IBM Upends Google’s Quantum Supremacy Claim

By Paul Smith-Goodson, Patrick Moorhead - November 5, 2019
Google’s Sycamore Quantum Computer

Quantum supremacy was supposed to be a significant benchmark to signal that quantum computers could finally solve problems beyond the capability of classical computers. The term "quantum supremacy" was first used in a 2012 paper by one of the world’s leading theoretical physicists, John Preskill, Professor of Theoretical Physics at Caltech. Last month, a leaked research paper declared that Google GOOGL +0% had attained quantum supremacy. According to the paper, Google’s 53-qubit quantum computer, called Sycamore, had solved a problem in a few minutes that would take a classical computer 10,000 years to solve.

IBM was the first and loudest to cry foul  

Dario Gil, head of IBM IBM +0% quantum research, described the claim of quantum supremacy as indefensible and misleading. In a written statement, he said, “Quantum computers are not 'supreme' against classical computers because of a laboratory experiment designed to essentially implement one very specific quantum sampling procedure with no practical applications.”

Yesterday, IBM published a paper that backed up that claim. The paper points out that Google made an error in estimating that a classical computer would require 10,000 years to solve the problem. In a discussion last week, Bob Sutor, Vice President, IBM Q Strategy & Ecosystem, also told me that IBM believed the 10,000-year estimate was an overstatement.

According to IBM’s blog, “an ideal simulation of the same task can be performed on a classical system in 2.5 days and with far greater fidelity." The blog post went on to say that 2.5 days is a worst-case estimate. Additional research could reduce the time even further.

Google overstated its 10,000-year estimate based on an erroneous assumption that the RAM requirements for running a quantum simulation of the problem in a classical computer would be prohibitively high.  Google used the time to offset the lack of space, resulting in its estimate of 10,000 years.

IBM used both RAM and hard drive space to run the quantum simulation on a classical computer. Additionally, it incorporated other conventional optimization techniques to improve performance.

Analysis of expected classical computing runtime vs circuit depth of “Google Sycamore Circuits”. The bottom (blue) line estimates the classical runtime for a 53-qubit processor (2.5 days for a circuit depth 20), and the upper line (orange) does so for a 54-qubit processor

There are other factors, as well. The number of computations that can be achieved on quantum computers is limited by how long qubits can maintain their quantum states--usually about a nanosecond. On a classical machine, there is no limit on how many times an algorithm can be run.

Next stop, Quantum Advantage  

It’s time for us to look past quantum supremacy and begin focusing research efforts toward a broader, more realistic capability called quantum advantage. Quantum advantage will exist when programmable quantum gate-computers reach a degree of technical maturity that allows them to solve some, but not all, real-world problems that classical computers can’t solve (or problems that classical machines require exponential time to solve).

It’s accepted that quantum machines will eventually be able to solve problems that classical machines can't solve. However, it's unlikely that quantum computers will replace classical computers. The long-term future of computing will be a hybrid between quantum and traditional machines.

+ posts

Paul Smith-Goodson is the Moor Insights & Strategy Vice President and Principal Analyst for quantum computing and artificial intelligence.  His early interest in quantum began while working on a joint AT&T and Bell Labs project and, during 360 overviews of Murray Hill advanced projects, Peter Shor provided an overview of his ground-breaking research in quantum error correction. 

+ posts

Patrick founded the firm based on his real-world world technology experiences with the understanding of what he wasn’t getting from analysts and consultants. Ten years later, Patrick is ranked #1 among technology industry analysts in terms of “power” (ARInsights)  in “press citations” (Apollo Research). Moorhead is a contributor at Forbes and frequently appears on CNBC. He is a broad-based analyst covering a wide variety of topics including the cloud, enterprise SaaS, collaboration, client computing, and semiconductors. He has 30 years of experience including 15 years of executive experience at high tech companies (NCR, AT&T, Compaq, now HP, and AMD) leading strategy, product management, product marketing, and corporate marketing, including three industry board appointments.