The future of GFX is looking rather small

Some info from a familiar name http://theovalich.wordpress.com/2008/11/18/tsmc-introduces-40nm-volume-production-advances-in-front-of-intel/

Looking forwards, by the time say Larrabee arrives , late 2009 or early 2010, a G200 would be 1/3 the saize today, or more precisely, well have cards out thatre better than 3 G280s in tri sli. You just gotta love the future, even as it shrinks, it seems less IS more
 

homerdog

Distinguished
Apr 16, 2007
1,700
0
19,780
Yes, this is how it goes. Silicon will reach its limits though; we will have to switch to something else (graphene?) if we are to keep Moore satisfied :p
 
G

Guest

Guest
Confidence issues? hehe... At a certain point things can only get so small... otherwise its just an ATOM... and you can't really tell an atom what to do... like you can't have a computer made out of a .4 nm transistor... it just doesn't work... you will always have to have some space... and then at that point all processors will do is instead of shrinking... they'll get larger... they'll build up and out
 

antiacid

Distinguished
Oct 21, 2008
456
0
18,790
Confidence issues? hehe... At a certain point things can only get so small... otherwise its just an ATOM... and you can't really tell an atom what to do... like you can't have a computer made out of a .4 nm transistor... it just doesn't work... you will always have to have some space... and then at that point all processors will do is instead of shrinking... they'll get larger... they'll build up and out


seeing as you don't really have a firm grasp on dimensions and sizes, here's a quick link for you:

http://www.webelements.com/copper/atom_sizes.html

Your average copper ion is in the double digits picometer range. That's 0.1 Ångstrom or 0.05 nm.

Your average copper atom is in the triple digit picometer range. That's 1.3 Ångstrom or 0.13 nm.

As of right now, we're looking at 45nm-32nm stuff. The intel's roadmap is planned until 10nm before 2019.

As for going smaller than that, there's big chances that a new "paradigm shift" will happen, ie DNA-based computing or something of the sort. Those ideas are tested in labs across the world as we speak. I wouldn't be worried that scientists and engineers will come up with new and innovative solution in the next decade :p