Researchers create laser light interconnects on silicon

  • Thread starter Thread starter Jan Panteltje
  • Start date Start date
John said:
EE Times has a page that announces a similar breakthrough every week,
but I haven't heard of any of these breakthroughs going commercial.

The laser makes heat, its driver makes a lot more, and the detector
speed will be dominated by capacitance. True, the middle part, the
optical waveguides, will be fast, but that's always been the easy
part.

The NYT article is full of obvious errors.

And didn't Intel just pull the plug on this technology? Yeah, here it
is:

http://biz.yahoo.com/bw/060911/20060911005589.html?.v=1

http://www.informationweek.com/news/showArticle.jhtml?articleID=192701025

I think they paid a few per cent of what Intel had invested; Intel
just dumped it to get it off the balance sheet.

John

Intel's making a huge mistake. With suitable technobabble,
they could've marketed it to the audio crowd. The term
"intrinsically laser purified audio" comes to mind ...

Ed
 
George said:
Oh, it's been longer than that - holographic was all set to replace core
memory back in uhh, ~1969.

Only a few years ago there was a holographic storage device that, had it
been built, would have stored a massive 100GB. Overtaken by history. Again.
 
Intel's making a huge mistake. With suitable technobabble,
they could've marketed it to the audio crowd. The term
"intrinsically laser purified audio" comes to mind ...

Ed

Intel is like Microsoft. They stumbled onto a big thing by accident,
and interpreted the event as skill. And everything else they try
doesn't work.

John
 
John said:
EE Times has a page that announces a similar breakthrough every week,
but I haven't heard of any of these breakthroughs going commercial.

The laser makes heat, its driver makes a lot more, and the detector
speed will be dominated by capacitance. True, the middle part, the
optical waveguides, will be fast, but that's always been the easy
part.

The NYT article is full of obvious errors.

And didn't Intel just pull the plug on this technology? Yeah, here it
is:

http://biz.yahoo.com/bw/060911/20060911005589.html?.v=1

http://www.informationweek.com/news/showArticle.jhtml?articleID=192701025

I think they paid a few per cent of what Intel had invested; Intel
just dumped it to get it off the balance sheet.

John

John Bowers of UCSB is a collaborator of mine...we're on the Intra-Chip
Optical Networks (ICON) team in the DARPA UNIQ program. He's a very
smart guy who knows his stuff, and this is a pretty nice piece of work.
However, he's not the one doing all the heavy breathing--it looks like
the Intel hype machine is what's behind that. (Those Intel hype guys
are _good_--better than the technology people, that's for sure.)

The gee-whiz terabit numbers are assuming things like 40 Gb/s per
wavelength, with 50 wavelengths per line, which isn't impossible at all.
The idea is that one 'wire' can pass many full-speed logic signals
simultaneously, which is a big win. You don't need terahertz logic
speeds to do this, which is a good thing since we're not going to have
terahertz logic speeds, ever.

You need a transistor f_max at least 10 times the logic speed, and you
can't make transistors with f_max of 20 THz. This is because the
maximum frequency that signals can propagate in a semiconductor is
proportional to the square root of the carrier density. The same is
true in plasmas, which is why this limit is called the 'plasma
frequency.' You can't dope semiconductors enough to get the plasma
frequency higher than 10-30 THz, so you can't make 20 THz transistors,
so you can't make 2 THz logic out of transistors. Not to mention that
at that speed, the region of the chip you can keep synchronous is about
30 microns square, on optimistic assumptions.

It's very helpful to have optical gain in all-photonic ICs, and these
InP-based devices may turn out to be an important part of the tool kit.
Wire is so much easier to make than these optical gizmos that we have
to have a really amazing advantage in real computer performance before
the chip guys will even talk to us. (I'd feel the same in their shoes.)
Fortunately it looks like we can do that.

Cheers,

Phil Hobbs
 
Phil said:
John Bowers of UCSB is a collaborator of mine...we're on the Intra-Chip
Optical Networks (ICON) team in the DARPA UNIQ program. He's a very
smart guy who knows his stuff, and this is a pretty nice piece of work.
However, he's not the one doing all the heavy breathing--it looks like
the Intel hype machine is what's behind that. (Those Intel hype guys
are _good_--better than the technology people, that's for sure.)

The gee-whiz terabit numbers are assuming things like 40 Gb/s per
wavelength, with 50 wavelengths per line, which isn't impossible at all.
The idea is that one 'wire' can pass many full-speed logic signals
simultaneously, which is a big win. You don't need terahertz logic
speeds to do this, which is a good thing since we're not going to have
terahertz logic speeds, ever.

You need a transistor f_max at least 10 times the logic speed, and you
can't make transistors with f_max of 20 THz. This is because the
maximum frequency that signals can propagate in a semiconductor is
proportional to the square root of the carrier density. The same is
true in plasmas, which is why this limit is called the 'plasma
frequency.' You can't dope semiconductors enough to get the plasma
frequency higher than 10-30 THz, so you can't make 20 THz transistors,
so you can't make 2 THz logic out of transistors. Not to mention that
at that speed, the region of the chip you can keep synchronous is about
30 microns square, on optimistic assumptions.

It's very helpful to have optical gain in all-photonic ICs, and these
InP-based devices may turn out to be an important part of the tool kit.
Wire is so much easier to make than these optical gizmos that we have
to have a really amazing advantage in real computer performance before
the chip guys will even talk to us. (I'd feel the same in their shoes.)
Fortunately it looks like we can do that.

Lots of 40GB/s channels down one optical waveguide does sound
interesting - serial busses are cute, and a multiplexed otpical serial
bus would be just that much cuter.
 
fammacd=! said:
Oh, it's been longer than that - holographic was all set to replace core
memory back in uhh, ~1969.

