i686 build of Fedora Core
H. Peter Anvin
hpa at zytor.com
Wed Jul 28 21:33:41 UTC 2004
Alan Cox wrote:
> On Wed, Jul 28, 2004 at 09:50:11PM +0200, Angelo wrote:
>
>>Sure, but what can do optimization without using the *hardcoded* cpu
>>instructions inserted in the architecture ?
>
>
> 90% of the optimisation work is scheduling instructions - picking which one
> to use and how to lay them out. The actual instructions added by 486 and
> later are not much use. Essentially it adds
>
> bswap - which nothing uses except kernel net code
> cmov - which is now slower than avoiding it
>
> and a collection of locked operations that are great for threading and
> are a big reason glibc has a i686 version since it contains all the pthreads
> stuff.
>
Actually, the big ones are MMX/MMXEXT/SSE-{1,2,3}. In particular sse2, since
sse2 can be used for all floating-point (-msse2 -mfpmath=sse), and on some
CPUs it can be 2x or more faster than using x87.
-hpa
More information about the devel
mailing list