| /external/libgav1/libgav1/src/dsp/arm/ |
| D | weight_mask_neon.cc | 43 const int16x8_t pred_0 = vld1q_s16(prediction_0); in WeightMask8_NEON() local
48 vreinterpretq_u16_s16(vabdq_s16(pred_0, pred_1)), kRoundingBits8bpp); in WeightMask8_NEON()
61 WeightMask8_NEON<mask_is_inverse>(pred_0, pred_1, mask)
65 pred_0 += 8; \
72 const auto* pred_0 = static_cast<const int16_t*>(prediction_0); in WeightMask8x8_NEON() local
84 const auto* pred_0 = static_cast<const int16_t*>(prediction_0); in WeightMask8x16_NEON() local
98 const auto* pred_0 = static_cast<const int16_t*>(prediction_0); in WeightMask8x32_NEON() local
113 WeightMask8_NEON<mask_is_inverse>(pred_0, pred_1, mask); \
114 WeightMask8_NEON<mask_is_inverse>(pred_0 + 8, pred_1 + 8, mask + 8)
118 pred_0 += 16; \
[all …]
|
| D | average_blend_neon.cc | 78 const auto* pred_0 = static_cast<const int16_t*>(prediction_0); in AverageBlend_NEON() local
84 const uint8x8_t result = AverageBlend8Row(pred_0, pred_1); in AverageBlend_NEON()
85 pred_0 += 8; in AverageBlend_NEON()
99 vst1_u8(dst, AverageBlend8Row(pred_0, pred_1)); in AverageBlend_NEON()
101 pred_0 += 8; in AverageBlend_NEON()
104 vst1_u8(dst, AverageBlend8Row(pred_0, pred_1)); in AverageBlend_NEON()
106 pred_0 += 8; in AverageBlend_NEON()
115 AverageBlendLargeRow(pred_0, pred_1, width, dst); in AverageBlend_NEON()
117 pred_0 += width; in AverageBlend_NEON()
120 AverageBlendLargeRow(pred_0, pred_1, width, dst); in AverageBlend_NEON()
[all …]
|
| D | mask_blend_neon.cc | 82 inline void WriteMaskBlendLine4x2(const int16_t* const pred_0, in WriteMaskBlendLine4x2() argument
87 const int16x8_t pred_val_0 = vld1q_s16(pred_0); in WriteMaskBlendLine4x2()
112 inline void MaskBlending4x4_NEON(const int16_t* pred_0, const int16_t* pred_1, in MaskBlending4x4_NEON() argument
120 WriteMaskBlendLine4x2(pred_0, pred_1, pred_mask_0, pred_mask_1, dst, in MaskBlending4x4_NEON()
124 pred_0 += 4 << 1; in MaskBlending4x4_NEON()
131 WriteMaskBlendLine4x2(pred_0, pred_1, pred_mask_0, pred_mask_1, dst, in MaskBlending4x4_NEON()
136 inline void MaskBlending4xH_NEON(const int16_t* pred_0, const int16_t* pred_1, in MaskBlending4xH_NEON() argument
143 pred_0, pred_1, mask, mask_stride, dst, dst_stride); in MaskBlending4xH_NEON()
153 WriteMaskBlendLine4x2(pred_0, pred_1, pred_mask_0, pred_mask_1, dst, in MaskBlending4xH_NEON()
155 pred_0 += 4 << 1; in MaskBlending4xH_NEON()
[all …]
|
| D | distance_weighted_blend_neon.cc | 136 const auto* pred_0 = static_cast<const int16_t*>(prediction_0); in DistanceWeightedBlend_NEON() local
143 DistanceWeightedBlendSmall_NEON<4, 4>(pred_0, pred_1, weights, dest, in DistanceWeightedBlend_NEON()
146 DistanceWeightedBlendSmall_NEON<4, 8>(pred_0, pred_1, weights, dest, in DistanceWeightedBlend_NEON()
150 DistanceWeightedBlendSmall_NEON<4, 16>(pred_0, pred_1, weights, dest, in DistanceWeightedBlend_NEON()
159 DistanceWeightedBlendSmall_NEON<8, 4>(pred_0, pred_1, weights, dest, in DistanceWeightedBlend_NEON()
163 DistanceWeightedBlendSmall_NEON<8, 8>(pred_0, pred_1, weights, dest, in DistanceWeightedBlend_NEON()
167 DistanceWeightedBlendSmall_NEON<8, 16>(pred_0, pred_1, weights, dest, in DistanceWeightedBlend_NEON()
172 DistanceWeightedBlendSmall_NEON<8, 32>(pred_0, pred_1, weights, dest, in DistanceWeightedBlend_NEON()
179 DistanceWeightedBlendLarge_NEON(pred_0, pred_1, weights, width, height, dest, in DistanceWeightedBlend_NEON()
275 const auto* pred_0 = static_cast<const uint16_t*>(prediction_0); in DistanceWeightedBlend_NEON() local
