(Image courtesy of SpiNNcloud)
Sandia National Laboratories has announced a partnership with AI and neuromorphic computing company, SpiNNcloud. Leveraging a next generation system, SpiNNaker2, this partnership seeks to advance neuromorphic architecture.
Funded through NNSA’s Advanced Simulation and Computing (ASC) program, this work will explore how neuromorphic computing can be leveraged for the nation’s nuclear deterrence missions. “As scientific computing and artificial intelligence systems push towards the limits of power consumption and cooling, it is imperative that the ASC program continues to push the boundaries of computing by investigating and helping to develop novel algorithms and hardware solutions,” said Thuc Hoang, Director of the ASC program. “Neuromorphic computing offers one approach to addressing these demanding problems.”
SpiNNaker is a contraction of ‘Spiking Neural Network Architecture,’ which is a brain inspired neuromorphic computer for large-scale, real-time modeling of brain-like applications. This technology can simulate large brain-like networks to enhance researchers’ understanding of the brain, as well as provide a framework to test the boundaries of current computing capabilities.
This computing platform was accomplished by building a neuromorphic architecture using a large-scale mesh composed of Arm Cores. These small and lightweight central processing units (CPU) are most notably used in cell phones due to their low-power footprint. SpiNNaker2 builds upon this scalable brain-like mesh and enhances the architecture with state-of-the-art distributed acceleration. The flexibility provided by its reconfigurability, scalability afforded by its real-time, large-scale mesh, and native support for hybrid acceleration of symbolic spiking and deep neural networks make SpiNNaker2 a unique computing platform. The system is suitable for integrating brain-inspired algorithms into a wide variety of computationally intensive problems.
“I have been fascinated by the brain for the past 20 years, and during this time, I have learned what to copy and what not to copy from the brain to design efficient systems,” said Dr. Christian Mayr, scientific lead at SpiNNcloud. “SpiNNaker2 is the outcome of more than a lifetime of research, and it is the living proof of efficiency through practical brain inspiration.”
As an emerging technology, it has unexplored potential applications. These include heterogeneous random walk applications, optimization approaches to solve computationally intensive problems, quantum emulation, energy-efficient deployment of large-scale networks, and acceleration of new algorithms such as event-based machine learning. Researchers at Sandia also hope to explore its applications in data analysis from scientific instruments, neural networks for remote sensing inference, enabling physical security processing at the sensor, performing scientific computing workloads, and physics informed neural networks.
“The AI and Machine Learning capabilities offered by a system like SpiNNaker2, has many potential impacts to innovation here at the laboratories,” said Rob Hoekstra, senior manager of the Extreme Scale Computing group at Sandia. “Continuing research into neuromorphic advantages over CPUs and GPUs [graphics processing units] for certain algorithms, could prove advantageous as large-scale neuromorphic computing is realized.”
Other commercial applications of this technology include drug discovery, deep neural network training, quantum computing emulation, optimization, large language models, and streaming edge AI. “Sandia’s vision of transforming the world with large-scale neuromorphic computing resonates with our expertise in building unparalleled systems,” said Hector Gonzalez, co-founder and co-CEO of SpiNNcloud. “Our technology is flexible enough to selectively apply brain-inspiration when it makes sense, and its efficient event-based operation is crucial to address today’s computational challenges. We are eager to integrate this technology into Sandia’s applications.”
(Image courtesy of SpiNNcloud)
Neuromorphic computing has potential to revolutionize multiple fields of computing. While artificial intelligence and machine learning are the obvious applications, a neuromorphic advantage has been shown for non-cognitive applications as well, such as numerical computing. The availability of large-scale neuromorphic systems fuels the opportunity to explore how computing inspired by the brain can usher in further breakthroughs.
“Exploration into large-scale neuromorphic computing is an exciting frontier for innovation,” said Craig Vineyard, principal member of the technical staff. “The partnership with SpiNNcloud will enable Sandia to push the boundaries of brain-based computing and research the impact algorithmic breakthroughs can have on our mission goals.”
About NNSA: Established by Congress in 2000, NNSA is a semi-autonomous agency within the U.S. Department of Energy responsible for enhancing national security through the military application of nuclear science. NNSA maintains and enhances the safety, security, and effectiveness of the U.S. nuclear weapons stockpile; works to reduce the global danger from weapons of mass destruction; provides the U.S. Navy with safe and militarily effective nuclear propulsion; and responds to nuclear and radiological emergencies in the United States and abroad.
May 8, 2024