I understand what you mean now, thing is that you wrote that voltage scaling was not as good in a separate paragraph, as such it could has been understood as being a statement that would hold for the whole TDP range.
Given the advertised TDP range of Carrizo on FP4 BGA (10-35W), it should be a good deal more efficient than Kaveri in that same TDP range, which will be great for notebook buyers as well as premium AiO buyers ($500-$550+ range). That is, assuming AMD can convince many OEMs to offer their chips in reasonably priced and equipped units.
AMD should seriously think about releasing either an AM1 version or an updated dual channel such plateform, given the MBs inherent lowish prices they could ask a nice premium for such an APU and still be very competitive for the APU + MB combo.
The main reason I agree with you here is that AMD already incurs the expense of maintaining SKUs for the AM1 line. They could EoL current products and bring in Carrizo/Carrizo-L. The absence of dual-channel RAM on AM1 is sort of a bummer, so I suspect Carrizo would be wasted on the platform.
EoLing the entire AM1 platform and replacing it with an update to accomodate dual-channel memory would be . . . more expensive, I think, and a risky maneuver considering AM1's apparent low volume.
Still, if AMD wants to fight against Intel's NUCs/UCFF units, then AM1 (or a hypothetical successor) is about the only platform they have right now equipped for that battle. Unless they start selling FP4 BGA units in that space, too . . .
CEO replaced, multiple SVP's heads rolled, Carrizo rollout scaled back to limited platforms.
The product isn't out but AMD has certainly recognized the problem(s) - Read, Byrne, LaForce, Naik...
Plenty of red flags indicating how well the product is expected to do by the very people entrusted with authority to make decisions for AMD's shareholders best interests (the very people who have the product in their hands now and are internally testing/validating the chip).
I'm not sure that Carrizo is to blame here. Everything is being scaled back except production of existing product lines. AMD appears to be taking a very conservative stance leading up to the Zen release. If anything, I would least suspect Carrizo of the Rory upheaval, since AMD is still (somewhat) willing to discuss its release in public. Mention of products like Nolan and Amur - particularly Amur - is scarce.
From the looks of the slides, AMD's decision to pull Carrizo from FM2+ seems more indicative of a genuine technical limitation of the design than anything else. Carrizo is being sold on efficiency which apparently doesn't translate to desktop-level TDPs.
As much as I like Kaveri, I suspect that Kaveri is more likely to blame for Rory & Friends getting canned than Carrizo. So much potential, so few sales.
I suspect it may be Seattle (and the whole ARM server business) which is throwing up red flags, and leading to the C-level cull. "Ambidextrous" was Rory's baby and Seattle was meant to be out 6 months ago, but it has sunk without a trace.
Seattle is another product that has received little attention as of late. Unlike Nolan or Amur, Seattle is supposedly already launched, or . . . something. So where is it? At this point I am genuinely surprised AMD is even bothering with ARM.