Lines Matching refs:q8
65 @ vadd.i16 q0, q8, q10
71 @ vhadd.s16 q0, q8, q10
85 @ Trashes: r12 q8 q9 q10 q11 q12 q13
96 vmov.i16 q8, #\add @ t1 will accumulate here
102 vmla.i16 q8, q12, d0[0] @ t1 = 17 * (temp1) + 4
108 vhadd.s16 q0, q8, q10 @ dst[0] = (t1 + t3) >> 1
109 vhsub.s16 q3, q8, q10 @ dst[3] = (t1 - t3) >> 1
161 vshl.i16 q8, q1, #4 @ t1|t2 = 16 * (src[8]|src[40])
169 vadd.i16 q8, q8, q2 @ t1|t2 += temp1|temp2
171 vmla.i16 q8, q1, d0[3] @ t1|t2 += 9 * (src[40]|src[56])
188 vmla.i16 q8, q3, d0[2] @ t1|t2 += 15 * src[24]|src[8]
209 vhadd.s16 q0, q8, q10 @ line[0,1] = (t1|t2 + 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
219 vshr.s16 q8, q0, #(\rshift - 1)
235 @ Input columns: q8, q9, q10, q11, q12, q13, q14, q15
236 @ Output columns: q8, q9, q10, q11, q12, q13, q14, q15
248 vadd.i16 q10, q8, q12 @ temp1 = src[0] + src[32]
249 vsub.i16 q14, q8, q12 @ temp2 = src[0] - src[32]
250 @ unused: q8, q12
251 vmov.i16 q8, #(\add / 2) @ t1 will accumulate here
257 vmla.i16 q8, q10, q0 @ t1 = 12/2 * temp1 + add
267 @ q8 old t1
298 vadd.i16 q11, q8, q1 @ t5 = t1 + t3
299 vsub.i16 q1, q8, q1 @ t8 = t1 - t3
302 @ unused: q8, q12
308 @ unused: q8
313 @ unused: q0, q8, q15
342 vadd.i16 q8, q11, q3 @ q8 = t5half + t1
358 vhadd.s16 q8, q11, q8 @ q8 = (t5half + t5half + t1) >> 1
384 vshr.s16 q8, q8, #(\rshift - 1) @ q8 = line[0]
397 vld1.64 {q8-q9}, [r0,:128]!
404 @ src[0] q8
417 transpose16_4x4 q8, q9, q10, q11, q12, q13, q14, q15
421 vst1.64 {q8-q9}, [r0,:128]!
549 vaddw.u8 q8, q8, d28 @ line[0,1] += dest[0,1]
554 vqmovun.s16 d16, q8 @ clip(line[0,1])
586 vmla.i16 q2, q13, q8 @ t1|t2 = 17 * (temp1|temp2) + add
1074 vaddw.u8 q8, q1, d1
1078 vqmovun.s16 d1, q8
1115 vaddw.u8 q8, q1, d1
1120 vqmovun.s16 d1, q8
1256 vmovl.u8 q8, d4 @ P2, P6
1259 vmls.i16 q3, q8, d0[1] @ 2*P1-5*P2, 2*P5-5*P6
1319 vshll.u8 q8, d1, #1 @ 2*P5
1330 vmls.i16 q8, q11, d0[1] @ 2*P5-5*P6
1333 vmla.i16 q8, q1, d0[1] @ 2*P5-5*P6+5*P7
1338 vmls.i16 q8, q10, d0[0] @ 2*P5-5*P6+5*P7-2*P8
1344 vrshr.s16 q8, q8, #3
1348 vabs.s16 q8, q8 @ a2
1351 vcge.s16 q14, q9, q8 @ test a1 >= a2
1354 vbsl q14, q8, q9 @ a3
1355 vcge.s16 q8, q15, q12 @ test a0 >= pq
1359 vorr q8, q11, q8 @ test clip == 0 || a0 >= pq
1361 vorr q9, q8, q12 @ test clip == 0 || a0 >= pq || a3 >= a0
1368 vorr q8, q8, q9
1373 vbic q0, q11, q8 @ set each d to zero if it should not be filtered
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]...
1427 vmovl.u8 q8, d17 @ P8
1433 vmls.i16 q11, q8, d0[0] @ 2*P5-5*P6+5*P7-2*P8
1434 vabs.s16 q8, q12
1437 vshr.s16 q8, q8, #1 @ clip
1441 vceq.i16 q13, q8, #0 @ test clip == 0
1460 vcge.s16 q10, q0, q8
1462 vbsl q10, q8, q0 @ FFMIN(d, clip)
1501 vshll.u8 q8, d4, #1 @ 2*P1[0..7]
1508 vmls.i16 q8, q10, d0[1] @ 2*P1[0..7]-5*P2[0..7]
1519 vmla.i16 q8, q14, d0[1] @ 2*P1[0..7]-5*P2[0..7]+5*P3[0..7]
1533 vmls.i16 q8, q14, d0[0] @ 2*P1[0..7]-5*P2[0..7]+5*P3[0..7]-2*P4[0..7]
1537 vrshr.s16 q8, q8, #3
1542 vabs.s16 q8, q8 @ a1[0..7]
1552 vcge.s16 q4, q8, q7 @ test a1[0..7] >= a2[0..7]
1556 vbsl q4, q7, q8 @ a3[0..7]
1558 vshr.s16 q8, q12, #8 @ clip_sign[8..15]
1576 vsub.i16 q5, q8, q5 @ clip_sign[8..15] - a0_sign[8..15]
1577 vceq.i16 q8, q7, #0 @ test clip[8..15] == 0
1580 vorr q8, q8, q13 @ test clip[8..15] == 0 || a0[8..15] >= pq
1587 vorr q4, q8, q4 @ test clip[8..15] == 0 || a0[8..15] >= pq || a3[8..15] >= a0[8..15]
1599 vorr q4, q8, q4
1700 vmovl.u8 q8, d1 @ P5[8..15]
1709 vsub.i16 q7, q6, q8 @ P4[8..15]-P5[8..15]
1713 vmla.i16 q15, q8, d0[1] @ 2*P3[8..15]-5*P4[8..15]+5*P5[8..15]
1772 vmla.i16 q8, q0, q7 @ invert d[8..15] depending on clip_sign[8..15] & a0_sign[8..15], or zero it if they match, and accumulate into P5
1777 vqmovun.s16 d5, q8
1830 vld1.8 {q8, q9}, [r0]!
1831 vbic q12, q8, q0
1832 vext.8 q13, q8, q9, #1
1833 vext.8 q14, q8, q9, #2
1834 vext.8 q15, q8, q9, #3
1866 2: vmov q8, q11
1870 vbic q12, q8, q0
1872 vext.8 q13, q8, q9, #1
1873 vext.8 q14, q8, q9, #2
1874 vext.8 q15, q8, q9, #3
1915 vst1.64 {q8}, [r2]!