I heard about it for mass storage for the Illiac-IV, about that
time. The real question is "why bother"? S/N of the media is the
limiting factor.
 
John said:
Intel is like Microsoft. They stumbled onto a big thing by accident,
and interpreted the event as skill. And everything else they try
doesn't work.


That's what they get for trying to maintain backward compatibility to
the 4004 processor.


--
Service to my country? Been there, Done that, and I've got my DD214 to
prove it.
Member of DAV #85.

Michael A. Terrell
Central Florida
 
Intel's making a huge mistake. With suitable technobabble,
they could've marketed it to the audio crowd. The term
"intrinsically laser purified audio" comes to mind ...
Now, that's good thinking:-)

Mati Meron | "When you argue with a fool,
(e-mail address removed) | chances are he is doing just the same"
 
"Data is recorded at 1.4 million bits per second, using a blue laser from
405 to 407 nm in wavelength."

That's truly impressive, man! Almost as fast as 1x CD recorder! Only 19 days
to write a 300 GB disk.
 
John Bowers of UCSB is a collaborator of mine...we're on the Intra-Chip
Optical Networks (ICON) team in the DARPA UNIQ program. He's a very
smart guy who knows his stuff, and this is a pretty nice piece of work.
However, he's not the one doing all the heavy breathing--it looks like
the Intel hype machine is what's behind that. (Those Intel hype guys
are _good_--better than the technology people, that's for sure.)

It's everywhere: newspapers, NPR radio, all the electronics web sites.
And they all, to various and usually severe degrees, get the
technology all tangled. Some of the bloopers I've heard include...

Calling it silicon lasers

Saying it will speed up chip-chip interconnects because light
propagates faster than electricity

Signals are fast on-chip but slow down between chips

Laptops get hot because of signals heating wires inside

This will be practically applied by the end of this decade.


"Lasers are already used to transmit high volumes of computer data
over longer distances — for example, between offices, cities and
across oceans — using fiber optic cables. But in computer chips, data
moves at great speed over the wires inside, then slows to a snail’s
pace when it is sent chip-to-chip inside a computer."

- the New York Times


John
 
I've been reading of these 'optical breakthroughs' for the past 25yrs.
Don't hold your breath.

There have been many: fiber optics, CDs and DVDs, CCDs and thermal
imagers, semiconductor lasers, laser printers, high-efficiency LEDs
from infrared to ultraviolet. But Intel hypes every obscure gadget as
the Next Big Thing, like their Raman "on-chip silicon laser" that
needs a table-sized external optical pump.

Pasting compound-semi lasers on top of silicon chips is interesting,
worthy of a journal article, but everybody is already declaring it to
be a miracle that will change the world. Unlikely.

John
 
"Data is recorded at 1.4 million bits per second, using a blue laser from
405 to 407 nm in wavelength."

That's truly impressive, man! Almost as fast as 1x CD recorder! Only 19 days
to write a 300 GB disk.


Yea, that is a good point.
But it took longer to make the movie.
Anyways from 1x CD-R to 32x CD-R took only a few years.... and 32x was the
limit because of centrifugal force, so as he had his 30 years, the next few
years should give exponential speed increase.
It is quite common for video render farms to run for days anyways.

But you could burn to blue-ray too.
However when you look at data carrier volume, this thing wins.
Say library of congress, how much space needed.
Storing in New York is more expensive then on the land ;-) maybe use these
things mainly in cities.
 
"Data is recorded at 1.4 million bits per second, using a blue laser from
405 to 407 nm in wavelength."

That's truly impressive, man! Almost as fast as 1x CD recorder! Only 19 days
to write a 300 GB disk.


Yea, that is a good point.
But it took longer to make the movie.
Anyways from 1x CD-R to 32x CD-R took only a few years.... and 32x was the
limit because of centrifugal force, so as he had his 30 years, the next few
years should give exponential speed increase.
It is quite common for video render farms to run for days anyways.

But you could burn to blue-ray too.
However when you look at data carrier volume, this thing wins.
Say library of congress, how much space needed.
Storing in New York is more expensive then on the land ;-) maybe use these
things mainly in cities.
 
(e-mail address removed) schreef:
Back to insulting people just for the sake of an insult? Still can't
get how to conduct a discussion in a civilized way? You really nead
to learn the basic manners before posting your childish insults.

Well, I was not insulting anyone.
You would not happen to be a Muslim fundamentalist now would you?
They get insulted by a self-inflicting reflective suicide-insult
mechanism.
like this:
http://groups.google.com/group/us.p...c2926c62b6f/98d4e474dd67e280#98d4e474dd67e280
'Moslems Threaten to Massacre Christians Over Claim of Islamic
Violence'

LOL
 
That's what they get for trying to maintain backward compatibility to
the 4004 processor.

No, they fail *miserably* every time they try yo break that
compatibility (e.g. iAPX-432, i960, Itanic). Compatability *is*
their one-trick pony.
 
martin griffith said:
What's Sealab 2021 like? Never heard of it

It's a spoof of the original, Sealab 2020. They take the original animation
and dub in new dialog for comic effect.

After they succeeded the guys doing it obtained high-end editing stations and
a budget such that, in more recent seasons, there's lots of entirely new
animation and much more "slicing and dicing" of scenes (part of the genius of
the initial seasons was that the animation ran pretty much the same as the
original, uncut) -- the fact that some of the scenes still came from Sealab
2020 no longer seems like a significant part of the episodes, which to me
makes them nowhere near as good as the first couple of seasons.

Amazon.Com would be happy to sell you the DVDs; it's on late at night on the
Cartoon Network in the US.
 
Back
Top