Intels 3rd gen chips, IVY BRIDGE use the 22nm process, ok good for them.
Rumored next gen GPU chips are rumored NOT to shrink from 28s.
Why isn't amd and nvidia shrinking their die size? they don't run on a tick-tock method, they could get alot more performance and efficiency by switching to 22s. What's holding them back? is it because in the world of economics, 22s arn't that economical to make or what?
companys should stop puting budget and dissapointment 1st, and switch them around.
I think that GPUs should lead in the race to single didget manufacturing processes because in most systems GPUs take up the most power.
smaller nm requires all new factories, with all new processes, and all new chip design. the R&D is expensive, and so is all the new equitment.
also, cpu's have much higher ghz, so there is more heat being built up, so the smaller nm matters quite a lot. since gpu's have a very simple core archetecture compared to cpu's theres not as much heat generated. (however there is a lot more of them).
the companies can sort of get by with a larger nm process.
they will need a lot of new specilized equipment so they need to make a new factory, though I really wish Nvidia would make a smaller nm, they are already so power effecient smaller nm would mean they would use even less
Probably causes more problems for said companys, they can still get more effeciancy out of 28nm, and 22nm can end up Generating more heat and a 2 slot cooler can only do so much, so id guess that theyre trying to figure out how to deal with heat/get more efeciencey out of that nm process they already have in other ways before they mot to a low nm as it could cause more problems than it solves.