View Full Version : Sun Researching Super Fast Laser Processors
Soulburner
03-24-2008, 04:31 PM
Sun wins $44 million contract from U.S. Defense agency to develop processors based on lasers which it claims will be 1000X faster than today's chips.
http://www.tomshardware.com/2008/03/24/sun_laser_chips/
While shrinking transistor sizes to ever decreasing nanometer lengths is a popular topic in the industry today, several companies are working towards removing physical copper connects altogether, replacing them with super high speed laser connections.
cajer
03-24-2008, 04:34 PM
Hopefully these can get to us, the consumer, faster unlike all of these other new technologies that we have been hearing about for years.
NeedMoMegaHurtZ
03-24-2008, 04:42 PM
"Fricken laser beams!" :p:
hecktic
03-24-2008, 04:59 PM
dam... lucky them
Spawne32
03-24-2008, 05:01 PM
we all knew this was coming, we are pretty much maxing out our limitations with electric(DC) based computers
Zytek_Fan
03-24-2008, 05:04 PM
They've been working on this for quite a while.
I've got an issue of Computer Power User from May 2006 that features this technology.
Omastar
03-24-2008, 05:55 PM
SOI technology is reaching the end of its useful life, isn't it?
fart_plume
03-24-2008, 06:06 PM
They've been talking about optical processors for 20 years or better. I'll believe it when I see it.
Bobsama
03-25-2008, 07:03 AM
I'll believe it when I can buy it.
iddqd
03-25-2008, 07:27 AM
I honestly don't see the allure of going from photons to electrons.... and then back to electrons.
Well, in any case, the oh-so-low speed of electrons in a semiconductor is by no means the limiting factory in transistors' performance. We can build transistors that switch 100, 1000, even a million times faster than the stuff we currently use. Nobody came up with a commercially viable way to switch the manufacturing process all at once, though.
quantum computers are supposed to be way faster than even these ;)
n-sanity
03-25-2008, 08:31 AM
"..with frickin' lasers attached to their head"
Lets hope they don't come up with gold penis shaped ram modules next...
For those that aren't clued in, austin powers is google-friendly.
I honestly don't see the allure of going from photons to electrons.... and then back to electrons.
Well, in any case, the oh-so-low speed of electrons in a semiconductor is by no means the limiting factory in transistors' performance. We can build transistors that switch 100, 1000, even a million times faster than the stuff we currently use. Nobody came up with a commercially viable way to switch the manufacturing process all at once, though.
Not with silicone we couldn't.
NH|Delph1
03-25-2008, 08:43 AM
So optical chip-to-chip I/O interconnects equals processors nowadays?
//Andreas
Nedjo
03-25-2008, 12:11 PM
what a waste of US tax payers money! putting laser diodes (or any other diodes) inside of the chip is bringing more problems then it solves existing. BTW existing problem isn't the speed of interconnections, but reaching the size of the Si atom!
Hornet331
03-25-2008, 12:26 PM
its funny how some companies need tax money while others fund the same thing by themself -> Silicon Photonics (http://techresearch.intel.com/articles/Tera-Scale/1419.htm)or Hybrid Silicon Laser (http://techresearch.intel.com/articles/Tera-Scale/1448.htm)
Origin_Unknown
03-25-2008, 12:30 PM
"Fricken laser beams!" :p:
"..with frickin' lasers attached to their head"
Lets hope they don't come up with gold penis shaped ram modules next...
For those that aren't clued in, austin powers is google-friendly.
Not with silicone we couldn't.
i cant resist :D
http://reuben.typepad.com/photos/photohosting/sharkswithlasers_001.jpg
NH|Delph1
03-25-2008, 01:55 PM
what a waste of US tax payers money! putting laser diodes (or any other diodes) inside of the chip is bringing more problems then it solves existing. BTW existing problem isn't the speed of interconnections, but reaching the size of the Si atom!
The original article is misleading. Sun is being paid to make optical interconnects, I.e. between chips, not inside of chips.
And you're partially wrong saying that the speed of the interconnect is the problem, it is a problem when you take it into context for what the research grant is really for.
There are many bottlenecks with the modern way of connecting processors and supercomputers, this is just one of them.
//Andreas