Light shines in quantum-computing arena.

For a few years, scientists have been predicting that computers exploiting the quantum properties of matter will carry out computations billions of times faster than today's supercomputers. Yet the technical challenges are so daunting that such quantum computers may not be feasible for decades.

Now, researchers have developed a new, yet less exotic computing method that may be as good as quantum computing for certain tasks, such as searching databases. The method relies entirely on classical physics, say lan Walmsley and his colleagues of the University of Rochester in New York. To convert their ideas into hardware, the Rochester scientists have built an optical device and successfully demonstrated the method.

The group reported its results at the Lasers and Electro-Optics/Quantum Electronics and Laser Science conference in Baltimore last week.

Researchers expect quantum processors to work incredibly fast thanks in part to particles' wavelike interactions, including interference. The processors would take advantage of another, stranger effect known as entanglement, in which two or more particles share one quantum state (SN: 8/21/00, p. 132).

Walmsley and his colleagues suspected that classical interference such as that between intersecting light waves, could lead to a computation method analogous to the interference aspect of quantum computing. Optical computers already exist, but their calculations take advantage of different intensities of light. Says Walmsley: "They don't exploit the fact that waves interfere."

His group built a prototype of an interference-based optical computer. Then, they searched a database using a famous quantum-computing algorithm invented by Lov K. Grover of Lucent Technologies' Bell Labs in Murray Hill, N.J. (SN: 11/29/99, p. 334). In 1997, Grover theorized that a quantum system relying on his interference-based algorithm would require only a single lightning-fast operation to find a target item.

The Rochester researchers simulate a 50-item database by vibrating a crystal of tellurium dioxide so that one small region of the crystal expands. Then, they fire a sub-trillionth-of-a-second pulse of light through a diffraction grating, which spreads the light into beams of various frequencies. The expanded region slightly alters the speed of the beam passing through it. The light rays exiting the crystal meet up and interact with a copy of the original pulse.

Because of interference, only the altered beam corresponding to the targeted database item passes on to a detector, the researchers say. The Rochester scheme thus completes its search in a single operation.

David A. Meyer of the University of California, San Diego and some other theorists Nad previously argued that a computer using classical physics can perform as well as any quantum computer in some calculations that involve only interference. Until now, however, there's been no actual interference-based classical computer for testing the idea. The Rochester team's work has experimentally verified this theory, Meyer says.

Walmsley says that his group is preparing experiments to see if the new approach might even work for some quantum calculations that are expected to require entanglement to be carried out quickly and efficiently.