The Cerebras CS-1 chip, containing 1.2 trillion transistors, is 200 times faster than a supercomputer.
The chip was so powerful in analyzing more than a million variables from fluctuating temperatures to 3D air movement that it was able to show what could happen faster than in real time.
Carbras CS-1 was developed in partnership with the US Department of Energy’s National Energy Technology Laboratory, which has been described as “the most powerful artificial intelligence computing system in the world.” The number of transistors running on the chip is about 22 times that of the recently announced Nvidia A100 80GB chip, designed for the latest supercomputers.
“This work opens the door to major breakthroughs in scientific computing performance,” said a blog for Carbras chip, adding that “Carbras CS-1 is the first system ever to demonstrate sufficient performance for faster-than-real time simulation. This means that when the chip is used to simulate a terminal.” Energy based on data on current operating conditions, it can tell you what will happen in the future faster than the laws of natural physics produce the same result. ”
The chip’s immense computing power will be used to train neural networks and perform high-fidelity simulations of real-world scenarios, such as simulating the best way to land a helicopter on a surface by modeling airflow patterns around its propellers.
The company has also tapped into a second-generation chip containing 2.6 trillion transistors, which will pave the way for more complex simulations in the real world.
This invention raises once again questions about computer simulation in the manner of “Matrix”, which is the movie that talks about the existence of a chip that can simulate many conditions and choose the best ones, that can predict the future and the nature of existence by simulating millions of possibilities.
The simulation hypothesis, developed by philosopher Nick Bostrom in 2003, states that in the future enormous amounts of computing power will be used to run a realistic simulation of the universe.
This hypothesis has since been popularized by Elon Musk, who claimed that there is a 99.99% chance that the universe we live in is a computer simulation.
“40 years ago we had a pong game,” Musk said at a 2016 conference, “Now we have a realistic 3D simulation game with millions of people playing simultaneously, which means millions of possibilities … if you assume you can apply this to life, then Games will become indistinguishable from reality. ”
Realistic universe-wide simulations are still elusive and may not be possible with conventional computers. Nevertheless, advances in ultra-powerful quantum computers could provide a potential pathway for their creation.