| /external/libvpx/libvpx/vp8/common/arm/neon/ |
| H A D | iwalsh_neon.c | 16 int16x8_t q0s16, q1s16, q2s16, q3s16; local
23 q1s16 = vld1q_s16(input + 8);
26 d4s16 = vadd_s16(vget_low_s16(q0s16), vget_high_s16(q1s16));
27 d6s16 = vadd_s16(vget_high_s16(q0s16), vget_low_s16(q1s16));
28 d5s16 = vsub_s16(vget_low_s16(q0s16), vget_high_s16(q1s16));
29 d7s16 = vsub_s16(vget_high_s16(q0s16), vget_low_s16(q1s16));
35 q1s16 = vsubq_s16(q2s16, q3s16);
38 vreinterpret_s32_s16(vget_low_s16(q1s16)));
40 vreinterpret_s32_s16(vget_high_s16(q1s16)));
57 q1s16
[all...] |
| H A D | shortidct4x4llm_neon.c | 27 int16x8_t q1s16, q2s16, q3s16, q4s16; local
37 q1s16 = vcombine_s16(d2, d4); // Swap d3 d4 here
43 d12 = vqadd_s16(vget_low_s16(q1s16), vget_high_s16(q1s16)); // a1
44 d13 = vqsub_s16(vget_low_s16(q1s16), vget_high_s16(q1s16)); // b1
68 q1s16 = vcombine_s16(v2tmp2.val[0], v2tmp3.val[0]);
74 d12 = vqadd_s16(vget_low_s16(q1s16), vget_high_s16(q1s16)); // a1
75 d13 = vqsub_s16(vget_low_s16(q1s16), vget_high_s1
[all...] |
| /external/libvpx/libvpx/vpx_dsp/arm/ |
| H A D | idct32x32_add_neon.c | 276 int16x8_t q0s16, q1s16, q4s16, q5s16; local
281 LOAD_FROM_OUTPUT(15, 30, 31, q0s16, q1s16);
282 q4s16 = vaddq_s16(q2s16, q1s16);
285 q7s16 = vsubq_s16(q2s16, q1s16);
289 LOAD_FROM_OUTPUT(1, 12, 13, q0s16, q1s16);
290 q2s16 = vaddq_s16(q10s16, q1s16);
293 q5s16 = vsubq_s16(q10s16, q1s16);
295 LOAD_FROM_OUTPUT(13, 18, 19, q0s16, q1s16);
296 q8s16 = vaddq_s16(q4s16, q1s16);
299 q7s16 = vsubq_s16(q4s16, q1s16);
378 int16x8_t q0s16, q1s16, q4s16, q5s16; local
466 int16x8_t q0s16, q1s16, q2s16, q3s16, q4s16, q5s16, q6s16, q7s16; local
[all...] |
| H A D | idct16x16_add_neon.c | 95 int16x8_t q0s16, q1s16, q2s16, q3s16, q4s16, q5s16, q6s16, q7s16; local
237 q1s16 = vaddq_s16(q9s16, q10s16);
262 q9s16 = vaddq_s16(q1s16, q6s16);
267 q14s16 = vsubq_s16(q1s16, q6s16);
340 int16x8_t q0s16, q1s16, q2s16, q3s16, q4s16, q5s16, q6s16, q7s16; local
428 q1s16 = vcombine_s16(d2s16, d3s16);
472 q9s16 = vsubq_s16(q0s16, q1s16);
473 q0s16 = vaddq_s16(q0s16, q1s16);
508 q1s16 = vcombine_s16(d2s16, d3s16);
534 q9s16 = vaddq_s16(q1s16, q2s1
890 int16x8_t q0s16, q1s16, q2s16, q4s16, q5s16, q6s16, q7s16; local
1034 int16x8_t q0s16, q1s16, q2s16, q3s16, q4s16, q5s16, q6s16, q7s16; local
[all...] |
| H A D | idct8x8_add_neon.c | 98 int16x8_t q0s16, q1s16, q2s16, q3s16, q4s16, q5s16, q6s16, q7s16; local
198 q1s16 = vaddq_s16(*q11s16, *q13s16);
231 *q9s16 = vaddq_s16(q1s16, q6s16);
236 *q14s16 = vsubq_s16(q1s16, q6s16);
363 int16x8_t q0s16, q1s16, q2s16, q3s16, q4s16, q5s16, q6s16, q7s16; local
383 q1s16 = vdupq_n_s16(cospi_4_64 * 2);
389 q7s16 = vqrdmulhq_s16(q9s16, q1s16);
391 q1s16 = vdupq_n_s16(cospi_12_64 * 2);
397 q6s16 = vqrdmulhq_s16(q11s16, q1s16);
400 q1s16
[all...] |
| H A D | intrapred_neon.c | 602 int16x8_t q1s16; local
611 q1s16 = vaddq_s16(vreinterpretq_s16_u16(q1u16),
613 d0u8 = vqmovun_s16(q1s16);
665 int16x8_t q0s16, q1s16, q8s16, q11s16; local
680 q1s16 = vaddq_s16(vreinterpretq_s16_u16(q0u16),
688 d2u8 = vqmovun_s16(q1s16);
701 q1s16 = vaddq_s16(vreinterpretq_s16_u16(q0u16),
709 d2u8 = vqmovun_s16(q1s16);
|
| /external/libvpx/libvpx/vp8/encoder/arm/neon/ |
| H A D | shortfdct_neon.c | 20 int16x8_t q0s16, q1s16; local
111 q1s16 = vcombine_s16(d2s16, d3s16);
114 vst1q_s16(output + 8, q1s16);
126 int16x8_t q0s16, q1s16, q2s16, q3s16; local
141 q1s16 = vld1q_s16(input);
149 v2tmp3 = vtrnq_s32(vreinterpretq_s32_s16(q1s16),
191 q1s16 = vcombine_s16(d2s16, d3s16);
200 v2tmp3 = vtrnq_s32(vreinterpretq_s32_s16(q1s16),
215 q1s16 = vsubq_s16(q11s16, q12s16);
222 d2s16 = vget_low_s16(q1s16);
[all...] |
| H A D | vp8_shortwalsh4x4_neon.c | 28 int16x8_t q0s16, q1s16; local
123 q1s16 = vcombine_s16(d2s16, d3s16);
126 vst1q_s16(output + 8, q1s16);
|
| /external/libvpx/libvpx/vp9/common/arm/neon/ |
| H A D | vp9_iht8x8_add_neon.c | 116 int16x8_t q0s16, q1s16, q2s16, q3s16, q4s16, q5s16, q6s16, q7s16; local
216 q1s16 = vaddq_s16(*q11s16, *q13s16);
249 *q9s16 = vaddq_s16(q1s16, q6s16);
254 *q14s16 = vsubq_s16(q1s16, q6s16);
|