Intel reasearchers build LASER on chip

[Cristatus]

Full story here

I highl doubt this will actually take off, but how long would something as new as this take to come down to microcomputer levels. I mean, come on, they've been talking about nano-computing since i was born (1985) (I don't know, but maybe I'm exageratting), but seriously, is this worth taking a note of. And while I have mentioned nano-computing, when can we expect that?

 

[f95toli]

Laser on chips have been around for a long time and are used commerically in optoelectronics, usually III-V materials such as GaN or GaAs are used (in heterostructures). What is new here is that they have managed to do it using silicon, this means that it should be easier to integrate with ordinary electronics.
So it is interesting but hardly revolutionary.

We have had nano-computing for quite some time now, everything on a modern microprocessor is done on the nanometer scale (the latest generation uses 90 nanometer linewidth). Using modern lithography (e-beam) we can make structures that are only a few nm across, the problem is that it is difficult to scale this up to commercial production.

 

[PowerEngineer]

My understanding is that Intel has been putting a lot of work into this. As the article says, they want to replace the electric wire connections to their CPU's with optical connections. 5-10 years?

 

[futuristicmonkey]

Originally posted by: PowerEngineer
My understanding is that Intel has been putting a lot of work into this. As the article says, they want to replace the electric wire connections to their CPU's with optical connections. 5-10 years?

This'll be like DDR2 then, eh?

I mean, DDR2 is fast, compared to DDR1, but it has so-so latencies, which are made up for by said speed.

So, while it may take a little longer to process the photons, the propagation speed (been waiting a while to use that ) of light (even thru anything) is A LOT faster than an electric signal, correct?

Or is this little 2:00 AM thought completely bonkers?

 

[unipidity]

I actually dont know how to calculate information propogation speed of a circuit. But since we are talking about an electric field... surely it would be at c? Hmm. Dont know enough about this.

 

[zakee00]

wow thats a cool idea, diffenately makes for a less power hungy and COOLER chip, i would think?

 

[Looney]

Originally posted by: logic1485
Full story here
I mean, come on, they've been talking about nano-computing since i was born (1985) (I don't know, but maybe I'm exageratting), but seriously, is this worth taking a note of. And while I have mentioned nano-computing, when can we expect that?

Um nanotechnology is used everywhere.

 

[bobsmith1492]

There really is no such thing as 'nanotechnology'. It's just a cool word used to get yourself a grant or make it look like you're really researching cool stuff. Anything really small can be called 'nanotechnology'.

 

[imhotepmp]

Originally posted by: futuristicmonkey

Originally posted by: PowerEngineer
My understanding is that Intel has been putting a lot of work into this. As the article says, they want to replace the electric wire connections to their CPU's with optical connections. 5-10 years?

This'll be like DDR2 then, eh?

I mean, DDR2 is fast, compared to DDR1, but it has so-so latencies, which are made up for by said speed.

So, while it may take a little longer to process the photons, the propagation speed (been waiting a while to use that ) of light (even thru anything) is A LOT faster than an electric signal, correct?

Or is this little 2:00 AM thought completely bonkers?

If by alot you mean ORDERS OF MAGNITUDE faster, then yes.

 

[f95toli]

It is NOT order of magntitudes faster. Electromagnetic waves travel through copper at about 80% of the speed of light. There are other advantages of using light to transmit signals (heat, multiplexing, distance etc) but it is not simply a matter of speed.

 

[matrix2]

I read about this for sometime now and i think it's a really good idea. Silicon as we know it will can only go so far. Look at intel prescott, you can fry an egg with that thing. using light is better because light can travel faster, look at fiber optics.

 

[unipidity]

But.. fiber optics is all about power loss (and the lack thereof), not transmission time.