[strike]By the trend of what GPU architecture AMD uses for it's APU (last year's architecture), it seems that they might use GCN with this one (I hope they do). Here's to hoping as well that they're able to integrate some of the improvements (whatever they are) in the upcoming desktop HD 8000 series, which is assumed to be still GCN I think.
What confuses me is, what is "Richland" exactly? By the way it looks, it seems like it's an APU itself (like Trinity), but for some reason it says on Wikipedia
that it's supposed to be the low-end Piledriver-based APU. This is probably outdated/erroneous info which Wikipedia is highly susceptible of having.
What still confuses me is, what happened to Kaveri? Is Richland just a 28nm transistor process shrink of Trinity i.e. still using Piledriver (plus other improvements like the GPU)? If that's so, would it come out in this same year as Kaveri? This entry in AMD's blog
seems to confirm at least that part about having the same power consumption that Wolfgang (or his source) speculated.
Wisecracker's info seems to answer these questions, but I'm thinking a lot of this is still uncertain.[/strike]
EDIT: Oops. Just got to see your addition of that slide/pic now Wisecracker. Looks pretty legit. Just wondering, where'd you get it? Straight from AMD's website or another site showing it from CES or something?