nanophotonic
Latest
IBM keeps light pulse bandwagon rolling, uses 'em for chip-to-chip communication
Lenovo loves its red mousing nipple, Apple digs its aluminum and IBM adores those light pulses. Nearly two full years after we heard this very company touting breakthroughs in science thanks to a nanophotonic switch, in flies a similar technique from Yorktown Heights that could "greatly further energy efficient computing." As the story goes, gurus at IBM have figured out how to replace electrical signals that communicate via copper wires between computer chips with tiny silicon circuits that chat using pulses of light. The device is called a nanophotonic avalanche photodetector, and according to Dr. T.C. Chen, this kind of embedded optical interconnection makes the "prospect of building power-efficient computer systems with performance at the Exaflop level" something that could be seen in the not-so-distant future. Reportedly, the avalanche photodetector demonstrated by IBM is the world's fastest device of its kind, able to receive optical information signals at 40Gbps and simultaneously multiply them tenfold. We know that's over some of your (read: our) heads, but there's a sufficiently nerdy video after the break that gets right down to the whos, whats, whys and wheres.
IBM's light pulse love affair continues with tiniest nanophotonic switch
Merely months after IBM first inserted silicon nanophotonics into our memory bank, and just weeks after we drooled all over ourselves reading about its green optical link, the mega-corp has chosen St. Patrick's Day to trumpet the development of the "world's tiniest nanophotonic switch." The device, which boasts a footprint "about 100 times smaller than the cross section of a human hair," is said to be a vital part of creating an on-chip optical network. More specifically, it's bringing the gurus behind it one step closer to conjuring up multi-core CPUs that transmit data with light pulses rather than relying on electrical signals on copper wires. This particular switch would essentially divert traffic within the network, ensuring that "optical messages from one processor core could efficiently get to any of the other cores on the chip." Keep on livin' the dream IBM -- just ping us when this stuff is anywhere near ready for the commercial market.
