| /external/libaom/libaom/aom_dsp/x86/ |
| D | loopfilter_sse2.c | 141 static AOM_FORCE_INLINE void filter4_sse2(__m128i *p1p0, __m128i *q1q0, in filter4_sse2() argument
152 ps1ps0_work = _mm_xor_si128(*p1p0, t80); /* ^ 0x80 */ in filter4_sse2()
192 static AOM_FORCE_INLINE void filter4_dual_sse2(__m128i *p1p0, __m128i *q1q0, in filter4_dual_sse2() argument
204 ps1ps0_work = _mm_xor_si128(*p1p0, t80); /* ^ 0x80 */ in filter4_dual_sse2()
247 __m128i q1p1, q0p0, p1p0, q1q0; in lpf_internal_4_sse2() local
255 p1p0 = _mm_unpacklo_epi32(q0p0, q1p1); in lpf_internal_4_sse2()
256 q1q0 = _mm_srli_si128(p1p0, 8); in lpf_internal_4_sse2()
261 __m128i abs_p1q1p0q0 = abs_diff(p1p0, q1q0); in lpf_internal_4_sse2()
284 filter4_sse2(&p1p0, &q1q0, &hev, &mask, q1q0_out, p1p0_out); in lpf_internal_4_sse2()
290 __m128i q1p1, q0p0, p1p0, q1q0; in lpf_internal_4_dual_sse2() local
[all …]
|
| D | highbd_loopfilter_sse2.c | 109 __m128i *p1p0, __m128i *q1q0, in highbd_hev_filter_mask_x_sse2() argument
119 *p1p0 = _mm_unpacklo_epi64(pq[0], pq[1]); in highbd_hev_filter_mask_x_sse2()
122 abs_p0q0_p1q1 = abs_diff16(*p1p0, *q1q0); in highbd_hev_filter_mask_x_sse2()
213 static AOM_FORCE_INLINE void highbd_filter4_sse2(__m128i *p1p0, __m128i *q1q0, in highbd_filter4_sse2() argument
228 ps1ps0_work = _mm_subs_epi16(*p1p0, *t80); in highbd_filter4_sse2()
339 __m128i p1p0, q1q0, abs_p1p0; in highbd_lpf_internal_14_sse2() local
341 highbd_hev_filter_mask_x_sse2(pq, 4, &p1p0, &q1q0, &abs_p1p0, &limit, &blimit, in highbd_lpf_internal_14_sse2()
346 highbd_filter4_sse2(&p1p0, &q1q0, &hevhev, &mask, &qs0qs1, &ps0ps1, &t80, bd); in highbd_lpf_internal_14_sse2()
743 __m128i p1p0, q1q0, abs_p1p0, ps1ps0, qs1qs0; in highbd_lpf_internal_6_sse2() local
757 highbd_hev_filter_mask_x_sse2(pq, 3, &p1p0, &q1q0, &abs_p1p0, &limit, &blimit, in highbd_lpf_internal_6_sse2()
[all …]
|
| /external/libgav1/libgav1/src/dsp/x86/ |
| D | loop_filter_sse4.cc | 52 inline __m128i CheckOuterThreshF4(const __m128i& q1q0, const __m128i& p1p0, in CheckOuterThreshF4() argument
56 const __m128i abs_pmq = AbsDiff(p1p0, q1q0); in CheckOuterThreshF4()
91 inline __m128i NeedsFilter4(const __m128i& q1q0, const __m128i& p1p0, in NeedsFilter4() argument
95 const __m128i outer_mask = CheckOuterThreshF4(q1q0, p1p0, outer_thresh); in NeedsFilter4()
160 const __m128i p1p0 = _mm_unpacklo_epi32(p0, p1); in Horizontal4() local
164 NeedsFilter4(q1q0, p1p0, qp1, qp0, v_outer_thresh, v_inner_thresh); in Horizontal4()
231 const __m128i p1p0 = _mm_shuffle_epi32(d0, 0x1); in Vertical4() local
235 NeedsFilter4(q1q0, p1p0, qp1, qp0, v_outer_thresh, v_inner_thresh); in Vertical4()
257 inline __m128i NeedsFilter6(const __m128i& q1q0, const __m128i& p1p0, in NeedsFilter6() argument
261 const __m128i outer_mask = CheckOuterThreshF4(q1q0, p1p0, outer_thresh); in NeedsFilter6()
[all …]
|
| /external/libaom/libaom/aom_dsp/arm/ |
| D | loopfilter_neon.c | 788 uint8x8_t UNINITIALIZED_IS_SAFE(p1p0), UNINITIALIZED_IS_SAFE(q0q1); in aom_lpf_vertical_4_neon()
795 load_unaligned_u8_4x4(src - 2, stride, (uint32x2_t *)&p1p0, in aom_lpf_vertical_4_neon()
798 transpose_u8_4x4(&p1p0, &q0q1); in aom_lpf_vertical_4_neon()
800 p1q0_p0q1 = vtrn_u32(vreinterpret_u32_u8(p1p0), vreinterpret_u32_u8(q0q1)); in aom_lpf_vertical_4_neon()
812 p1p0 = vreinterpret_u8_u32(p1p0_q1q0.val[0]); in aom_lpf_vertical_4_neon()
815 transpose_u8_4x4(&p1p0, &q0q1); in aom_lpf_vertical_4_neon()
817 store_unaligned_u8_4x1(src - 2, p1p0, 0); in aom_lpf_vertical_4_neon()
819 store_unaligned_u8_4x1((src - 2) + 2 * stride, p1p0, 1); in aom_lpf_vertical_4_neon()
|
| /external/libvpx/libvpx/vpx_dsp/x86/ |
| D | loopfilter_sse2.c | 28 const __m128i abs_p1q1p0q0 = abs_diff(p1p0, q1q0); \
67 ps1ps0 = _mm_xor_si128(p1p0, t80); /* ^ 0x80 */ \
116 __m128i q1p1, q0p0, p3p2, p2p1, p1p0, q3q2, q2q1, q1q0, ps1ps0, qs1qs0; in vpx_lpf_horizontal_4_sse2() local
127 p1p0 = _mm_unpacklo_epi64(q0p0, q1p1); in vpx_lpf_horizontal_4_sse2()
151 __m128i q1p1, q0p0, p3p2, p2p1, p1p0, q3q2, q2q1, q1q0, ps1ps0, qs1qs0; in vpx_lpf_vertical_4_sse2() local
172 p1p0 = _mm_unpacklo_epi16(q1q0, x1); in vpx_lpf_vertical_4_sse2()
176 p3p2 = _mm_unpacklo_epi32(p1p0, x0); in vpx_lpf_vertical_4_sse2()
178 p1p0 = _mm_unpackhi_epi32(p1p0, x0); in vpx_lpf_vertical_4_sse2()
180 p1p0 = _mm_unpackhi_epi64(p1p0, _mm_slli_si128(p1p0, 8)); // swap lo and high in vpx_lpf_vertical_4_sse2()
191 q0p0 = _mm_unpacklo_epi64(p1p0, q1q0); in vpx_lpf_vertical_4_sse2()
[all …]
|