By: Megol (golem960.delete@this.gmail.com), August 13, 2014 9:25 am
Room: Moderated Discussions
Megol (golem960.delete@this.gmail.com) on August 12, 2014 5:48 am wrote:
> anon (anon.delete@this.anon.com) on August 11, 2014 8:46 pm wrote:
> > Oh wow you solved everything. Lucky x86 has zero other decoding difficulties except for variable lengths.
> >
>
> Yes as far as I know. But I guess you are willing to give us all other problems?
>
> BTW: Things are only a problem if they hinder scaling when running actual code. If not I'm still correct.
>
As you haven't answered this yet (most likely because listing all x86 scalability problems take time) I'll again ask another question that you didn't touch earlier:
If the decode problems are specific to x86 why are most studies in scalability limits done on RISC architectures? And why do they show that going wider than two wide gives rapidly dimishing returns?
> anon (anon.delete@this.anon.com) on August 11, 2014 8:46 pm wrote:
> > Oh wow you solved everything. Lucky x86 has zero other decoding difficulties except for variable lengths.
> >
>
> Yes as far as I know. But I guess you are willing to give us all other problems?
>
> BTW: Things are only a problem if they hinder scaling when running actual code. If not I'm still correct.
>
As you haven't answered this yet (most likely because listing all x86 scalability problems take time) I'll again ask another question that you didn't touch earlier:
If the decode problems are specific to x86 why are most studies in scalability limits done on RISC architectures? And why do they show that going wider than two wide gives rapidly dimishing returns?
