Originally Posted by PiOfPie
Plus Intel has Alpha's IP, so they'll have a reasonably good idea of what parts of the 21264 he views as a success and will be able to counter accordingly. Bulldozer heavily curbed from the 21264, so the the parts of Zen adopted from the construction cores will be pretty easy for Intel to infer, as well.
The only "blueprints" that Intel don't have are: a) the Cat Cores and b) any "secret weapons" that Keller's thought of while working on the Ax chips and since then.
Your statement assumes he used anything of AMD's current hardware, or philosophy.
We now have a quote (the most recent) that he was given freedom to build from scratch, starting over entirely. If that is a case no one outside of the team developing it have any idea what it will be. At least not beyond known limitations in engineering. He isn't going to reinvent the wheel, but that doesn't mean Intel or anyone can have an idea worth voicing.
The closest guess anyone, right now, could make is that it will fall between current performance from AMD and faster than what Intel has. Why? If it was truly a ground-up, truly new-new architecture, the sky is the limit. You are talking about the possible scenario where a very desperate company said "Here is a mountain of cash, save us."
. That would have created a situation of one of the best Architects, with a massive budget, and all his years of knowledge to current, building a brand new architecture.
Why is it hard to believe that one of the best, with access to the best technologies, and (hopefully) the budget to back it, with freedom, couldn't create the best processor? If Zen truly is a ground up build, that is a departure from what AMD has been doing.