/external/libvpx/libvpx/vp8/common/arm/neon/ |
shortidct4x4llm_neon.c | 27 int16x8_t q1s16, q2s16, q3s16, q4s16; local 40 q3s16 = vqdmulhq_n_s16(q2s16, sinpi8sqrt2); 46 q3s16 = vshrq_n_s16(q3s16, 1); 49 q3s16 = vqaddq_s16(q3s16, q2s16); 52 d10 = vqsub_s16(vget_low_s16(q3s16), vget_high_s16(q4s16)); // c1 53 d11 = vqadd_s16(vget_high_s16(q3s16), vget_low_s16(q4s16)); // d1 71 q3s16 = vqdmulhq_n_s16(q2s16, sinpi8sqrt2); 77 q3s16 = vshrq_n_s16(q3s16, 1) [all...] |
iwalsh_neon.c | 16 int16x8_t q0s16, q1s16, q2s16, q3s16; local 32 q3s16 = vcombine_s16(d6s16, d7s16); 34 q0s16 = vaddq_s16(q2s16, q3s16); 35 q1s16 = vsubq_s16(q2s16, q3s16); 52 q3s16 = vcombine_s16(d6s16, d7s16); 56 q0s16 = vaddq_s16(q2s16, q3s16); 57 q1s16 = vsubq_s16(q2s16, q3s16);
|
loopfiltersimplehorizontaledge_neon.c | 21 int16x8_t q2s16, q3s16, q13s16; local 54 q3s16 = vsubl_s8(vget_high_s8(vreinterpretq_s8_u8(q7u8)), 61 q3s16 = vmulq_s16(q3s16, q13s16); 67 q3s16 = vaddw_s8(q3s16, vget_high_s8(q4s8)); 70 d9s8 = vqmovn_s16(q3s16);
|
sixtappredict_neon.c | 38 int16x8_t q3s16, q4s16, q5s16, q6s16, q7s16; local 112 q3s16 = vreinterpretq_s16_u16(q3u16); 117 q5s16 = vqaddq_s16(q5s16, q3s16); 381 q3s16 = vreinterpretq_s16_u16(q3u16); 386 q5s16 = vqaddq_s16(q5s16, q3s16); 416 int16x8_t q3s16, q4s16, q5s16, q6s16, q7s16; local 486 q3s16 = vreinterpretq_s16_u16(q3u16); 495 q7s16 = vqaddq_s16(q7s16, q3s16); 598 q3s16 = vreinterpretq_s16_u16(q3u16); 607 q7s16 = vqaddq_s16(q7s16, q3s16); 819 int16x8_t q3s16, q4s16, q5s16, q6s16, q7s16; local 1295 int16x8_t q3s16, q4s16, q5s16, q6s16, q7s16, q8s16, q9s16, q10s16; local [all...] |
/external/libvpx/libvpx/vpx_dsp/arm/ |
idct32x32_add_neon.c | 265 int16x8_t q3s16, 283 q5s16 = vaddq_s16(q3s16, q0s16); 284 q6s16 = vsubq_s16(q3s16, q0s16); 291 q3s16 = vaddq_s16(q11s16, q0s16); 305 q5s16 = vaddq_s16(q3s16, q0s16); 306 q6s16 = vsubq_s16(q3s16, q0s16); 313 q3s16 = vaddq_s16(q13s16, q0s16); 327 q5s16 = vaddq_s16(q3s16, q0s16); 328 q6s16 = vsubq_s16(q3s16, q0s16); 335 q3s16 = vaddq_s16(q15s16, q0s16) 466 int16x8_t q0s16, q1s16, q2s16, q3s16, q4s16, q5s16, q6s16, q7s16; local [all...] |
idct16x16_add_neon.c | 95 int16x8_t q0s16, q1s16, q2s16, q3s16, q4s16, q5s16, q6s16, q7s16; local 239 q3s16 = vsubq_s16(q8s16, q11s16); 264 q11s16 = vaddq_s16(q3s16, q4s16); 265 q12s16 = vsubq_s16(q3s16, q4s16); 340 int16x8_t q0s16, q1s16, q2s16, q3s16, q4s16, q5s16, q6s16, q7s16; local 468 q3s16 = vcombine_s16(d6s16, d7s16); 474 q10s16 = vsubq_s16(q3s16, q2s16); 475 q11s16 = vaddq_s16(q2s16, q3s16); 511 q3s16 = q11s16; 533 q8s16 = vaddq_s16(q0s16, q3s16); 1034 int16x8_t q0s16, q1s16, q2s16, q3s16, q4s16, q5s16, q6s16, q7s16; local [all...] |
idct8x8_add_neon.c | 98 int16x8_t q0s16, q1s16, q2s16, q3s16, q4s16, q5s16, q6s16, q7s16; local 200 q3s16 = vsubq_s16(*q9s16, *q15s16); 233 *q11s16 = vaddq_s16(q3s16, q4s16); 234 *q12s16 = vsubq_s16(q3s16, q4s16); 363 int16x8_t q0s16, q1s16, q2s16, q3s16, q4s16, q5s16, q6s16, q7s16; local 414 q3s16 = vsubq_s16(q9s16, q15s16); 448 q11s16 = vaddq_s16(q3s16, q4s16); 449 q12s16 = vsubq_s16(q3s16, q4s16);
|
/external/libvpx/libvpx/vp9/common/arm/neon/ |
vp9_iht4x4_add_neon.c | 59 int16x8_t *q3s16) { 62 *q3s16 = vdupq_n_s16(sinpi_3_9); 111 int16x8_t *q3s16, 117 d6s16 = vget_low_s16(*q3s16); 162 int16x8_t q3s16, q8s16, q9s16; local 180 GENERATE_SINE_CONSTANTS(&d3s16, &d4s16, &d5s16, &q3s16); 189 IADST4x4_1D(&d3s16, &d4s16, &d5s16, &q3s16, &q8s16, &q9s16); 194 GENERATE_SINE_CONSTANTS(&d3s16, &d4s16, &d5s16, &q3s16); 197 IADST4x4_1D(&d3s16, &d4s16, &d5s16, &q3s16, &q8s16, &q9s16); 207 GENERATE_SINE_CONSTANTS(&d3s16, &d4s16, &d5s16, &q3s16); [all...] |
vp9_iht8x8_add_neon.c | 116 int16x8_t q0s16, q1s16, q2s16, q3s16, q4s16, q5s16, q6s16, q7s16; local 218 q3s16 = vsubq_s16(*q9s16, *q15s16); 251 *q11s16 = vaddq_s16(q3s16, q4s16); 252 *q12s16 = vsubq_s16(q3s16, q4s16);
|
/external/libvpx/libvpx/vp8/encoder/arm/neon/ |
shortfdct_neon.c | 126 int16x8_t q0s16, q1s16, q2s16, q3s16; local 145 q3s16 = vld1q_s16(input); 150 vreinterpretq_s32_s16(q3s16)); 192 q3s16 = vcombine_s16(d6s16, d7s16); 201 vreinterpretq_s32_s16(q3s16)); 262 q3s16 = vcombine_s16(d6s16, d7s16); 267 vst1q_s16(output + 24, q3s16);
|