No, there is not such possibility. I really think he believes in that nonsense.
Sunnyvale 2010-2012:
- Ambitious derivative of Bulldozer...
- ...later to root cause competitive performance gaps. (Pretty sure this direct quote means to d
Boston Design Center (2004 - 2014):
- K9 Construction core(Bulldozer v0.3)- 90nm
- K10 Construction core(Bulldozer v0.7)- 65nm
- Bulldozer- 45nm
- Bulldozer/Piledriver- 32nm
- Steamroller A/Steamroller B/Excavator- 28nm
May 2013:
FP256 is added to AMD64 guides.
Sunnyvale 2013-2015:
- Aspen/K12 starts.
- Bulldozer returned to Boston.
- Aspen/K12 ends.
Boston Design Center (2015/2016-Present):
- NG Bulldozer starts @ Boston.
- (Late 2014) // 2004 Bulldozer Team Lead returns to AMD.
- (2016-Present) // Pre-silicon validation of Bulldozer family cores. ((seronx reacts: OH NO, how did I miss this?!))
Sunnyvale (2016-Present):
- NG Bulldozer starts @ Sunnyvale. // alludes to a new new generation core after Aspen/K12.
November 2013 v0.9 - 2016 Roadmap:
Merlin Falcon (28nm-15h) -> ---- (20nm-15h)
Peregrine Falcon (28nm-16h) -> ---- (20nm-15h)
Those two 20nm skus will later be stop-gapped by Bristol Ridge/Stoney Ridge Falconized skus. ((22FDX is placed by 20nm LPM as Advanced FDSOI:
http://i.imgur.com/yrzjA0v.jpg)bb=0v)
Not once has Bulldozer ever been completely cancelled by AMD.
http://semiaccurate.com/2012/11/19/amd-kills-off-big-cores-kaveri-steamroller-and-excavator/
It never dies, it will always haunt you, it hunts you, it will find you, and you will be ah darn AMD is doing 15h again.
Given AMD spokesperson and various other references...
- 15h is well and alive.
- It will target cheap SKUs. (Need to get a return on all that wasted effort on Aspen/K12)