Mail Archives: pgcc/2001/02/20/14:28:21
On Tue, Feb 20, 2001 at 09:10:30PM +0000, Nick Kurshev <nickols_k AT mail DOT ru> wrote:
> You declare that manuals contain the lie.
> Well, I did my own investigation and results say that you are wrong, please see below:
Well, I have yet to see that, but you are fere to claim and tell whatever
you like :(
> 1. Official manual don't contain a lie (may be mistakes and errata only) and I still prone to trust them.
most of the optimization tips (like avoidong push and the like) are wrong,
you may claim whatever you like, but please go elsewhere -> I am not the
only one who fell into that trap, you might ask the coutnless others that
did implement amd's tips for gcc and found that the code only gets slower.
For a starter, browse the gcc lists.
> 2. K6 has very slow MMX unit when it used as storage unit only. But MMX technology was introduced as
it's also slow as it lacks the parallelism other cpus (including the k7)
support for mmx. so yes, the mxx unix is slow.
> >> Avoid long instruction length Use x86 instructions that are
> >> less than eight bytes in length. An x86 instruction that is longer
> >>
> >> Clear registers using MOV reg, 0 instead of XOR reg, reg
> >
> >is mov 0, reg longer than seven bytes?
>
> See TEST#4 and TEST#5.
What is your pointm that doesn't answer my question. Indeed, mov reg, 0
is not longer than 7 bytes, so if it's true what you claim this is yet
another bug in the amd manual.
--
-----==- |
----==-- _ |
---==---(_)__ __ ____ __ Marc Lehmann +--
--==---/ / _ \/ // /\ \/ / pcg AT goof DOT com |e|
-=====/_/_//_/\_,_/ /_/\_\ XX11-RIPE --+
The choice of a GNU generation |
|
- Raw text -