static uint8_t tinyAverageMMX(uint8_t *dst, uint8_t *src1, uint8_t *src2,uint32_t l)
{
int delta;
uint32_t ww,rr;
uint8_t *s1,*s2,*d1;
int a1,a2;
s1=src1;
s2=src2;
d1=dst;
ww=l>>2;
rr=l&3;
#ifdef GCC_2_95_X
__asm__(
"pxor %mm7,%mm7"
::
);
#else
__asm__(
"pxor %%mm7,%%mm7"
::
);
#endif
for(int y=0;y<ww;y++)
{
__asm__(
"movd (%0),%%mm0 \n"
"movd (%1),%%mm1 \n"
"punpcklbw %%mm7,%%mm0 \n"
"punpcklbw %%mm7,%%mm1 \n"
"paddw %%mm1,%%mm0 \n"
"psrlw $1,%%mm0 \n"
"packuswb %%mm0, %%mm0\n"
"movd %%mm0,(%2) \n"
: : "r" (s1),"r" (s2),"r"(d1)
);
s1+=4;
s2+=4;
d1+=4;
}
__asm__(
"emms\n"
::
);
if(rr) tinyAverage(d1, s1, s2,rr);
return 1;
}
static uint8_t averageF(ADMImage *tgt,ADMImage *src,ADMImage *srcP)
{
int delta;
uint32_t ww,hh;
uint8_t *s1,*s2,*d1;
int a1,a2,t1;
tinyAverage(YPLANE(tgt),YPLANE(src),YPLANE(srcP),tgt->_width,tgt->_height);
tinyAverage(UPLANE(tgt),UPLANE(src),UPLANE(srcP),tgt->_width>>1,tgt->_height>>1);
tinyAverage(VPLANE(tgt),VPLANE(src),VPLANE(srcP),tgt->_width>>1,tgt->_height>>1);
return 1;
}
static uint8_t tinyAverageMMX(uint8_t *dst, uint8_t *src1, uint8_t *src2,uint32_t l)
{
int delta;
uint32_t ww,rr;
uint8_t *s1,*s2,*d1;
int a1,a2;
s1=src1;
s2=src2;
d1=dst;
ww=l>>2;
rr=l&3;
ADM_CLEAR_MM7();
for(int y=0; y<ww; y++)
{
__asm__ volatile(
"movd (%0),%%mm0 \n"
"movd (%1),%%mm1 \n"
"punpcklbw %%mm7,%%mm0 \n"
"punpcklbw %%mm7,%%mm1 \n"
"paddw %%mm1,%%mm0 \n"
"psrlw $1,%%mm0 \n"
"packuswb %%mm0, %%mm0\n"
"movd %%mm0,(%2) \n"
: : "r" (s1),"r" (s2),"r"(d1)
:"memory"
);
s1+=4;
s2+=4;
d1+=4;
}
ADM_EMMS();
if(rr) tinyAverage(d1, s1, s2,rr);
return 1;
}