| /external/libaom/libaom/aom_dsp/x86/ |
| D | loopfilter_sse2.c | 30 __m128i *q1p1, __m128i *q2p2, in transpose_pq_14_sse2() argument
72 *q2p2 = _mm_unpackhi_epi32( in transpose_pq_14_sse2()
395 __m128i *q6p6, __m128i *q5p5, __m128i *q4p4, __m128i *q3p3, __m128i *q2p2, in lpf_internal_14_dual_sse2() argument
434 work = _mm_max_epu8(abs_diff(*q2p2, *q1p1), abs_diff(*q3p3, *q2p2)); in lpf_internal_14_dual_sse2()
451 flat = _mm_max_epu8(abs_diff(*q2p2, *q0p0), abs_diff(*q3p3, *q0p0)); in lpf_internal_14_dual_sse2()
477 p2_16 = _mm_unpacklo_epi8(*q2p2, zero); in lpf_internal_14_dual_sse2()
482 q2_16 = _mm_unpackhi_epi8(*q2p2, zero); in lpf_internal_14_dual_sse2()
571 *q2p2 = _mm_andnot_si128(flat, *q2p2); in lpf_internal_14_dual_sse2()
573 *q2p2 = _mm_or_si128(*q2p2, flat_q2p2); in lpf_internal_14_dual_sse2()
679 *q2p2 = _mm_andnot_si128(flat2, *q2p2); in lpf_internal_14_dual_sse2()
[all …]
|
| /external/libvpx/libvpx/vpx_dsp/x86/ |
| D | loopfilter_avx2.c | 23 __m128i q7p7, q6p6, q5p5, q4p4, q3p3, q2p2, q1p1, q0p0, p0q0, p1q1; in vpx_lpf_horizontal_16_avx2() local
38 q2p2 = _mm_loadl_epi64((__m128i *)(s - 3 * pitch)); in vpx_lpf_horizontal_16_avx2()
39 q2p2 = _mm_castps_si128( in vpx_lpf_horizontal_16_avx2()
40 _mm_loadh_pi(_mm_castsi128_ps(q2p2), (__m64 *)(s + 2 * pitch))); in vpx_lpf_horizontal_16_avx2()
75 _mm_or_si128(_mm_subs_epu8(q2p2, q1p1), _mm_subs_epu8(q1p1, q2p2)), in vpx_lpf_horizontal_16_avx2()
76 _mm_or_si128(_mm_subs_epu8(q3p3, q2p2), _mm_subs_epu8(q2p2, q3p3))); in vpx_lpf_horizontal_16_avx2()
131 _mm_or_si128(_mm_subs_epu8(q2p2, q0p0), _mm_subs_epu8(q0p0, q2p2)), in vpx_lpf_horizontal_16_avx2()
182 p2_16 = _mm_unpacklo_epi8(q2p2, zero); in vpx_lpf_horizontal_16_avx2()
187 q2_16 = _mm_unpackhi_epi8(q2p2, zero); in vpx_lpf_horizontal_16_avx2()
309 q2p2 = _mm_andnot_si128(flat, q2p2); in vpx_lpf_horizontal_16_avx2()
[all …]
|
| D | loopfilter_sse2.c | 241 __m128i q7p7, q6p6, q5p5, q4p4, q3p3, q2p2, q1p1, q0p0, p0q0, p1q1; in vpx_lpf_horizontal_16_sse2() local
250 q2p2 = _mm_loadl_epi64((__m128i *)(s - 3 * pitch)); in vpx_lpf_horizontal_16_sse2()
251 q2p2 = _mm_castps_si128( in vpx_lpf_horizontal_16_sse2()
252 _mm_loadh_pi(_mm_castsi128_ps(q2p2), (__m64 *)(s + 2 * pitch))); in vpx_lpf_horizontal_16_sse2()
283 work = _mm_max_epu8(abs_diff(q2p2, q1p1), abs_diff(q3p3, q2p2)); in vpx_lpf_horizontal_16_sse2()
337 flat = _mm_max_epu8(abs_diff(q2p2, q0p0), abs_diff(q3p3, q0p0)); in vpx_lpf_horizontal_16_sse2()
380 p2_16 = _mm_unpacklo_epi8(q2p2, zero); in vpx_lpf_horizontal_16_sse2()
385 q2_16 = _mm_unpackhi_epi8(q2p2, zero); in vpx_lpf_horizontal_16_sse2()
507 q2p2 = _mm_andnot_si128(flat, q2p2); in vpx_lpf_horizontal_16_sse2()
509 q2p2 = _mm_or_si128(q2p2, flat_q2p2); in vpx_lpf_horizontal_16_sse2()
[all …]
|
| /external/libaom/libaom/aom_dsp/arm/ |
| D | loopfilter_neon.c | 159 uint8x8_t q0p0, q1p1, q2p2; in lpf_14_neon() local
227 q2p2 = vreinterpret_u8_u32(vrev64_u32(vreinterpret_u32_u8(*p2q2))); in lpf_14_neon()
244 out_pq0 = vaddw_u8(out_pq0, q2p2); in lpf_14_neon()
286 qp_sum = vaddl_u8(q2p2, q1p1); in lpf_14_neon()
408 uint8x8_t q0p0, q1p1, q2p2; in lpf_8_neon() local
417 q2p2 = vreinterpret_u8_u32(vrev64_u32(vreinterpret_u32_u8(*p2q2))); in lpf_8_neon()
428 out_pq0 = vaddw_u8(out_pq0, q2p2); in lpf_8_neon()
|
| /external/libgav1/libgav1/src/dsp/arm/ |
| D | loop_filter_neon.cc | 566 const uint8x8_t q2p2 = Transpose32(p2q2); in Filter8() local
567 sum = vaddq_u16(vaddl_u8(p0q0, q2p2), sum); in Filter8()
796 const uint8x8_t q2p2 = Transpose32(p2q2); in Filter14() local
797 sum = vaddq_u16(vaddl_u8(p2q2, q2p2), sum); in Filter14()
|
| /external/libgav1/libgav1/src/dsp/x86/ |
| D | loop_filter_sse4.cc | 915 __m128i* q0p0, __m128i* q1p1, __m128i* q2p2, in DualTranspose8x4To4x8() argument
945 *q2p2 = _mm_unpackhi_epi32(_mm_slli_si128(ww1, 4), ww2); in DualTranspose8x4To4x8()
|