[all …]
|
| D | intrapred_smooth_neon.cc | 63 const uint32x4_t pred_0 = vaddl_u16(weighted_top, weighted_left); in CalculatePred() local
65 const uint32x4_t pred_2 = vaddq_u32(pred_0, pred_1); in CalculatePred()
308 const uint8x16_t pred_0 = in SmoothVertical16PlusxN_NEON() local
310 vst1q_u8(dst, pred_0); in SmoothVertical16PlusxN_NEON()
422 const uint8x16_t pred_0 = CalculateHorizontalWeightsAndPred( in SmoothHorizontal16PlusxN_NEON() local
424 vst1q_u8(dst, pred_0); in SmoothHorizontal16PlusxN_NEON()
|
| /external/libgav1/libgav1/src/dsp/x86/ |
| D | average_blend_sse4.cc | 40 const __m128i pred_0 = LoadLo8(prediction_0); in AverageBlend4Row() local
42 __m128i res = _mm_add_epi16(pred_0, pred_1); in AverageBlend4Row()
49 const __m128i pred_0 = LoadAligned16(prediction_0); in AverageBlend8Row() local
51 __m128i res = _mm_add_epi16(pred_0, pred_1); in AverageBlend8Row()
78 const auto* pred_0 = static_cast<const int16_t*>(prediction_0); in AverageBlend_SSE4_1() local
86 AverageBlend4Row(pred_0, pred_1, dst); in AverageBlend_SSE4_1()
88 pred_0 += width; in AverageBlend_SSE4_1()
91 AverageBlend4Row(pred_0, pred_1, dst); in AverageBlend_SSE4_1()
93 pred_0 += width; in AverageBlend_SSE4_1()
103 AverageBlend8Row(pred_0, pred_1, dst); in AverageBlend_SSE4_1()
[all …]
|
| D | weight_mask_sse4.cc | 80 WeightMask16_SSE4<mask_is_inverse, false>(pred_0, pred_1, mask, mask_stride)
84 pred_0 += 8 << 1; \
91 const auto* pred_0 = static_cast<const int16_t*>(prediction_0); in WeightMask8x8_SSE4() local
103 const auto* pred_0 = static_cast<const int16_t*>(prediction_0); in WeightMask8x16_SSE4() local
117 const auto* pred_0 = static_cast<const int16_t*>(prediction_0); in WeightMask8x32_SSE4() local
129 WeightMask16_SSE4<mask_is_inverse, true>(pred_0, pred_1, mask, mask_stride)
133 pred_0 += 16; \
140 const auto* pred_0 = static_cast<const int16_t*>(prediction_0); in WeightMask16x8_SSE4() local
152 const auto* pred_0 = static_cast<const int16_t*>(prediction_0); in WeightMask16x16_SSE4() local
166 const auto* pred_0 = static_cast<const int16_t*>(prediction_0); in WeightMask16x32_SSE4() local
[all …]
|
| D | distance_weighted_blend_sse4.cc | 57 const int16_t* pred_0, const int16_t* pred_1, const uint8_t weight_0, in DistanceWeightedBlend4xH_SSE4_1() argument
64 const __m128i src_00 = LoadLo8(pred_0); in DistanceWeightedBlend4xH_SSE4_1()
66 pred_0 += 4; in DistanceWeightedBlend4xH_SSE4_1()
68 __m128i src_0 = LoadHi8(src_00, pred_0); in DistanceWeightedBlend4xH_SSE4_1()
70 pred_0 += 4; in DistanceWeightedBlend4xH_SSE4_1()
74 const __m128i src_01 = LoadLo8(pred_0); in DistanceWeightedBlend4xH_SSE4_1()
76 pred_0 += 4; in DistanceWeightedBlend4xH_SSE4_1()
78 src_0 = LoadHi8(src_01, pred_0); in DistanceWeightedBlend4xH_SSE4_1()
80 pred_0 += 4; in DistanceWeightedBlend4xH_SSE4_1()
101 const int16_t* pred_0, const int16_t* pred_1, const uint8_t weight_0, in DistanceWeightedBlend8xH_SSE4_1() argument
[all …]
|
| D | mask_blend_sse4.cc | 119 inline void WriteMaskBlendLine4x2(const int16_t* const pred_0, in WriteMaskBlendLine4x2() argument
124 const __m128i pred_val_0 = LoadAligned16(pred_0); in WriteMaskBlendLine4x2()
148 inline void MaskBlending4x4_SSE4(const int16_t* pred_0, const int16_t* pred_1, in MaskBlending4x4_SSE4() argument
156 WriteMaskBlendLine4x2(pred_0, pred_1, pred_mask_0, pred_mask_1, dst, in MaskBlending4x4_SSE4()
158 pred_0 += 4 << 1; in MaskBlending4x4_SSE4()
165 WriteMaskBlendLine4x2(pred_0, pred_1, pred_mask_0, pred_mask_1, dst, in MaskBlending4x4_SSE4()
