IBM is working on a new supercomputer inspired by the human brain.
Experts have long maintained that the human brain is the most powerful supercomputer in the world, and now a team of researchers is building a new, neural network-based supercomputer inspired by the organ, writes Danny Palmer for ZDNet.
IBM teams up with nuclear security facility
IBM is working in conjunction with the Lawrence Livermore National Laboratory (LLNL), a federal government research facility in California. The facility is meant to ensure the safety, security, and reliability of the United States nuclear deterrent.
Scientists from the two organizations are working on what they say is a “first of a kind” supercomputing platform inspired by the human brain, which will be used for deep learning. the neural-network will use IBM TrueNorth neurosynaptic computer chips.
The idea behind the processors is that they will help computers perform cognitive tasks more efficiently than normal chips. Such tasks include recognizing patterns and processing senses.
TrueNorth chips to provide incredible efficiency
A single TrueNorth processor is made up of 5.4 billion transistors fixed together to make 1 million digital neurons that use 256 million electrical synapses to communicate with each other. This allows the chips to be incredibly efficient.
The platform will use 16 TrueNorth chips in total, processing the equivalent of 16 million neurons and 4 billion synapses while only using 2.5 watts of power, the same amount of energy as a tablet.
“The low-power consumption of these brain-inspired processors reflects the industry’s desire and a creative approach to reducing power consumption in all components for future systems as we set our sights on exascale computing,” says Michel McCoy, LLNL program director for Weapon Simulation and Computing.
Researchers excited by potential uses of technology
Scientists at the Lawrence Livermore National Laboratory will use the new system to “explore new computing capabilities” concerned with the National Nuclear Security Administration’s (NNSA) missions in cybersecurity, control of US nuclear weapons, and potentially the management of non-proliferation agreements.
In order to do so the researchers will use the NNSA’s Advanced Simulation and Computing (ASC) program. The software will be used to evaluate machine learning applications and deep learning algorithms, while carrying out feasibility studies on the program.
“Neuromorphic computing opens very exciting new possibilities and is consistent with what we see as the future of the high-performance computing and simulation at the heart of our national security missions,” says Jim Brase, LLNL deputy associate director for data science, who says that neuromorphic computing could “change how we do science”.
“The delivery of this advanced computing platform represents a major milestone as we enter the next era of cognitive computing,” says Dharmendra S. Modha, IBM Fellow and chief scientist of brain-inspired computing at IBM Research.
The collaboration will help both LLNL and IBM to push brain-inspired computing “to enable future systems that deliver unprecedented capability and throughput, while helping to minimize the capital, operating and programming costs,” according to Modha.
Computer scientists at Lawrence Livermore will work with IBM Research, partners from across the Department of Energy, and universities in order to push neurosynaptic architecture forward.
These are exciting times for computer scientists, and it might not be long before we see a major breakthrough in terms of the power and efficiency of computer systems. As it stands the U.S. nuclear program is controlled by seriously outdated technology, so it’s good to see the agencies taking steps towards changing that situation.