Forefront | Dean of Intel Research: It takes ten years for quantum computing to meet practical problems

no thumb

In the research of quantum computing, Intel is in the first echelon with IBM and Google.When IBM and Google struggled for the issue of quantum hegemony, Intel looked very Buddhist.According to Rich Uhlig, Dean of Intel Research, quantum hegemony is not important. What matters is quantum practicality.Google’s recent paper published in Nature has announced the realization of quantum hegemony.Google claims that its quantum processor takes about 200 seconds to complete the task that traditional computers have to complete in 10,000 years. On a specific task, quantum computers greatly surpass the computing power of classical computers to achieve “quantum superiority.”IBM questioned Google’s claim that “quantum hegemony” was flawed and manipulated the test results.Rich Uhlig acknowledged Google’s achievement and called it “exciting news.”Uhlig said that to prove quantum hegemony, two conditions are required: 1. It has to find a sufficiently complex subject; 2. Quantum computing surpasses traditional computers in this topic.Google found such a topic and proved the superiority of quantum computers.But Uhlig also pointed out that the problem of Google research is almost non-practical, and we will let quantum computing solve practical problems in the future.The topic that Google researchers have assigned to quantum computers is “to prove that the number produced by a random number generator is indeed random.”This is an extremely complex purely mathematical problem with little practical value.Uhlig believes that it takes another ten years to encounter practical problems. Quantum chemistry and material modeling are potential directions. Max-Cut algorithm optimization and cracking encryption algorithms require higher qubits.This is much more conservative than the time IBM predicted.Norishige Morimoto, dean of the IBM Tokyo Research Institute and vice president of the world, said that the commercialization of quantum computing only takes three to five years.To visually understand how quantum realism can be achieved, Intel uses high-performance quantum simulators to predict that quantum computers outperform supercomputer nodes when solving Max-Cut optimization problems.Max-Cut is an algorithm whose complexity increases exponentially with the increase of the number of variables. It is widely used in various fields from traffic management to electronic design. Therefore, Max-Cut is selected as a test case.In the research of quantum computing, Intel has laid out two lines at the same time: superconducting qubit chip and silicon spin qubit chip.According to Uhlig, Intel currently invests more resources in silicon spin qubit technology.The size of the spin qubit is much smaller than that of the quasi-quantum, and it has a larger miniaturization advantage than the superconducting qubit.Quantum computing is still at a very early stage, Uhlig said it is still in the “first mile of the marathon run”, and the real commercialization is still far away.Current quantum computing only reaches dozens of qubits, and millions of qubits may be needed to solve complex problems.However, the industry has begun an initial attempt to commercialize quantum computing.Both Microsoft and IBM have recently announced plans to open quantum computing cloud services.However, these services are not intended to solve problems, but are more universal.Microsoft said developers and businesses should now be exposed to quantum algorithms and hardware to help the industry understand the benefits of this technology.It is worth mentioning that quantum computing and CPU classic computing are not substitutes for each other, but to solve the extremely complicated problem that classical computing cannot solve.Earlier, Intel CTO Michael Mayberry said in an interview that the development of quantum computing will not make existing other computing methods obsolete, and our commonly used CPUs are still useful.