On Thu, 06 Dec 2001 14:20:56 +0100, "Thorsten Froehlich"
<tho### [at] trfde> wrote:

>In article <3c0f635a.4744504@news.povray.org> , ken### [at] uniplanit (Angelo 
>'kENpEX' Pesce) wrote:
>
>> Premature. Now I want to know, are U going to change ray-sphere
>> intersection? Do U see new techniques to do 3d vector cross product?
>> dot product? Matrix-vector product?
>
>No, but neither of those will even benefit a single cycle if you do them in
>hand written assembler.  They are too small and linear to offer any improved
>speed.  Any instruction scheduler of a reasonable compiler (which are the
>compilers the POV-Team is using) will be better than a human optimizing the
>same code.  To the contrary, most compilers have global optimization
>strategies that you effectively kill by adding bits and pieces of inline
>assembler because most compilers won't for good reasons not touch (read:
>optimize) user supplied hand written assembler code...
Well... Actually my asm routines tell me something different... Mabye
you should read something like agner fog asm optimization. Many guys
think that asm is not useful at all for those things... Well, at last
on intel plattform you're really wrong...

Post a reply to this message