[x264-devel] Re: [PATCH] IDCT8 and SA8D routines in Altivec

[Guillaume POIRIER]

Sorry, I missunderstood your previous post.

Manuel Rommel a écrit :

> And I have a question about the use of VEC_ABS in your SA8D code, for 
> example:
> 
>> +    vec_s16_t abs6v = VEC_ABS(sa8d6v);
> 
> As far as I understand, this gets transformed to
> 
> vec_s16_t abs6v =     sa8d6v     = vec_max( sa8d6v, vec_sub( zero_s16v, 
> sa8d6v ) );
> 
> This shouldn't alter the output but it could be possible to get rid of 
> some instructions, if I understand it right.

No, it shouldn't change anything if the compiler does its job (remember, 
PPC has 2 op, 1 result instruction set, so operands don't always get 
overwritten like on x86. Remember also that when you code with 
intrinsics, there are lots of different passes in the compiler before it 
emits ASM, so this kind of optimization aren't needed).

Guillaume

-- 
This is the x264-devel mailing-list
To unsubscribe, go to: http://developers.videolan.org/lists.html