My guess would be that they are completely focused on Zen and having it be usable through the entire product stack. Introducing Jaguar to OEMs at 14nm might mess with the messaging they are trying to get across. It may be good short term, but perhaps they don't view it as a viable long term play with the potential to disrupt their long term planning.
Just a guess. And like I said, until someone officially comments on it or someone does a teardown of the XBO-S, we won't know if it's still 28 nm or 14/16 nm. If it's 28 nm, what'd they do so that the smaller size and more congested interior of the XBO-S didn't interfere with the low acoustics of the XBO? Again, things we won't know until we can see a tear down.
I'm pretty interested in seeing what they did with the overall interior design of the XBO-S.
Regards,
SB
May be AMD is holding on its announcement, in my view they need those cores mighty Intel decide to stick to its Atom CPU walking away from some previous claim on the matter.
I would still find weird for them to have the product ready before Zen (the OneS launch in august, right,), the same applies to the GPU to a lesser extend. My guts bend toward 28nm so my reasoning, even though I am surprised that MSFT went through the effort of simply updating Durango but lithography having following the path it followed. We are going to know soon enough but it is interesting to see what a durango 2.0 could look like based on late AMD latest hardware.
based on some comments it seems that the chips has some headroom (on the gpu side made available to HDR games) and its position as a potent media hub is also a hint a proper power management (finally). that is something that AMD newer IPS can deliver even at 28nm.
The same comment about the headroom (on the GPU side) wrt to resolution scaling have me to think of AMD late color compression feature, it could allow to make more of the 32MB, bandwidth is not the constrain here but gain in storage are possibly welcome.
Those comment about the smoother gameplay have also me thinking of slight turbo mechanism on the CPU side.
So it could be possible that Durango 2.0 is durango built upon Puma+ and the GCN version used in Tonga and Fiji. Proper power management along with architectural improvements could explain a lower power consumption as well as the smoother performances and the "headroom" available to some titles.
Among other things up upscaling quality could be a bit better, encoding (so both recording and streaming) should be more power efficient (may be better quality).
Msft did not make a good at selling this product, it looks like a nice step up, good price, but they let geeks with lots of questions.