86     __m128i c0, c2, c4, c6, c8;  in rsdIntrinsicConvolve3x3_K()  local
93     c0 = _mm_shuffle_epi32(x, 0x00);  in rsdIntrinsicConvolve3x3_K()
116         o0 = _mm_madd_epi16(_mm_unpacklo_epi16(p0, p1), c0);  in rsdIntrinsicConvolve3x3_K()
117         o1 = _mm_madd_epi16(_mm_unpacklo_epi16(p1, p2), c0);  in rsdIntrinsicConvolve3x3_K()
154     __m128i c0, c1, c2, c3;  in rsdIntrinsicColorMatrix4x4_K()  local
160     c0 = _mm_loadl_epi64((const __m128i *)(coef+0));  in rsdIntrinsicColorMatrix4x4_K()
162     c0 = _mm_unpacklo_epi16(c0, c1);  in rsdIntrinsicColorMatrix4x4_K()
173         x2 = _mm_madd_epi16(xy, _mm_shuffle_epi32(c0, 0x00));  in rsdIntrinsicColorMatrix4x4_K()
174         y2 = _mm_madd_epi16(xy, _mm_shuffle_epi32(c0, 0x55));  in rsdIntrinsicColorMatrix4x4_K()
175         z2 = _mm_madd_epi16(xy, _mm_shuffle_epi32(c0, 0xaa));  in rsdIntrinsicColorMatrix4x4_K()
176         w2 = _mm_madd_epi16(xy, _mm_shuffle_epi32(c0, 0xff));  in rsdIntrinsicColorMatrix4x4_K()
210     __m128i c0, c1, c2, c3;  in rsdIntrinsicColorMatrix3x3_K()  local
216     c0 = _mm_loadl_epi64((const __m128i *)(coef+0));  in rsdIntrinsicColorMatrix3x3_K()
218     c0 = _mm_unpacklo_epi16(c0, c1);  in rsdIntrinsicColorMatrix3x3_K()
229         x2 =  _mm_madd_epi16(xy, _mm_shuffle_epi32(c0, 0x00));  in rsdIntrinsicColorMatrix3x3_K()
230         y2 =  _mm_madd_epi16(xy, _mm_shuffle_epi32(c0, 0x55));  in rsdIntrinsicColorMatrix3x3_K()
231         z2 =  _mm_madd_epi16(xy, _mm_shuffle_epi32(c0, 0xaa));  in rsdIntrinsicColorMatrix3x3_K()
262     __m128i c0, c1, c2, c3;  in rsdIntrinsicColorMatrixDot_K()  local
268     c0 = _mm_loadl_epi64((const __m128i *)(coef+0));  in rsdIntrinsicColorMatrixDot_K()
269     c0 = _mm_shufflelo_epi16(c0, 0);  in rsdIntrinsicColorMatrixDot_K()
272     c0 = _mm_unpacklo_epi16(c0, c1);  in rsdIntrinsicColorMatrixDot_K()
286         x2 =  _mm_madd_epi16(xy, c0);  in rsdIntrinsicColorMatrixDot_K()
417     __m128i c0, c1, c2, c3, c4;  in rsdIntrinsicYuv_K()  local
422     c0 = _mm_set1_epi32(param[0]);  /*  298 */  in rsdIntrinsicYuv_K()
443         Y = mullo_epi32(Y, c0);  in rsdIntrinsicYuv_K()
479     __m128i c0, c1, c2, c3, c4;  in rsdIntrinsicYuvR_K()  local
484     c0 = _mm_set1_epi32(param[0]);  /*  298 */  in rsdIntrinsicYuvR_K()
505         Y = mullo_epi32(Y, c0);  in rsdIntrinsicYuvR_K()
541     __m128i c0, c1, c2, c3, c4;  in rsdIntrinsicYuv2_K()  local
546     c0 = _mm_set1_epi32(param[0]);  /*  298 */  in rsdIntrinsicYuv2_K()
566         Y = mullo_epi32(Y, c0);  in rsdIntrinsicYuv2_K()
604     __m128i c0, c2, c4, c6, c8, c10, c12;  in rsdIntrinsicConvolve5x5_K()  local
616     c0  = _mm_shuffle_epi32(x, 0x00);  in rsdIntrinsicConvolve5x5_K()
689         o0 =                   _mm_madd_epi16( _mm_unpacklo_epi16(p0, p1),  c0);  in rsdIntrinsicConvolve5x5_K()
704         o1 =                   _mm_madd_epi16( _mm_unpacklo_epi16(p1, p2),  c0);  in rsdIntrinsicConvolve5x5_K()
719         o2 =                   _mm_madd_epi16( _mm_unpacklo_epi16(p2,p3),  c0);  in rsdIntrinsicConvolve5x5_K()
734         o3 =                   _mm_madd_epi16( _mm_unpacklo_epi16(p3,p4),  c0);  in rsdIntrinsicConvolve5x5_K()