Shift from 90 to 80 nanometre is no risky move

WE SPOKE with some smart chaps who make the chips we are making the fuss about. It is always best to hear stuff from the horse's mouth and that is what we achieved this time. The chap said that Nvidia's shrinkage from 90 to 80 nanometre is not that risky move at all. We originally wrote about it here

The chips chaps call the transition from 90 to 80 nanometre an optical shrink and it does not cause the big risk especially for complex digital chips such as GPUs.

When you go to 80 nanometre you have to design your logic and even the floor plan at 90 nanometre and all you have to do is the optical shrink in the fab. At the litho stage of chip development you just produce a smaller geometry and that's it. If you want to go from 90 to 65 nanometre you have to remake the chip virtually from scratch and to make all transistors, gates and logic 65 nanometre compliant.

The optical shrink from 90 to 80 nanometre is cheaper and faster than going to a next the processor node, such as 65 nanometre. You can keep your 90 nanometre tools and still make the 80 nanometre chips and you don't have to worry about all the pain and suffering of the transition to 65 nanometre.

So the transition from 90 to 80 nanometre for Nvidia's G80 would not be that dramatic. Yet Nvidia obviously didn't want to take any chances and it wants to be out on the market ASAP, probably in late Q3 - by the end of September, or early days of October. µ