The day before yesterday, the TOP500 list of global high-performance computers, which was announced twice a year, was released as scheduled. On this new list, Japan’s Fuyue once again won the world’s No. 1 ranking. This is also the fourth time that Fuyue has ranked No. 1 in the world.
According to the published list, Fuyue achieved a computing speed of 4.42 billion times per second, which is about three times the 1.49 billion times of the second-ranked American “apex” supercomputer, and the fourth-ranked Taihu Lake. The 930,000,000,000,000,000,000,000,000th of the light is more than four times, which is really powerful.
However, although these supercomputers are powerful, they are still not enough in front of quantum computers. Recently, Hao Huan, authorized by Tsinghua University, confirmed this to everyone.
At the SC Supercomputing Conference held a few days ago, Fu Haohuan, a professor at Tsinghua University, used China’s Shenwei supercomputing to simulate a quantum computer that can be scaled up to 42 million CPU cores and has a performance of 44 billion quadrillions per second.
How fast is this 44 billion quadrillion times, compared with Fuyue’s 4.422 billion times, it is 10 times as fast. In terms of this property, it takes only 304 seconds on the current quantum computer that traditional computers need to operate for 10,000 years.
However, because it is simulated, it is actually nothing in front of a real quantum computer, because Google’s previous 54-qubit Sycamore quantum computer only needs 200 seconds to complete the performance of traditional computing for 10,000 years, which is already better than This is much stronger.
At the Quantum Summit on November 15, IBM also unveiled the veil of its own new generation of quantum computers. This quantum computer is called “Eagle”. IBM claims that this is the world’s first quantum computer with more than 100 operable quantum computers. The number of bits, as high as 127, is more than twice that of Google Sycamore, and the increase in qubits can increase the performance exponentially, rather than a simple multiple.
According to IBM, it plans to produce a 433-qubit “Osprey” chip and a 1,121-qubit “Condor” chip in 2022, and a computer with more than 1,000 qubits in 2 years. If this is true, the current These supercomputers are really far behind such a computer, and they are completely incomparable.
It can be seen that the development of supercomputing now may not be as good as the direct development of quantum computers, because the performance of quantum computers is simply not comparable to the current traditional supercomputing, what do you think?