IBM NorthPole - Neural Inference at the Frontier of Energy, Space, and Time

NorthPole outperforms all prevalent architectures, even those that use more-advanced technology processes.

Edit Page

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.

Social share preview for IBM NorthPole - Neural inference at the frontier of energy, space, and time

Upcoming Workshops

No workshops are currently scheduled. Check back soon for new events!

Are you an expert in a neuromorphic topic? We invite you to share your knowledge with our community. Hosting a workshop is a great way to engage with peers and share your work.

IBM NorthPole - Neural inference at the frontier of energy, space, and time

About the Speaker

Dr. Carlos Ortega-Otero is an Sr. Research Staff Member at IBM driven by a passion in Circuit Design, Neuromorphic Chip Architectures, Low-Power Circuits and Physical Design optimizations. He earned his Ph.D. from Cornell University under the guidance of Prof. Rajit Manohar.

Throughout his career, he has worked in groundbreaking projects, including Ultra-Low Power Asynchronous Sensor Network nodes, Medical Implantable Wireless Sensors, The TrueNorth Brain-Inspired Chip, and the NorthPole Project. At IBM, Carlos works under the leadership of Dr. Dharmendra Modha in the Brain-Inspired Computing Group.

He plays key roles in Architecture, Specification, Digital Implementation, Physical Design, Timing Signoff, and Manufacturing teams of the NorthPole Project. Carlos is proud to be part of the Brain-Inspired Computing Group at IBM that continues to shape the future of Integrated Circuits and AI.

Inspired? Share your work.

Share your expertise with the community by speaking at a workshop, student talk, or hacking hour. It’s a great way to get feedback and help others learn.

Learn How to Present

Related Workshops

From C/C++ to Dynamically Scheduled Circuits

From C/C++ to Dynamically Scheduled Circuits

Explore the journey from C/C++ to Dynamically Scheduled Circuits with Lana Josipović, an expert in high-level synthesis and reconfigurable computing. Join her recorded workshop session on innovative hardware design techniques.

Accelerating Inference and Training at the Edge

Accelerating Inference and Training at the Edge

Join us for a talk by Maxence Ernoult, Research Scientist at Rain, on accelerating inference and training at the edge.

Low-power Spiking Neural Network Processing Systems for Extreme-Edge Applications

Low-power Spiking Neural Network Processing Systems for Extreme-Edge Applications

Join Dr. Federico Corradi as he explores low-power spiking neural network processing systems, offering insights into energy-efficient computing for extreme-edge applications.