Breakthrough Edge AI Inference Performance Using Northpole in 3U Vpx Form Factor

We present preliminary results demonstrating AI (artificial intelligence) inference using the IBM AIU NorthPole Chip [1], [2] incorporated into a compact, rugged 3U VPX form factor module (NP-VPX) [3]. NP-VPX allows NorthPole to be used in edge applications with stringent cooling requirements, high-speed switch fabrics, and rugged environments. NP-VPX processes 965 frames per second (fps) with a Yolo-v4 network with <tex xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">$640 \times 640$</tex> pixel images at 73.5 W at full-precision accuracy, achieving 13.2 frames / J (fps/W). NP-VPX processes over 40,300 fps with a ResNet-50 network with <tex xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">$224 \times 224$</tex> pixel images at 65.9 W at full-precision accuracy, achieving 611 frames <tex xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">$/ \mathrm{J}$</tex>.

Authors

Publication Details

Published
2024-09-23
DOI
https://doi.org/10.1109/hpec62836.2024.10938474
Citations
2
Primary Topic
Optical Systems and Laser Technology
Type
article
Field-Weighted Citation Impact
0.42
Controls
|||
ALL TIME
JAN
FEB
MAR
APR
MAY
JUN
JUL
AUG
SEP
OCT
article

Breakthrough Edge AI Inference Performance Using Northpole in 3U Vpx Form Factor

Filipp Akopyan, Guillaume Garreau, Myron Flickner, Michael E. Criscolo et al.
2 citations
Optical Systems and Laser Technology
0.42
article

Breakthrough Edge AI Inference Performance Using Northpole in 3U Vpx Form Factor

Filipp Akopyan, Guillaume Garreau, Myron Flickner, Michael E. Criscolo, Takanori Ueda, Rafael Vieira de Sousa, Rathinakumar Appuswamy, Michael DeBole, Peter J. Carlson, E. G. Colgan, Brian Taba, W. P. Risk, Megumi Ito, Andrew S. Cassidy, Deepika Bablani, Ignacio Terrizzano, Steven K. Esser, Jeffrey L. McKinstry, Nathaniel J. McClatchey, Dharmendra S. Modha, Yutaka Nakamura, Pallab Datta, Carlos Ortega Otero, Jeffrey A. Kusnitz, Heinz Baier, Jun Sawada, Jennifer L. Klamo, John V. Arthur, Neil McGlohon, Rajamohan Gandhasri, Jay Sivagnaname, Alexander Andreopoulos, Arnon Amir, Tapan K. Nayak, P. Mann, Daniel F. Smith, Kai Schleupen
article en
2 citations

Abstract

We present preliminary results demonstrating AI (artificial intelligence) inference using the IBM AIU NorthPole Chip [1], [2] incorporated into a compact, rugged 3U VPX form factor module (NP-VPX) [3]. NP-VPX allows NorthPole to be used in edge applications with stringent cooling requirements, high-speed switch fabrics, and rugged environments. NP-VPX processes 965 frames per second (fps) with a Yolo-v4 network with <tex xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">$640 \times 640$</tex> pixel images at 73.5 W at full-precision accuracy, achieving 13.2 frames / J (fps/W). NP-VPX processes over 40,300 fps with a ResNet-50 network with <tex xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">$224 \times 224$</tex> pixel images at 65.9 W at full-precision accuracy, achieving 611 frames <tex xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">$/ \mathrm{J}$</tex>.

Openalex Percentile: Top 35%
Optical Systems and Laser Technology
0.42
AI Navigator

Ask Laika to Summarize, Analyze, and Connect papers live on the map.

Summarize Papers & Methodologies

Extract key findings, datasets, and comparative methods across publications.

Benchmark Rankings & Visual Analytics

Rank top research institutions, authors, funders, topics, and journals by Field-Weighted Citation Impact (FWCI) and paper volume with instant charts.

Connect Distant Disciplines

Bridge topological clusters on the map to find hidden collaborative intersections.