Inside Global Foundries...

I always though it was going to go:

32nm -> 22nm -> 16nm -> 11nm -> magic pixies

Where did 28nm come from? Who decides these lithography dimensions, some consortium of companies or some law of optics/physics? (just curious)
 
Every node, you divide the gate length by sqrt(2) to reduce the area per tranny in half. The long running sequence is indeed what you say. Recently half nodes have become popular as it is becoming harder and harder to shrink further. 16nm and beyond, conventional mosfets may have to go.

28 nm came in because let's face it, 22nm is close to the last stop (if not the last stop).
 
Back
Top