All this nomenclature aside, the big question is, will 14nm be a big improvement? There are two very real possibilities: 1. It's inherently harder to shrink the transistor and get performance nowadays so it will be yet another "~5%" (10 max). Or.. 2) Haswell has been 'trying things' so to speak when it didn't perform spectacularly (since they had the chance to slack since AMD/TSMC is so far behind) hence to really unleash a 20% boost at least. I do suspect the latter may happen. It has been too boring per release for so long. It might have been at least partly artificial.
post #31 of 53
8/26/13 at 10:52am