static av_always_inline void h264_idct_dc_add_internal(uint8_t *dst, int16_t *block, int stride, int size)
{
vec_s16 dc16;
vec_u8 dcplus, dcminus, v0, v1, v2, v3, aligner;
vec_s32 v_dc32;
LOAD_ZERO;
DECLARE_ALIGNED(16, int, dc);
int i;
dc = (block[0] + 32) >> 6;
block[0] = 0;
v_dc32 = vec_lde(0, &dc);
dc16 = VEC_SPLAT16((vec_s16)v_dc32, 1);
if (size == 4)
dc16 = VEC_SLD16(dc16, zero_s16v, 8);
dcplus = vec_packsu(dc16, zero_s16v);
dcminus = vec_packsu(vec_sub(zero_s16v, dc16), zero_s16v);
#if HAVE_BIGENDIAN
aligner = vec_lvsr(0, dst);
dcplus = vec_perm(dcplus, dcplus, aligner);
dcminus = vec_perm(dcminus, dcminus, aligner);
#endif
for (i = 0; i < size; i += 4) {
v0 = DST_LD(0, dst+0*stride);
v1 = DST_LD(0, dst+1*stride);
v2 = DST_LD(0, dst+2*stride);
v3 = DST_LD(0, dst+3*stride);
v0 = vec_adds(v0, dcplus);
v1 = vec_adds(v1, dcplus);
v2 = vec_adds(v2, dcplus);
v3 = vec_adds(v3, dcplus);
v0 = vec_subs(v0, dcminus);
v1 = vec_subs(v1, dcminus);
v2 = vec_subs(v2, dcminus);
v3 = vec_subs(v3, dcminus);
VEC_ST(v0, 0, dst+0*stride);
VEC_ST(v1, 0, dst+1*stride);
VEC_ST(v2, 0, dst+2*stride);
VEC_ST(v3, 0, dst+3*stride);
dst += 4*stride;
}
}
static void yuv2planeX_16_altivec(const int16_t *filter, int filterSize,
const int16_t **src, uint8_t *dest,
const uint8_t *dither, int offset, int x)
{
register int i, j;
LOCAL_ALIGNED(16, int, val, [16]);
vector signed int vo1, vo2, vo3, vo4;
vector unsigned short vs1, vs2;
vector unsigned char vf;
vector unsigned int altivec_vectorShiftInt19 =
vec_add(vec_splat_u32(10), vec_splat_u32(9));
for (i = 0; i < 16; i++)
val[i] = dither[(x + i + offset) & 7] << 12;
vo1 = vec_ld(0, val);
vo2 = vec_ld(16, val);
vo3 = vec_ld(32, val);
vo4 = vec_ld(48, val);
for (j = 0; j < filterSize; j++) {
unsigned int joffset=j<<1;
unsigned int xoffset=x<<1;
vector unsigned char perm;
vector signed short l1,vLumFilter;
LOAD_FILTER(vLumFilter,filter);
vLumFilter = vec_splat(vLumFilter, 0);
LOAD_L1(l1,src[j],perm);
yuv2planeX_8(vo1, vo2, l1, src[j], x, perm, vLumFilter);
yuv2planeX_8(vo3, vo4, l1, src[j], x + 8, perm, vLumFilter);
}
vo1 = vec_sra(vo1, altivec_vectorShiftInt19);
vo2 = vec_sra(vo2, altivec_vectorShiftInt19);
vo3 = vec_sra(vo3, altivec_vectorShiftInt19);
vo4 = vec_sra(vo4, altivec_vectorShiftInt19);
vs1 = vec_packsu(vo1, vo2);
vs2 = vec_packsu(vo3, vo4);
vf = vec_packsu(vs1, vs2);
VEC_ST(vf, 0, dest);
}
