Scientists Cross Photonic Hurdle to Faster Computer Chips
Scientists have made yet another big step toward replacing silicon chips with photonic circuits.
Research groups at the California Institute of Technology and the University of California, San Diego claim to have discovered a path to light signals on a silicon chip from reflecting backwards and interfering with its operation.
Transporting information in computer chips relies on a technique that allows current to flow in only one direction that avoids interference through a scattered beam that would make a signal and a chip unstable. So far, the model used in silicon chips, which relies on a diode to isolate electric signals, has not been possible. However the researchers say they have developed a metallic-silicon optical waveguide system to channel light so it travels in different patterns depending on its propagation direction. When traveling forward, the pattern is symmetric, but asymmetric when reflected backwards. According to a paper published in the journal Science, backscattered light is dissipated as a result.
"This discovery will help to realize a long-term goal of combining electronics with photonics to enable scalable, energy-efficient and cost-effective technology that will have a tremendous impact on such information systems as supercomputers, the Internet, and data centers," said Yeshaiahu Fainman, professor and chair of the UC San Diego Department of Electrical and Computer Engineering. "Computer technology will be able to handle a lot more data, faster and at lower cost, which will benefit large-scale business and government users as well as gadget-loving consumers."
Probably as a result of being sucked into your computer's black hole!
Folding@home cluster running on light cures death.
Sign me up!
This is still in the realm of applied physics, not "technology", so I'd hold your undue pessimism until later.
Circuit: http://en.wikipedia.org/wiki/Electronic_circuit
Transistor: http://en.wikipedia.org/wiki/Transistor
An electronic circuit is a collection of interconnected electronic elements. A transistor is one possible element. Or was this a troll?
Probably as a result of being sucked into your computer's black hole!
Of it though does bring to "light" many ideas of use to say at times, like that of lighted sources for use. And with as well is it more say "power" frinedly. And if so, What little more of use from what it normally would be have for anything of use?
Then on say the sack of more, what more of is better then more of off what more would be?
Otherwise sounds interesting and for wait on a avialible use to say on the idea, yes? Price outside of course, but still not at the sametime.
Otherwise something that does something with itself to be more or less of something that still wasnt much of something in some way, probably finds its fancy, or not, usually would imply dumb, but maybe for "strings" of articles then one finds some smarts?? Yes??
Light travels one direction... pattern reflected asymmetrically... reflected pattern travels from CPU to disco ball... disco party in room while gaming.
ujelly?
I agree, however photonic based comptuers wil totally make sense for some applications two words come to mind space exploration , one of the biggest hurdles to the space program is gettign newer tech that is reliable , and every thign i've read on photonic computing suggest it would take alot more punnishment than current transistor technologies.
I believe, but don't quote me,that the issue before with back scatter was canceling out/distorting the data stream by changing the phase. The information in the system is probably in the back scatter and I imagine the closer the back scatter function is to random (and therefore the less information carried), the better the system works.
The barriers to space exploration (other than adequate funding) are in the short term a lack of Pu-238 (the only suitable nucleide) for the RTG's that power long range probes beyond practical solar power range, and, in the long term, cheap-which means reusable- ferry vehicles to orbit as well as better, perhaps more exotic engine technologies for faster and longer range travel so that probes can be sent more consistently and technological advance on earth can be maximized sooner (i.e. no 90's tech orbiting Jupiter in this decade). We desperately need to develop powerful and efficient designs that leverage the far greater energy storage of nuclear power for these applications. Computing solutions, as of right now, are perhaps the only non-barrier. Current technology is far more than powerful enough, can be made very energy efficient (especially compared to the other equipment on the mission), is resilient, and because it works 'in the background' is easy to shield from cosmic radiation. The sensitivity of the exposed sensor to cosmic radiation is far more of an engineering concern. Due to the specialized tasks (no or little multitasking, no rendering especially for probes) the computers aboard these missions probably need to be no more powerful than 1/10 of the processing power of an iPhone. Any legitimate number crunching analysis of the data can be done by supercomputers on earth once its beamed back.
I hate the lack of an edit button. So hard to read in the little box b4 posting.