Article: AMD's Mobile Strategy
By: anon (firstname.lastname@example.org), December 20, 2011 7:35 pm
Room: Moderated Discussions
>How is LEA handled in most designs? Is it single uop?
In the original K8 it was multiple macro-ops, but they added an optimization in one of the steppings to rewrite it to one.
(that's a cool thing about the much maligned x86
frontends: they can do tricks like that relatively easily.
I also like the loop buffer the newer Intel cores have)
>Yeah, I think it will also help to see a new uarch from Intel. That may give a
>better idea of what's feasible, because I think Intel has avoided uarch changes to focus on system level issues.
You already have another low power intel uarch coming:
MIC. While the total power is huge there so many cores
that each has only a small power budget (150/50/fraction
the gigantic vector unit leaves over for integer) ?