Lines Matching refs:q3
83 @ Input columns: q0 q1 q2 q3
84 @ Output columns: q0 q1 q2 q3
100 vmul.i16 q11, q3, d0[1] @ t4 = 22 * (src[3])
105 vmla.i16 q10, q3, d1[0] @ t3 += 10 * src[3]
109 vhsub.s16 q3, q8, q10 @ dst[3] = (t1 - t3) >> 1
115 vshr.s16 q3, q3, #(\rshift - 1) @ dst[3] >>= (rshift - 1)
157 @ q3 src[24]|src[56]
163 vshl.i16 q12, q3, #4 @ temp3a|temp4a = 16 * src[24]|src[56]
164 vswp d6, d7 @ q3 = src[56]|src[24]
166 vshl.i16 q2, q3, #2 @ temp1|temp2 = 4 * (src[56]|src[24])
167 vswp d3, d6 @ q1 = src[40]|src[56], q3 = src[8]|src[24]
170 vmul.i16 q12, q3, d0[3] @ temp3|temp4 = 9 * src[8]|src[24]
172 vswp d6, d7 @ q3 = src[24]|src[8]
188 vmla.i16 q8, q3, d0[2] @ t1|t2 += 15 * src[24]|src[8]
202 vadd.i16 q3, q15, q10 @ line[7,6] = t5|t6 + 1
212 vhsub.s16 q3, q3, q8 @ line[7,6] = (t5|t6 - t1|t2 + 1) >> 1
215 vhsub.s16 q3, q10, q8 @ line[7,6] = (t5|t6 - t1|t2) >> 1
221 vshr.s16 q11, q3, #(\rshift - 1)
243 vshl.i16 q3, q14, #3 @ temp4 = 16/2 * src[48]
254 vsub.i16 q2, q2, q3 @ t4 = 6/2 * src[16] - 16/2 * src[48]
256 @ unused: q3
260 @ unused: q3, q10, q14
266 @ q3
276 @ unused: q3, q10, q14
278 vshl.i16 q3, q9, #4 @ t1 = 16 * src[8]
285 vmla.i16 q3, q11, d0[0] @ t1 += 15 * src[24]
290 vmla.i16 q3, q13, d0[1] @ t1 += 9 * src[40]
309 vmla.i16 q3, q15, d1[1] @ t1 += 4 * src[56]
316 @ t1 q3
342 vadd.i16 q8, q11, q3 @ q8 = t5half + t1
344 vsub.i16 q3, q3, q12 @ q3 = t1 - 1
359 vsub.i16 q3, q11, q3 @ q3 = t5half - t1 + 1
367 vhadd.s16 q3, q11, q3 @ q3 = (t5half + t5half - t1 + 1) >> 1
392 vshr.s16 q15, q3, #(\rshift - 1) @ q15 = line[7]
432 vld1.64 {q2-q3}, [r2,:128]
434 transpose16 q0, q1, q2, q3 @ transpose rows to columns
468 transpose16 q0, q1, q2, q3 @ turn columns into rows
474 @ row[3] q3
499 vaddw.u8 q3, q3, d31 @ line[3] += dest[3]
504 vqmovun.s16 d3, q3 @ line[3]
532 @ dst[3] = q3
534 transpose16 q0, q1, q2, q3 @ Transpose rows (registers) into columns
585 vmul.i16 q3, q14, q1 @ t3|t4 = 22 * (src[1]|src[3])
590 vadd.i16 q0, q2, q3 @ dst[0,2] = (t1|t2 + t3|t4)
591 vsub.i16 q1, q2, q3 @ dst[3,1] = (t1|t2 - t3|t4)
792 @ q3 [hv] src[stride * 2]
864 vext.16 q3, q0, q1, #3
1073 vaddw.u8 q3, q1, d0
1077 vqmovun.s16 d0, q3
1114 vaddw.u8 q3, q1, d0
1118 vqmovun.s16 d0, q3
1155 vaddw.u8 q3, q1, d1
1158 vqmovun.s16 d1, q3
1191 vmovl.u8 q3, d7 @ P4
1231 vqmovun.s16 d0, q3
1255 vshll.u8 q3, d2, #1 @ 2*P1, 2*P5
1259 vmls.i16 q3, q8, d0[1] @ 2*P1-5*P2, 2*P5-5*P6
1262 vmla.i16 q3, q9, d0[1] @ 2*P1-5*P2+5*P3, 2*P5-5*P6+5*P7
1263 vmls.i16 q3, q2, d0[0] @ 2*P1-5*P2+5*P3-2*P4, 2*P5-5*P6+5*P7-2*P8
1269 vrshr.s16 q3, q3, #3
1273 vabs.s16 q3, q3 @ a1, a2
1331 vmovl.u8 q3, d7 @ P4
1336 vsub.i16 q13, q3, q1 @ P4-P5
1337 vmls.i16 q2, q3, d0[1] @ 2*P3-5*P4
1341 vmls.i16 q9, q3, d0[0] @ 2*P1-5*P2+5*P3-2*P4
1374 vmls.i16 q3, q0, q2 @ invert d depending on clip_sign & a0_sign, or zero it if they match, and accumulate into P4
1376 vqmovun.s16 d0, q3
1410 vtrn.8 q3, q8 @ P1[4], P1[5], P3[4]... P1[6], P1[7], P3[6]... P2[4], P2[5], P4[4]... P2[6], P2[7], P4[6]...
