|
|
Nature Electronics
2 publications, 5.71%
|
|
|
Nature Communications
2 publications, 5.71%
|
|
|
Nanomaterials
2 publications, 5.71%
|
|
|
Electronics (Switzerland)
2 publications, 5.71%
|
|
|
Science
1 publication, 2.86%
|
|
|
ACS Applied Nano Materials
1 publication, 2.86%
|
|
|
Nature
1 publication, 2.86%
|
|
|
JPSJ News and Comments
1 publication, 2.86%
|
|
|
Nature Reviews Electrical Engineering
1 publication, 2.86%
|
|
|
Journal of Computer Science and Technology
1 publication, 2.86%
|
|
|
Current Opinion in Structural Biology
1 publication, 2.86%
|
|
|
Modern Electronic Materials
1 publication, 2.86%
|
|
|
Neural Networks
1 publication, 2.86%
|
|
|
NDT
1 publication, 2.86%
|
|
|
Proceedings of the IEEE
1 publication, 2.86%
|
|
|
Advanced Functional Materials
1 publication, 2.86%
|
|
|
Results in Engineering
1 publication, 2.86%
|
|
|
Pattern Analysis and Applications
1 publication, 2.86%
|
|
|
Lecture Notes in Computer Science
1 publication, 2.86%
|
|
|
Mesoscience and Nanotechnology
1 publication, 2.86%
|
|
|
IEEE Transactions on Computational Social Systems
1 publication, 2.86%
|
|
|
IEEE Embedded Systems Letters
1 publication, 2.86%
|
|
|
InfoMat
1 publication, 2.86%
|
|
|
1
2
|
Publishers
|
1
2
3
4
5
6
7
8
9
|
|
|
Springer Nature
9 publications, 25.71%
|
|
|
Institute of Electrical and Electronics Engineers (IEEE)
8 publications, 22.86%
|
|
|
MDPI
5 publications, 14.29%
|
|
|
Elsevier
3 publications, 8.57%
|
|
|
Wiley
2 publications, 5.71%
|
|
|
Cold Spring Harbor Laboratory
2 publications, 5.71%
|
|
|
American Association for the Advancement of Science (AAAS)
1 publication, 2.86%
|
|
|
American Chemical Society (ACS)
1 publication, 2.86%
|
|
|
Physical Society of Japan
1 publication, 2.86%
|
|
|
Pensoft Publishers
1 publication, 2.86%
|
|
|
SPIE-Intl Soc Optical Eng
1 publication, 2.86%
|
|
|
Treatise
1 publication, 2.86%
|
|
|
1
2
3
4
5
6
7
8
9
|
- We do not take into account publications without a DOI.
- Statistics recalculated only for publications connected to researchers, organizations and labs registered on the platform.
- Statistics recalculated weekly.
Are you a researcher?
Create a profile to get free access to personal recommendations for colleagues and new articles.
Metrics
Cite this
GOST |
RIS |
BibTex |
MLA
Cite this
GOST
Copy
Modha D. S. et al. Neural inference at the frontier of energy, space, and time // Science. 2023. Vol. 382. No. 6668. pp. 329-335.
GOST all authors (up to 50)
Copy
Modha D. S., Akopyan F., Andreopoulos A., Appuswamy R., Arthur J. V., Cassidy A., Datta P., Debole M. V., Esser S. K., Otero C. O., Sawada J., Taba B., Amir A., Bablani D., Carlson P. J., Flickner M. D., Gandhasri R., Garreau G. J., Ito M., Klamo J. L., Kusnitz J. A., Mcclatchey N. J., McKinstry J. L., Nakamura Y., NAYAK T. K., Risk W. P., Schleupen K., Shaw B., Sivagnaname J., Smith D. F., Terrizzano I., Ueda T. Neural inference at the frontier of energy, space, and time // Science. 2023. Vol. 382. No. 6668. pp. 329-335.
Cite this
RIS
Copy
TY - JOUR
DO - 10.1126/science.adh1174
UR - https://doi.org/10.1126/science.adh1174
TI - Neural inference at the frontier of energy, space, and time
T2 - Science
AU - Modha, Dharmendra S.
AU - Akopyan, Filipp
AU - Andreopoulos, Alexander
AU - Appuswamy, Rathinakumar
AU - Arthur, John V
AU - Cassidy, Andrew
AU - Datta, Pallab
AU - Debole, Michael V
AU - Esser, Steven K.
AU - Otero, Carlos Ortega
AU - Sawada, Jun
AU - Taba, Brian
AU - Amir, Arnon
AU - Bablani, Deepika
AU - Carlson, Peter J
AU - Flickner, Myron D
AU - Gandhasri, Rajamohan
AU - Garreau, Guillaume J.
AU - Ito, Megumi
AU - Klamo, Jennifer L.
AU - Kusnitz, Jeffrey A.
AU - Mcclatchey, Nathaniel J
AU - McKinstry, Jeffrey L.
AU - Nakamura, Yutaka
AU - NAYAK, TAPAN K.
AU - Risk, William P.
AU - Schleupen, Kai
AU - Shaw, Ben
AU - Sivagnaname, Jay
AU - Smith, Daniel F
AU - Terrizzano, Ignacio
AU - Ueda, Takanori
PY - 2023
DA - 2023/10/20
PB - American Association for the Advancement of Science (AAAS)
SP - 329-335
IS - 6668
VL - 382
SN - 0036-8075
SN - 1095-9203
ER -
Cite this
BibTex (up to 50 authors)
Copy
@article{2023_Modha,
author = {Dharmendra S. Modha and Filipp Akopyan and Alexander Andreopoulos and Rathinakumar Appuswamy and John V Arthur and Andrew Cassidy and Pallab Datta and Michael V Debole and Steven K. Esser and Carlos Ortega Otero and Jun Sawada and Brian Taba and Arnon Amir and Deepika Bablani and Peter J Carlson and Myron D Flickner and Rajamohan Gandhasri and Guillaume J. Garreau and Megumi Ito and Jennifer L. Klamo and Jeffrey A. Kusnitz and Nathaniel J Mcclatchey and Jeffrey L. McKinstry and Yutaka Nakamura and TAPAN K. NAYAK and William P. Risk and Kai Schleupen and Ben Shaw and Jay Sivagnaname and Daniel F Smith and Ignacio Terrizzano and Takanori Ueda},
title = {Neural inference at the frontier of energy, space, and time},
journal = {Science},
year = {2023},
volume = {382},
publisher = {American Association for the Advancement of Science (AAAS)},
month = {oct},
url = {https://doi.org/10.1126/science.adh1174},
number = {6668},
pages = {329--335},
doi = {10.1126/science.adh1174}
}
Cite this
MLA
Copy
Modha, Dharmendra S., et al. “Neural inference at the frontier of energy, space, and time.” Science, vol. 382, no. 6668, Oct. 2023, pp. 329-335. https://doi.org/10.1126/science.adh1174.