IBM is marking its 100th anniversary by celebrating its record in technology innovation. It created dynamic RAM, the disk drive and the magnetic strips used on credit cards, among many other inventions. It is one of the most inventive companies in the world.
But the computing industry is moving to a new future as disruptive and as radical as the era that began with the introduction of silicon chips, and that future is quantum computing. These are systems that use the behavior of subatomic particles to conduct calculations now performed with transistors on a chip.
This future may be anywhere from 10 to 20 or more years away. But if the potential of quantum computing is fully realized, it may trigger a development rush in chip and hardware design reminiscent of what Silicon Valley experienced decades ago.
"Think about the game changer we're now approaching," said Bernard Meyerson, IBM's vice president of innovation and a fellow at the company. It's Meyerson's job to help make sure that IBM's last 100 years aren't remembered as its best ones. That's one of the reasons he talks about the changes in the chip world.
Following Moore's Law, and then shrinking by another factor of 10 from the leading-edge processors of today, these transistors will be so small that "you cross into a quantum mechanical regime of operation -- there's no precedent for that," Meyerson said.
But even once this shrink limit is reached, in about 10 years, progress will continue as engineers build tightly coupled systems with massive levels of integration on blocks of chips, as well as make improvements in memory, caching and speed processing, Meyerson said.
Those advances will extend the time frame to 20 years. But after that, "you better have a helluva trick up your sleeve," said Meyerson. One of those tricks may be quantum computing.
IBM researchers have studied the theory and potential of quantum computing for years, and more recently they have been experimenting with the concepts, said Bill Gallagher, the senior manager of quantum computing at IBM Research.
"It's one of our most significant fundamental research projects now, and may be one of the largest fundamental ones," said Gallagher, There's been "good progress, but a long way to go," he said.
An ordinary computer is a collection of bits that can either be a 0 or a 1. But quantum bits can hold those states, 0 and 1, simultaneously. Instead of doing a calculation one after the other, the processing power in a quantum computer can increase exponentially. Two quantum bits, or qubits, can hold four distinct states, which can be processed simultaneously, three qubits can hold eight and 10 qubits can hold 1,024 states. In time, researchers expect machines with thousands of qubits.
But the subatomic world of quantum computing is daunting. Approaches to maintaining "quantum coherence," a stable state for the interaction of atoms and electrons running calculations, include processing at temperatures near absolute zero, or -459.67 degrees Fahrenheit, to reduce thermal interference, and using superconducting metals. Lengthening the amount of time that a coherent state can be maintained is one of the challenges facing researchers.