DisplayPort and CMOS Photonics: Luxtera leads the way

Haven't read through it all yet (cool article though), but shouldn't the "Megabits/second" be near the vertical axis (first graph, second graph is fine)? I was confused for a few moments because normally you have the x-axis title correspond to what's changing on the x-scale...
 
Haven't read through it all yet (cool article though), but shouldn't the "Megabits/second" be near the vertical axis (first graph, second graph is fine)? I was confused for a few moments because normally you have the x-axis title correspond to what's changing on the x-scale...
It's either a bug in the graph generator or simply that I am bad at graphs :D

Thanks for the report, we're looking into it.
 
:) It's a really minor thing, but here's what I meant (in case my post didn't make sense):

graphwl9.jpg



Anyhoo, great article. I'm all for moving away from copper. :)

Finally, since the interconnect uses fiber optic cable compared to copper wire, the hybrid cables will be one-tenth the size and weight of their copper counterparts.

Would they then be similar in size/weight as those home theater optical audio cables? I'd personally feel safer with a cable that's resistant to bending, but I guess that's all in implementation later down the road.

Interchip communication sounds really exciting. :cool:

And now the long wait....
 
Ethernet

Does anyone else think that we should just wait for 100 Gb Ethernet over optical fiber, and then just use that as the next display standard? 100 Gb Ethernet would have plenty of bandwidth, plus it would have great economies of scale, and be much more flexible than any existing display standard. You could easily define an interface to control monitors over the Ethernet connection; you could even have a web interface to your monitors much like routers have. You could have one port on your video card, then hook up 10 monitors through a hub. Making a display wall would be simple. You could have three computers at your desk and swap monitors between them using software like <a href="http://synergy2.sourceforge.net/">Synergy</a> without needing a separate KVM. How awesome would that be?
 
Interesting.

I hadn't really heard of CMOS photonics until Arun mentioned it a few days ago.

Arun said:
[...]
And what about the Dalian fab, which will be coming online in 2010 on 90nm, no less than 3 process nodes behind everything else by that point.

My theory there is that Intel plans to dedicate that fab to southbridges but, more importantly, to CMOS photonics. And yes, I just implied that I believe China will soon be the world's largest producer (albeit via Intel) of CMOS photonics in the world. If that is indeed the case, I'm not exactly surprised that Intel got such a generous deal with the government, and I wonder what will be the reaction of the press once they realize that, if it is indeed correct.
[...]
 
Hey Spy Hunter, sorry for the slow approval in the moderation queue, I guess nobody realzied this post was still waiting! :)
Does anyone else think that we should just wait for 100 Gb Ethernet over optical fiber, and then just use that as the next display standard?
100 Gb would certainly give a truckload of flexibility, but I am sketical it could be made economical enough. Remember that the CMOS photonics would either be a separate chip (lower integration...) or the GPU and the device's transceiver would need to be manufactured in SOI. So if SOI becomes mainstream for other reasons in the next 5 years, then it's a possibility. Otherwise, it's not IMO. And even then, I presume the die size will be far from negligible for a 100Gbps CMOS Photonicsdesign...

Furthermore, Ethernet is two-way. I'm not completely sure it makes sense to have a two-way connection to the screen, although I guess that'd allow you to connect it to the same router/hub as your internet connection's. Whether that's actually desirable, I don't know though, hehe.
I hadn't really heard of CMOS photonics until Arun mentioned it a few days ago.
Yeah, the thought that Intel Dalian might actually be reserved for that purpose came to mind while I was contributing to this article. Southbridges are also a possibility, but this makes even more sense in my mind. It would definitely be very interesting if Intel bought Soitec or a similar company.
 
Does anyone else think that we should just wait for 100 Gb Ethernet over optical fiber, and then just use that as the next display standard? 100 Gb Ethernet would have plenty of bandwidth, plus it would have great economies of scale, and be much more flexible than any existing display standard. You could easily define an interface to control monitors over the Ethernet connection; you could even have a web interface to your monitors much like routers have. You could have one port on your video card, then hook up 10 monitors through a hub. Making a display wall would be simple. You could have three computers at your desk and swap monitors between them using software like <a href="http://synergy2.sourceforge.net/">Synergy</a> without needing a separate KVM. How awesome would that be?

A veritable geek fantasy. Excuse me while I go drool over the possibility of 10 paralleled boxes in the basement run to my 52" television.

Though, I must say, for the *really* geeky, even if you do this by mounting a very lightweight 22" LCD on a lapdesk type surface, add a keyboard and a 14' DVI cable...it just begins to seem all too possible to recline while connected to a very high powered box :D

Back on track...fascinating article. I haven't read all of it (I didn't look at the front page, so this is the first I noticed of it), but will do so. Excellent look at exciting tech.

There *has* to be a better way to connect equipment in the future that allows more flexibility.

EDIT: Great article. If there isn't already a massive demand for these solutions....about 10 minutes after the concept sinks in, you can envision about 100 different ways you would *love* to use it.
 
Last edited by a moderator:
Back
Top