starsmine
Im just frustrated with how unclear they have been with Bristol ridge, March slides show it being something other then zeden, May slides show it being Zen.
If the leaked claims are accurate, Bristol Ridge is just an attempt to both unify codenames and allow a version of Carrizo on desktop before Zen shows up. So of course, AMD will want to get proper return on investment and won't push the Zen APU for another year after that. Considering previous debuts, the October date for the Zen 8-core makes sense, and the APU in spring of 2017 also makes sense. But if you're interested in a laptop, I suppose you've weighed out battery versus performance as Intel will push Cannonlake out by then, over Intel always overcharges for their IGP. By the time the next architecture at 10nm is pushed out, Zen+ should show, by then you might graduate.
AlexisRO
single threaded performance at stock greater than my Sandy at 4.5G (at least 15% improvement).
I'm sure that isn't going to happen, what you're describing would be faster at stock than even Skylake; it isn't realistic and you can't hold that against AMD since I doubt even Intel can satisfy that.
Going by AMD's own optimistic slides, Zen would match the IPC of Sandy Bridge (Piledriver is 80% behind Haswell but Steamroller adds 20%, Excavator adds 5% and Zen add 40%).
Except we don't know what frequency they will set for stock speeds, I doubt the Zen quad w/HT will debut at 3.4GHz to really match 2600K. Therefore, let's say it debuted at 4GHz, that would match a stock 3770K or 3820K. It is going to meet up with an Ivy Bridge class processor in CPU-bound apps.
If you do any GPU-bound, I don't think there is any point in upgrading your CPU.
AlexisRO
Multi threaded performance, considering the 8c 16t, as close to a 5960X or above.
I'd imagine an 8-core Ivy Bridge-EP equivalency, like Xeon E5-2667 v2.
NOTE: Just because the probability of AMD not beating Intel is very high doesn't mean Intel won't do something about it. They have enjoyed a luxury of being absolutely dominant for almost a decade, they will attempt to maintain it buy furthering the gap, thus will push either Skylake-E early (by cancelling Broadwell-E and EOL'ing X99) or push even higher frequencies for the Skylake Refresh known as "Kaby Lake" until the next architecture after Cannonlake in 2018 that will push the IPC higher on the order of Core 2. We just might see a 6-core dual-channel option in LGA1151 platform, but we're talking about a post 2017 reaction to Zen and Zen+.
But if Zen flops and doesn't even beat Excavator, then Intel will give us quads in mainstream and 5-10% gains per generation forever... something fanatics have settled on and accepted.