[rescue] SGI, MIPS, and IA-64

[Joshua D Boyd]

On Fri, Feb 15, 2002 at 05:28:29PM -0500, Dave McGuire wrote:
>>I think the bundles are cool (a bundle has 3 41bit instructions in them, which
>>are guaranteed to execute simultaneously, to my understanding), plus for the
>>most part it looks like IA-64 can execute 2 bundles at once (meaning 6 parallel
>>instructions).
> 
>   Wow, six whole instructions...one more than the (several year old)
> MIPS R10K.  Isn't Intel wonderful?

Err, the R10k only has 5 execution units.  I'm assuming that they operate
in parallel, but I can't seem to find any sources that say this explicitly.

I do see out of order execution on the R10k (I think that IA-64 chucked this,
but I'm being to lazy to look it up).  I see speculative branching.  IA-64
does that.  At one time they were going to add hints to it so that you can
hint at which branch is more likely to make it a bit less speculative, but
again I'm being too lazy to look it up.   They both are certainly complicated
chips though, one of the reasons why I still consider the 68030, 68040, r3k and
r4k to be my favorite chips, even if they aren't as fast.


-- 
Joshua D. Boyd