A new system created by British researchers enables personal computers to be empowered to provide them with the same capabilities as supercomputers.
Get this technological feat at a much more accessible value than that required to acquire one of these enormous teams, which are currently only available to large companies and organizations.
The innovation, developed by researchers at the University of Sussex, in the United Kingdom, would allow a single personal computer to acquire the ability to simulate brain models of almost unlimited size.
According to a release, scientists used the latest Graphics Processing Units (GPUs), which can be found on the market, to shape the new technology.
A new age
A new age Traditional format supercomputers and quantum computers promise to revolutionize computing in the coming decades.
With capacities and calculation speeds unimaginable in the framework of current technology, they will allow the arrival of a new era in the field of information, relaunching in turn the development of other related sectors such as Artificial Intelligence, robotics or biotechnology.
For example, more advanced supercomputers they are currently able to process more than 400 billion calculations per second.
According to the work of British scientists, the new technology could allow a conventional computer to achieve equivalent performance without requiring large structures.
These new “home supercomputers” would have a practically unlimited capacity to simulate highly complex neural networks, for example to run simulations of mammalian brains that are used in different studies in the area of neuroscience.
In the research, which was published in the journal Nature Computational Science, the experts stressed that the new technology could make new large-scale brain simulation research viable, for example, those aimed at advancing the understanding and treatment of neurological disorders.
Neural networks The innovation is based on previous models, such as the work of the American researcher Eugene Izhikevich, who developed a similar method for large-scale brain simulation in 2006, but with the limitations of the computers available at the time.
Now, British researchers used a similar scheme on a modern GPU installed in a home computer, developing a model of the visual cortex of a macaque monkey. Until now, a neural network of this complexity could only run on a gigantic supercomputer.
The new development not only matches the performance of supercomputers, but also surpasses some of them by up to 35%, both in terms of start-up time and response time in the calculations performed.
At the same time, the solution created requires 10 times less energy than a supercomputer, thus generating an advantage in the field of environmental sustainability.
Considering that large-scale simulations of neural network models are an important tool to improve our understanding of brain dynamics and functions, this new technology represents a great opportunity for researchers, teachers and students to use advanced technologies at their work tables.
At the same time, specialists argue that the availability of greater processing power in home computers may arouse curiosity and interest in these topics in the general public.
“This research is a game changer for computational neuroscience and artificial intelligence researchers, as they can now simulate brain circuits on their local workstations by simulating brain circuits,” Professor Nowotny, Professor of Computer Science at the University of Sussex, explains. at the same time that it allows outsiders to turn their gaming PC into a supercomputer and run large neural networks.
Larger GPU-accelerated brain simulations with procedural connectivity. Knight, J.C., Nowotny, T. Nature Computational Science (2021).DOI:https://doi.org/10.1038/s43588-020-00022-7
Photo: @awmleer and Unsplash.
Eddie is an Australian news reporter with over 9 years in the industry and has published on Forbes and tech crunch.