An IBM research team has developed a new kind of optical modulator that converts a digital electrical signal carried on a wire into a series of light pulses that can be carried on a silicon nanophotonic waveguide. Essentially what this means is that going forward, a series of multicore computer chips could be connected via these modulators.
The new technology aims to enable a power-efficient method to connect hundreds or thousands of cores together on a tiny chip by eliminating the wires required to connect them. Using light instead of wires to send information between the cores can be 100 times faster and use 10 times less power than wires.
To see how it works, check out the video:
Giant supercomputers of today will, in the future, shrink in size, thanks to this development. Or perhaps we have humongous computing power in our laptops. From a broadband perspective, we can think about putting optical routing on a chip, lowering costs and power requirements, and have a major impact in terms of boosting bandwidth.
Haven’t they been talking about light relay breakthroughs for the better part of a year now? They should be using it in the next Intel processors by now, right? =P
Well they have been talking about it for a while, and lots of people working on it, but looks like these guys have finally done it at a level where this technology can be commercialized.
The video no longer works!
It’s working on our end, Sunny — maybe try again?
If it is going to be this way further, we will have good computers in our mobile phones. Such a tendency still is progressing. I am greeting
a 100-1000 times reduction in modulator size has NOT been the factor holding up integration of optics w/ Electronics. In addition to this on/off switch, you still need a light source (laser), a receiver to interpret it back in to electronic signals (the modulator can be used for this function as well but may be inefficient) and the supporting electronics (which needs to go as fast as the optics it is driving so no savings there perhas not even power savings). All these incremental improvements help, I just do not think this is “the” one. Last fall’s silicon hybrid laser (UCSB + Intel) is a bigger impact breakthrough… still miles to go before we have something that is easily integrate with the rest of processor/electronics and used just like we use electrical I/O or interconnects… damn photons eh – they are lean and mean and go fast and can carry tera/peta/bytes of information but are really hard to get going (laser) or turn a corner (switch/shutters) or easily transfer the information they are carrying in to electrons.
rohit,
while you are right about that, this is part of an ongoing shift that we have talked about in the past. i think it be interesting to see how these technologies – various breakthroughs – come together.
i am getting increasingly impressed by the progress being made on this front.