IBM ups pressure with quantum computer prototype
IBM is increasing the pressure on Google in the battle to commercialize quantum computing technology.
Quantum computers hold the promise of being able to solve difficult problems from fields such as chemistry and material science that are currently beyond the reach of the most powerful conventional supercomputers. They may also one day render some current encryption techniques obsolete.
IBM said Friday it has created a prototype 50 qubit quantum computer. A machine this size is believed to be close to the threshold at which it could perform tasks beyond the reach of conventional supercomputers — a major milestone in computer science that researchers in the field refer to as “quantum supremacy.”
In a statement, IBM said it “aims to demonstrate capabilities beyond today’s classical systems” with quantum systems this size.
Friday’s announcement puts IBM in a neck-and-neck race with Google, which has said that it plans to show a similarly sized machine capable of achieving this milestone by the end of the year.
Today’s quantum computers remain too small and too errorprone to outperform conventional supercomputers at most tasks, but the technology is advancing rapidly. A number of companies — including IBM, Google, Microsoft, Canada’s D-Wave Systems Inc. and California-based startup Rigetti Computing — are pushing to create machines that businesses can use.
The Institute for Quantum Computing at the University of Waterloo also is a leader in quantum computing research.
Jonathan Breeze, a research fellow working on advanced materials at Imperial College London and not affiliated with any of the companies developing quantum computing, said that practical applications of quantum computers will depend largely on being able to reduce the error-rate in their calculations.
“There is much debate about how errors scale with the number of qubits,” Breeze said. If errors grow exponentially as more qubits are added, as some suspect, then the technology may fall far short of expectations. “The exciting thing is that the technology is now developing at such a rapid rate that we should be able to answer that question soon,” Breeze said.
While IBM and its rivals are keen to lead the breakthroughs in quantum computing, they are also increasingly eager to win over new customers. Google has offered science labs and artificial intelligence researchers early access to its quantum machines.
IBM said Friday it’s making a 20 qubit quantum computer available to paying customers through its cloud computing platform by the end of 2017. The company had previously offered customers access to a 17 qubit machine. It also gave researchers access to run experiments on both a five and a 16 qubit quantum computer over the internet for free.
The financial industry has also been taking a strong interest in the power of quantum computing. The second investment made by the CME Group’s venture arm ever made was in 1QB Information Technologies Inc., a quantum computing software company in Vancouver and backed by the Royal Bank of Scotland. Goldman Sachs is an investor in D-Wave Systems.
A conventional computer uses tiny semiconductors to store information in a binary format — either a 0 or 1 — called bits. Quantum computers use different techniques — often involving materials cooled to temperatures colder than those found in outer space — to create processing units, called qubits, which exhibit quantum mechanical properties. Qubits, for instance, can represent both a 0 and 1 simultaneously.
And unlike standard bits, which are meant to process information discretely from one another, quantum bits affect one another. These properties would theoretically give a quantum computer vastly more power than a conventional machine. But the qubits in all of today’s quantum computers are able to remain in a quantum state for only fractions of a second. As they fall out of this state, errors creep into their calculations. IBM said Friday that it has managed to keep its qubits in a quantum state for 90 microseconds, which is twice what it could do in smaller systems unveiled six months ago.