1429 vmovl.u8 q3, d6 @ P5
1431 vsub.i16 q12, q2, q3 @ P4-P5
1438 vmla.i16 q1, q3, d0[1] @ 2*P3-5*P4+5*P5
1466 vmla.i16 q3, q0, q1 @ invert d depending on clip_sign & a0_sign, or zero it if they match, and accumulate into P5
1468 vqmovun.s16 d1, q3
1496 vld1.64 {q3}, [r3 :128], r1 @ P2
1510 vmovl.u8 q3, d7 @ P2[8..15]
1515 vmls.i16 q2, q3, d0[1] @ 2*P1[8..15]-5*P2[8..15]
1516 vshll.u8 q3, d10, #1 @ 2*P3[0..7]
1525 vmls.i16 q3, q14, d0[1] @ 2*P3[0..7]-5*P4[0..7]
1536 vmla.i16 q3, q6, d0[1] @ 2*P3[0..7]-5*P4[0..7]+5*P5[0..7]
1546 vmls.i16 q3, q12, d0[0] @ 2*P3[0..7]-5*P4[0..7]+5*P5[0..7]-2*P6[0..7]
1554 vrshr.s16 q3, q3, #3
1563 vabs.s16 q2, q3 @ a0[0..7]
1565 vshr.s16 q3, q3, #8 @ a0_sign[0..7]
1566 vsub.i16 q3, q15, q3 @ clip_sign[0..7] - a0_sign[0..7]
1602 vmls.i16 q14, q2, q3 @ invert d[0..7] depending on clip_sign[0..7] & a0_sign[0..7], or zero it if they match, and accumulate into P4[0..7]
1607 vmla.i16 q6, q2, q3 @ invert d[0..7] depending on clip_sign[0..7] & a0_sign[0..7], or zero it if they match, and accumulate into P5[0..7]
1683 vmovl.u8 q3, d7 @ P6[8..15]
1690 vmls.i16 q13, q3, d0[1] @ 2*P5[8..15]-5*P6[8..15]
1716 vmls.i16 q15, q3, d0[0] @ 2*P3[8..15]-5*P4[8..15]+5*P5[8..15]-2*P6[8..15]
1717 vabs.s16 q3, q11 @ a2[0..7]
1722 vcge.s16 q13, q10, q3 @ test a1[0..7] >= a2[0..7]
1727 vbsl q13, q3, q10 @ a3[0..7]
1728 vcge.s16 q3, q2, q12 @ test a1[8..15] >= a2[8.15]
1731 vbsl q3, q12, q2 @ a3[8..15]
1738 vqsub.u16 q15, q10, q3 @ a0[8..15] >= a3[8..15] ? a0[8..15]-a3[8..15] : 0 (a0 > a3 in all cases where filtering is enabled, so makes more sense to subtract this way round than the opposite and then taking the abs)
1739 vcge.s16 q3, q3, q10 @ test a3[8..15] >= a0[8..15]
1756 vorr q3, q10, q3 @ test clip[8..15] == 0 || a0[8..15] >= pq || a3[8..15] >= a0[8..15]
1871 vorr q3, q13, q15
1896 vorr q3, q13, q15