170 inline void MaskBlending4xH_SSE4(const int16_t* pred_0, const int16_t* pred_1, in MaskBlending4xH_SSE4() argument
177 pred_0, pred_1, mask, mask_stride, dst, dst_stride); in MaskBlending4xH_SSE4()
187 WriteMaskBlendLine4x2(pred_0, pred_1, pred_mask_0, pred_mask_1, dst, in MaskBlending4xH_SSE4()
189 pred_0 += 4 << 1; in MaskBlending4xH_SSE4()
[all …]
|
| /external/libgav1/libgav1/src/dsp/ |
| D | average_blend.cc | 38 const auto* pred_0 = static_cast<const PredType*>(prediction_0); in AverageBlend_C() local
48 int res = pred_0[x] + pred_1[x]; in AverageBlend_C()
56 pred_0 += width; in AverageBlend_C()
|
| D | distance_weighted_blend.cc | 39 const auto* pred_0 = static_cast<const PredType*>(prediction_0); in DistanceWeightedBlend_C() local
50 int res = pred_0[x] * weight_0 + pred_1[x] * weight_1; in DistanceWeightedBlend_C()
58 pred_0 += width; in DistanceWeightedBlend_C()
|
| D | weight_mask.cc | 36 const auto* pred_0 = static_cast<const PredType*>(prediction_0); in WeightMask_C() local
44 std::abs(pred_0[x] - pred_1[x]), rounding_bits); in WeightMask_C()
49 pred_0 += width; in WeightMask_C()
|
| D | mask_blend.cc | 54 const auto* pred_0 = static_cast<const PredType*>(prediction_0); in MaskBlend_C() local
69 mask_value * pred_1[x] + (64 - mask_value) * pred_0[x], 6)); in MaskBlend_C()
72 int res = (mask_value * pred_0[x] + (64 - mask_value) * pred_1[x]) >> 6; in MaskBlend_C()
82 pred_0 += width; in MaskBlend_C()
|
| /external/libhevc/encoder/arm/ |
| D | ihevce_ssd_calculator_neon.c | 140 uint8x16_t src_0, pred_0, src_1, pred_1, abs_0, abs_1; in ihevce_1x32_ssd_computer_neon() local
146 pred_0 = vld1q_u8(pu1_pred); in ihevce_1x32_ssd_computer_neon()
153 pred_0 = vld2q_u8(pu1_pred).val[chroma_plane]; in ihevce_1x32_ssd_computer_neon()
157 abs_0 = vabdq_u8(src_0, pred_0); in ihevce_1x32_ssd_computer_neon()
180 uint8x16_t pred_0, pred_1, pred_2, pred_3; in ihevce_1x64_ssd_computer_neon() local
188 pred_0 = vld1q_u8(pu1_pred); in ihevce_1x64_ssd_computer_neon()
199 pred_0 = vld2q_u8(pu1_pred).val[chroma_plane]; in ihevce_1x64_ssd_computer_neon()
207 abs_0 = vabdq_u8(src_0, pred_0); in ihevce_1x64_ssd_computer_neon()
|
| D | ihevce_sad_compute_neon.c | 144 const uint8x16_t pred_0 = vld1q_u8(pu1_pred); in ihevce_32xn_sad_computer_neon() local
148 abs_0 = vabal_u8(abs_0, vget_low_u8(src_0), vget_low_u8(pred_0)); in ihevce_32xn_sad_computer_neon()
149 abs_0 = vabal_u8(abs_0, vget_high_u8(src_0), vget_high_u8(pred_0)); in ihevce_32xn_sad_computer_neon()
178 const uint8x16_t pred_0 = vld1q_u8(pu1_pred); in ihevce_64xn_sad_computer_neon() local
186 abs_0 = vabal_u8(abs_0, vget_low_u8(src_0), vget_low_u8(pred_0)); in ihevce_64xn_sad_computer_neon()
187 abs_0 = vabal_u8(abs_0, vget_high_u8(src_0), vget_high_u8(pred_0)); in ihevce_64xn_sad_computer_neon()
|
| D | ihevce_ssd_and_sad_calculator_neon.c | 186 const uint8x16_t pred_0 = vld1q_u8(pu1_recon); in ihevce_ssd_and_sad_calculator_neon() local
190 abs_l = vabd_u8(vget_low_u8(src_0), vget_low_u8(pred_0)); in ihevce_ssd_and_sad_calculator_neon()
191 abs_h = vabd_u8(vget_high_u8(src_0), vget_high_u8(pred_0)); in ihevce_ssd_and_sad_calculator_neon()
245 const uint8x16_t pred_0 = vld1q_u8(pu1_recon); in ihevce_ssd_and_sad_calculator_neon() local
253 abs_l = vabd_u8(vget_low_u8(src_0), vget_low_u8(pred_0)); in ihevce_ssd_and_sad_calculator_neon()
254 abs_h = vabd_u8(vget_high_u8(src_0), vget_high_u8(pred_0)); in ihevce_ssd_and_sad_calculator_neon()
|