[extropy-chat] Moore != AI (was Re: Taiwan)

[Eugen Leitl]

On Thu, Mar 31, 2005 at 06:00:45AM -0800, Mike Lorrey wrote:

> I was under the impression we were going to shift to photonic circuits
> within 3-5 years. That alone is a huge reduction in size and power
> requirements. What gives?

Photonics is one of the candidates to solve the signalling problems (data
rate, power, die area), but mostly off-die. There are alternative solutions
(e.g. wafer stacking, recently proposed by Intel) proposed, which solve the 
memory bandwidth problems by providing millions of lanes between CPU cores
and memory blocks. 

Photonics for switching is only interesting for purely photonic networks
(which avoid costly and slow photonics-electronics-photonics path). Such
switches are relatively big, and are power hogs. Very little integration
density there.

Photonic modulators and waveguides connecting subunits on-die are probably
more than a decade off. You're probably accurate with 3-5 years estimate 
for first commercial silicon including fully integrated on-die photonics.

Nobody knows what the first successful product for this will be, though.

-- 
Eugen* Leitl <a href="http://leitl.org">leitl</a>
______________________________________________________________
ICBM: 48.07078, 11.61144            http://www.leitl.org
8B29F6BE: 099D 78BA 2FD3 B014 B08A  7779 75B0 2443 8B29 F6BE
http://moleculardevices.org         http://nanomachines.net
-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: application/pgp-signature
Size: 198 bytes
Desc: not available
URL: <http://lists.extropy.org/pipermail/extropy-chat/attachments/20050331/1dd13aa7/attachment.bin>