Linus Torvalds (torvalds.delete@this.linux-foundation.org) on May 19, 2013 11:24 am wrote:
> Ricardo B (ricardo.b.delete@this.xxxxxx.xx) on May 19, 2013 7:20 am wrote:
> >
> > It mostly depends on where in the single threaded power/performance curve the design is aiming for.
>
> No, I think it's bullshit today, and it was bullshit five years ago.
>
> The reason Atom started out in-order wasn't because that was a particularly energy-efficient
> design, it was because it was a quick-and-dirty "let's get something small running, and two
> guys in this small team worked on the P54 and can whip up a prototype in three days".
>

Sounds very LRB. Not at all like a Silverthorne.

> Then the "prototype in three days" ended up being six months to actually get something
> that works, but by golly, it actually did, so when they sold this concept of Atom
> to management, you had a quick hack that you needed to make excuses for.
>
> And the excuse was "OoO is too power-hungry".
>
> And people lapped it up like it was Moses bringing down the tablets from the mountain.
>
> The real disgrace is how it then took Intel five years to decide to do it
> right. Even though I guarantee that anybody with an IQ over room temperature
> knew that the original Atom sucked the moment they got to play with it.
>
> Linus