[MPlayer-dev-eng] [PATCH] Make mp3lib SIMD optimizations work on AMD64, the Finale

Dominik 'Rathann' Mierzejewski dominik at rangers.eu.org
Thu Jun 7 11:38:55 CEST 2007


On Thursday, 07 June 2007 at 03:42, Zuxy Meng wrote:
> Hi,
> 
> 2007/6/6, Reimar Döffinger <Reimar.Doeffinger at stud.uni-karlsruhe.de>:
> >Hello,
> >On Wed, Jun 06, 2007 at 05:11:59PM +0800, Zuxy Meng wrote:
> >[...]
> >>  SRCS_COMMON = sr1.c
> >> +ifeq ($(TARGET_ARCH_X86),yes)
> >> +SRCS_COMMON-$(TARGET_MMX)     += decode_MMX.c
> >> +SRCS_COMMON-$(TARGET_SSE)     += dct64_sse.c
> >> +endif
> >
> >Is the TARGET_ARCH_X86 check really needed? I'd say TARGET_MMX or
> >TARGET_SSE implies it...
> 
> Sure they aren't needed. Hence the improved patch.
> 
> -- 
> Zuxy
> Beauty is truth,
> While truth is beauty.
> PGP KeyID: E8555ED6

> Index: mp3lib/sr1.c
> ===================================================================
> --- mp3lib/sr1.c	?????? 23483??
> +++ mp3lib/sr1.c	????????????
> @@ -32,8 +32,9 @@
>  
>  #include "libvo/fastmemcpy.h"
>  
> -#ifdef ARCH_X86_32
> -#define CAN_COMPILE_X86_ASM
> +#ifdef ARCH_X86_64
> +#undef HAVE_3DNOW
> +#undef HAVE_3DNOWEX
>  #endif

I still don't like the above. It's unintuitive, because my AMD64 has both
3DNow and 3DNowExt.

Regards,
R.

-- 
MPlayer developer and RPMs maintainer: http://mplayerhq.hu http://rpm.livna.org
There should be a science of discontent. People need hard times and
oppression to develop psychic muscles.
	-- from "Collected Sayings of Muad'Dib" by the Princess Irulan



More information about the MPlayer-dev-eng mailing list