At Intel Corp, Anne Matsuura is the director of Quantum & Molecular Technologies, and Mike Davies is the director of the Neuromorphic Computing Lab. Together they research ways to create computers that can calculate in a realm close to human reasoning and simultaneous processing. This rather than instruction-based reasoning as computers do today. Or in a single, line by line computing, one task at a time, process of today’s computers. They call these studies Neuromorphic Computing, and Quantum Computing.
They are introducing these groundbreaking technologies at the Intel Labs, Intel Innovations 2022 press release and event Wednesday. The event will showcase a Quantum Software Development Kit (SDK) and the Loihi Neuromorphic Research Chip and attendant boards and modules.
Anne Matsuura has undertaken a different kind of computer, with a different kind of software. Quantum Computers use qubits, instead of binary, on-off switches called bits we use in digital computers. The Quantum computer sees in simultaneous actions, whereas digital computers see only two variables. For example, a digital computer sees a coin as either heads or tails, one or the other, while a quantum computer sees both heads and tails at the same time, much like we would.
For this type of computer to work there needs to specialized software. Since there are relatively few quantum computers, the quantum SDK is designed as a platform to be simulated on a digital computer. The SDK uses a quantum runtime environment optimized for executing hybrid quantum-classical algorithms. That would mean that quantum simultaneous computing can be simulated in real-time for algorithms for something like financial modeling.
The goal of the Quantum SDK in its beta release is to advance the science behind Quantum computing, create new instances of software that can use the quantum process, and eventually create devices that use quantum computing software. This is a giant leap forward to computers that can see the world in holistic fashion rather than linear.
Mike Davies is heading up the efforts to create a Neuromorphic computing intuitive system with the Loihi 2 second generation research chip. The Loihi 2 is a neural chip that uses the Lava open-source software framework. Neural chips are meant to mimic thought process in the brain, much like the neurons and synapsis handling millions of computations a second.
The Loihi 2 is able to work in multiple instances, where four Loihi 2 chips are used in a series on one module. The modules can then be set up in series that can carry dozens of interwoven systems, creating an ecosystem that would mimic the intuitive structures of thousands of algorithms inferences per second. The development board that can be installed in series is the Kapoho Point, an eight chip, two sided (4 chips on each side) Loihi 2 system.
The boards were being sent out to the member community involved in the research, Air Force Research Laboratory was the first to receive the Kapoho Point boards and have startling results in solving optimization problems. The boards can solve problems with up to eight million variables, while 10 times the speed of other neural systems, and are running on 1,000 times lower energy. This could mean that an AI system could run inference choices in computing at speeds that mimic brainwaves.
Research teams had over 700 researchers, 500 Ph.D.’s 100+ principal engineers and fellows, and 30+ technical disciplines. Research began five years ago and is ongoing. There is a research community of 180 companies, aiding in these studies.
