[Ffmpeg-devel] [PATCH] Snow mmx+sse2 asm optimizations

Michael Niedermayer michaelni
Sat Mar 11 01:50:32 CET 2006


Hi

On Fri, Mar 10, 2006 at 05:29:10PM -0500, Robert Edele wrote:
> 
> > OK, I'll wait. Could you look over Robert's patch then?... It (should be) 
> > production ready...
> > 
> I think I've figured out what Michael wants. I think he wanted the
> add_yblock function to be trimmed down. I've cleaned them up and
> offloaded the repetitive code into marcos, shrinking the patch by about
> 16kB while marginally speeding it up.
> 
> Michael, if there's more you want done before you're willing to commit
> it, please speak up.

[...]

> +             "add $"PTR_SIZE"*1, %1          \n\t"\
> +             "add %%"REG_c", %0              \n\t"\
> +             "dec %%"REG_b"                  \n\t"\
> +             "jnz 1b                         \n\t"\
> +             :\
> +             :\
> +             "m"(dst8),"m"(dst_array),"rm"((long)(src_x<<2)),"m"(obmc),"a"(block),"m"((long)b_h),"rm"((long)src_stride):\
> +             "%"REG_b"","%"REG_c"","%"REG_S"","%"REG_D"","%"REG_d"");

%0 and %1 are changed but on the input list

[...]

-- 
Michael





More information about the ffmpeg-devel mailing list