A new exascale supercomputer in China, developed by Sunway, has trained an AI model dubbed bagualu with 174 trillion parameters.
Scientists in China say they have been able to run an artificial intelligence model as sophisticated as a human brain on their most powerful supercomputer, a report from the South China Morning Pos t reveals.
According to the report, this puts China’s Newest Generation Sunway supercomputer on the same level as the U.S. Department of Energy’s Frontier, which was named the world’s most powerful supercomputer earlier this month.
As a point of reference, Frontier is the first machine to have demonstrated it can perform more than one quintillion calculations per second.
The talk is provided on a Free/Donation basis. If you would like to support my work then you can paypal me at this link:
https://paypal.me/wai69
Or to support me longer term Patreon me at: https://www.patreon.com/waihtsang.
Unfortunately my internet link went down in the second Q&A session at the end and the recording cut off. Shame, loads of great information came out about FPGA/ASIC implementations, AI for the VR/AR, C/C++ and a whole load of other riveting and most interesting techie stuff. But thankfully the main part of the talk was recorded.
Summary: Researchers present an all-atom molecular dynamic simulation of synaptic vesicle fusion.
Source: Texas Advanced Computing Center.
Let’s think for a second about thought—specifically, the physics of neurons in the brain.
Qubits are a basic building block for quantum computers, but they’re also notoriously fragile—tricky to observe without erasing their information in the process. Now, new research from the University of Colorado Boulder and the National Institute of Standards and Technology (NIST) could be a leap forward for handling qubits with a light touch.
In the study, a team of physicists demonstrated that it could read out the signals from a type of qubit called a superconducting qubit using laser light, and without destroying the qubit at the same time.
The group’s results could be a major step toward building a quantum internet, the researchers say. Such a network would link up dozens or even hundreds of quantum chips, allowing engineers to solve problems that are beyond the reach of even the fastest supercomputers around today. They could also, theoretically, use a similar set of tools to send unbreakable codes over long distances.
Canadian quantum computer company, Xanadu, has used its photonic quantum computer chip, Borealis, to solve a problem in 36 microseconds versus classical supercomputers taking 9,000 years. This is 7,884 trillion times faster. This runtime advantage is more than 50 million times larger than that of earlier photonic demonstrations.
An earlier quantum photonic computer used a static chip. The Borealis optical elements can all be readily programmed.
Continue reading “Xanadu Photonic Quantum Chip Solves Trillions of Times Faster” »
The first exascale computer has officially arrived.
The world’s fastest supercomputer performed more than a quintillion calculations per second, entering the realm of exascale computing. That’s according to a ranking of the world’s speediest supercomputers called the TOP500, announced on May 30. The computer, known as Frontier, is the first exascale computer to be included on the biannual list.
Ars Technica’s Chris Lee has spent a good portion of his adult life playing with lasers, so he’s a big fan of photon-based quantum computing. Even as various forms of physical hardware like superconducting wires and trapped ions made progress, it was possible to find him gushing about an optical quantum computer put together by a Canadian startup called Xanadu. But, in the year since Xanadu described its hardware, companies using that other technology continued to make progress by cutting down error rates, exploring new technologies, and upping the qubit count.
But the advantage of optical quantum computing didn’t go away, and now Xanadu is back with a reminder that it still hasn’t gone away. Thanks to some tweaks to the design it described a year ago, Xanadu is now able to sometimes perform operations with more than 200 qubits. And it has shown that simulating the behavior of just one of those operations on a supercomputer would take 9,000 years, while its optical quantum computer can do them in just a few-dozen milliseconds.
This is an entirely contrived benchmark: Just as Google’s quantum computer did, the quantum computer is just being itself while the supercomputer is trying to simulate it. The news here is more about the potential of Xanadu’s hardware to scale.
