Performance exceeds hanging Suzuki 241 million times! Google once again demonstrates the quantum advantage!

Now, in a paper published online by Google researchers on the arXiv, the company's latest technology "surpasses the capabilities of existing classical supercomputers.

 

Google announced that it has developed a quantum computer that can immediately perform calculations that would take the best existing supercomputers 47 years to complete - a breakthrough that undoubtedly establishes that experimental quantum machines can outperform their traditional classical rivals.

 

 

"Phase transition in Random Circuit Sampling"

 

 

Google Quantum Computer

 

Proponents of quantum computers argue that the technology relies on a special state of quantum physics to create powerful machines that can even fight climate change and create breakthrough medicines; however, they also have the potential to disrupt today's encryption systems, making them a threat to national security.

 

Four years ago, Google claimed to be the first company to achieve "quantum computing superiority": a landmark node where quantum computers surpass existing machines. This was challenged at the time by competitors who argued that Google was exaggerating the differences between its machines and traditional supercomputers.

 

This time, the new paper Phase transition in Random Circuit Sampling shows a more powerful device that aims to end that debate.

 

While the 2019 machine has 53 quantum bits, Google's next-generation device will have 70 quantum bits. Adding more quantum bits can exponentially increase the power of a quantum computer, meaning the new machine is 241 million times more powerful than the 2019 machine.

 

Frontier, the world's leading supercomputer, takes 6.18 seconds to match the calculations of Google's 2019 53-quantum-bit computer; by comparison, Frontier takes 47.2 years to match its latest quantum efficacy, the researchers said.

 

 

Noise from random circuit sampling

 

 

Noise-induced phase transition

 

 

Estimated simulated computational cost with Frontier's specifications used for estimation

 

 

demonstration of quantum dominance. The results show that the current experiment far exceeds the memory capacity of Frontier.

 

Google's paper shows how larger quantum computers can manage "noise" - interference that can disrupt the fragile state in which quantum bits operate, making it difficult to continue computations.

 

The experiment proposes a new random circuit sampling scheme with 70 quantum bits and 24 cycles, with an estimated fidelity of 1.7×10-3, implying an increase in circuit size of about 60 percent for the same fidelity. In the paper, the researchers say, "We conclude that our demonstration is firmly in the regime of quantum computing beyond the classical one. And some rival machines are measuring on randomized tasks." Critics say this favors quantum computers and lacks any practical value beyond academic research.

 

The industry also has high praise for the results.

 

Steve Brierley, chief executive of Cambridge-based quantum company Riverlane, said, "This is an important milestone. The debate about whether we have reached or can indeed reach quantum computing superiority is now settled."

 

Sebastian Weidt, chief executive of Brighton-based start-up Universal Quantum, said quantum computers need to demonstrate more practical capabilities. "This is a very good demonstration of the quantum advantage. While it's a great achievement academically, the algorithms used don't really have practical applications in the real world. We really have to move into the era of practical quantum computing - an era where quantum computers with thousands of quantum bits really start to provide value to society in a way that classical computers never could. "

 

Reference Links:

[1]https://www.telegraph.co.uk/business/2023/07/02/google-quantum-computer-breakthrough-instant-calculations/

[2]https://arxiv.org/abs/2304.11119

[3]https://www.msn.com/en-us/news/technology/breakthrough-quantum-computer-instantly-makes-calculations-that-take-rivals-47-years/ar-AA1dkHu6

2023-07-04