NEUROPULS project

• Post by: manager
• September 15, 2023
• No Comment

NEUROPULS: NEUROmorphic energy-efficient secure accelerators based on Phase change materials aUgmented siLicon photonicS.

Unlock the future of edge computing with NEUROPULS, a pioneering project that aims to revolutionize the processing of massive data streams in self-driving vehicles, IoT devices, and Industry 4.0 applications. NEUROPULS introduces energy-efficient and secure neuromorphic accelerators built on cutting-edge technologies, including phase change materials and augmented silicon photonics.

As the demand for data processing at the edge grows exponentially, conventional computing systems face significant limitations. NEUROPULS addresses these challenges by embracing neuromorphic computing principles, providing improved energy efficiency, reduced latency, and enhanced bandwidth density for critical applications.

Security is paramount in edge computing. NEUROPULS integrates novel security layers based on photonic Physical Unclonable Functions (PUFs), eliminating vulnerabilities associated with memory-stored secret keys. This innovative approach enhances the security of edge devices, ensuring protection against memory-accessing attacks and cyber threats.

NEUROPULS leverages CMOS-compatible platforms with augmented silicon photonics. It introduces non-volatile phase change materials for synapses/neurons and III-V materials for on-chip spiking sources, enabling the creation of novel neuromorphic accelerators. These accelerators feature RISC-V compliant interfaces, ensuring seamless adoption and programmability.

NEUROPULS employs a groundbreaking full-system simulation platform for design space exploration. This platform ensures optimal performance, allowing the project to achieve two orders of magnitude improvement in energy efficiency.

NEUROPULS considers three key use cases for benchmarking, demonstrating its transformative impact on edge computing. Explore how NEUROPULS reshapes the landscape of energy-efficient and secure edge-computing systems for a smarter, safer future.

In summary, NEUROPULS represents a paradigm shift in edge computing, combining neuromorphic computing principles, photonic PUFs, and advanced materials to create energy-efficient, secure, and high-performance neuromorphic accelerators. Join us on this journey to redefine the possibilities of edge computing.

If you find this project interesting, also check out the other projects of our team. 

Project’s details:

Duration: 2023-2026
Coordinator: Centre National De La Recherche Scientifique CNRS

Partners: Centre National de la Recherche Scientifique CNRS (FR), Commissariat a l’Energie Atomique et aux Energies Alternatives (FR), Universiteit Gent (BE), Politecnico di Torino (IT), INESC ID – Instituto de Engenhariade Sistemas e Computadores, Investigacao e Desenvolvimento em Lisboa (PT), Barcelona Supercomputing Center Centro Nacional de Supercomputacion (ES), Hewlett Packard Enterprise Belgium (BE), Albora Technologies SL (ES), Argotech AS (CZ), Universita Degli Studi di Verona (IT), Technische Universitat Berlin (DE)
Funded by: European Commission Horizon Europe Framework Programme (HORIZON), within the Digital and emerging technologies for competitiveness and fit for the green deal (HORIZON-CL4-2022-DIGITAL-EMERGING-01), topic Ultra-low-power, secure processors for edge computing (RIA) (HORIZON-CL4-2022-DIGITAL-EMERGING-01-01)
Grant Agreement No.: Project 101070238

Follow us

Publications

• F. Pavanello et al., “NEUROPULS: NEUROmorphic energy-efficient secure accelerators based on Phase change materials aUgmented siLicon photonicS,” 2023 IEEE European Test Symposium (ETS), Venezia, Italy, 2023, pp. 1-6, doi: 10.1109/ETS56758.2023.10173974.
• F. Pavanello et al., “Special Session: Neuromorphic hardware design and reliability from traditional CMOS to emerging technologies,” 2023 IEEE 41st VLSI Test Symposium (VTS), San Diego, CA, USA, 2023, pp. 1-10, doi: 10.1109/VTS56346.2023.10139932.

Related readings:

• N/A