/external/libaom/libaom/aom_dsp/x86/ |
D | blend_mask_sse4.h | 29 const __m128i *m, const __m128i *v_round_offset, const __m128i *v_maxval, in blend_a64_d16_mask_w4_sse41() argument 31 const __m128i max_minus_m = _mm_sub_epi16(*v_maxval, *m); in blend_a64_d16_mask_w4_sse41() 47 const __m128i *m, const __m128i *v_round_offset, const __m128i *v_maxval, in blend_a64_d16_mask_w8_sse41() argument 49 const __m128i max_minus_m = _mm_sub_epi16(*v_maxval, *m); in blend_a64_d16_mask_w8_sse41() 69 const __m128i v_maxval = _mm_set1_epi16(AOM_BLEND_A64_MAX_ALPHA); in aom_lowbd_blend_a64_d16_mask_subw0_subh0_w4_sse4_1() local 74 blend_a64_d16_mask_w4_sse41(dst, src0, src1, &m, round_offset, &v_maxval, in aom_lowbd_blend_a64_d16_mask_subw0_subh0_w4_sse4_1() 88 const __m128i v_maxval = _mm_set1_epi16(AOM_BLEND_A64_MAX_ALPHA); in aom_lowbd_blend_a64_d16_mask_subw0_subh0_w8_sse4_1() local 92 blend_a64_d16_mask_w8_sse41(dst, src0, src1, &m, round_offset, &v_maxval, in aom_lowbd_blend_a64_d16_mask_subw0_subh0_w8_sse4_1() 106 const __m128i v_maxval = _mm_set1_epi16(AOM_BLEND_A64_MAX_ALPHA); in aom_lowbd_blend_a64_d16_mask_subw1_subh1_w4_sse4_1() local 117 blend_a64_d16_mask_w4_sse41(dst, src0, src1, &m, round_offset, &v_maxval, in aom_lowbd_blend_a64_d16_mask_subw1_subh1_w4_sse4_1() [all …]
|
D | blend_a64_mask_avx2.c | 30 const __m256i *m0, const __m256i *v_round_offset, const __m256i *v_maxval, in blend_a64_d16_mask_w16_avx2() argument 32 const __m256i max_minus_m0 = _mm256_sub_epi16(*v_maxval, *m0); in blend_a64_d16_mask_w16_avx2() 52 const __m256i *v_maxval, int shift) { in blend_a64_d16_mask_w32_avx2() argument 53 const __m256i max_minus_m0 = _mm256_sub_epi16(*v_maxval, *m0); in blend_a64_d16_mask_w32_avx2() 54 const __m256i max_minus_m1 = _mm256_sub_epi16(*v_maxval, *m1); in blend_a64_d16_mask_w32_avx2() 87 const __m256i v_maxval = _mm256_set1_epi16(AOM_BLEND_A64_MAX_ALPHA); in lowbd_blend_a64_d16_mask_subw0_subh0_w16_avx2() local 92 blend_a64_d16_mask_w16_avx2(dst, src0, src1, &m0, round_offset, &v_maxval, in lowbd_blend_a64_d16_mask_subw0_subh0_w16_avx2() 106 const __m256i v_maxval = _mm256_set1_epi16(AOM_BLEND_A64_MAX_ALPHA); in lowbd_blend_a64_d16_mask_subw0_subh0_w32_avx2() local 114 round_offset, &v_maxval, shift); in lowbd_blend_a64_d16_mask_subw0_subh0_w32_avx2() 128 const __m256i v_maxval = _mm256_set1_epi16(AOM_BLEND_A64_MAX_ALPHA); in lowbd_blend_a64_d16_mask_subw1_subh1_w16_avx2() local [all …]
|
D | blend_a64_mask_sse4.c | 877 const __m128i *v_maxval, int shift) { in blend_a64_d16_mask_w16_sse41() argument 878 const __m128i max_minus_m0 = _mm_sub_epi16(*v_maxval, *m0); in blend_a64_d16_mask_w16_sse41() 879 const __m128i max_minus_m1 = _mm_sub_epi16(*v_maxval, *m1); in blend_a64_d16_mask_w16_sse41() 908 const __m128i v_maxval = _mm_set1_epi16(AOM_BLEND_A64_MAX_ALPHA); in lowbd_blend_a64_d16_mask_subw0_subh0_w16_sse4_1() local 916 round_offset, &v_maxval, shift); in lowbd_blend_a64_d16_mask_subw0_subh0_w16_sse4_1() 930 const __m128i v_maxval = _mm_set1_epi16(AOM_BLEND_A64_MAX_ALPHA); in lowbd_blend_a64_d16_mask_subw1_subh1_w16_sse4_1() local 948 round_offset, &v_maxval, shift); in lowbd_blend_a64_d16_mask_subw1_subh1_w16_sse4_1() 962 const __m128i v_maxval = _mm_set1_epi16(AOM_BLEND_A64_MAX_ALPHA); in lowbd_blend_a64_d16_mask_subw1_subh0_w16_sse4_1() local 975 round_offset, &v_maxval, shift); in lowbd_blend_a64_d16_mask_subw1_subh0_w16_sse4_1() 989 const __m128i v_maxval = _mm_set1_epi16(AOM_BLEND_A64_MAX_ALPHA); in lowbd_blend_a64_d16_mask_subw0_subh1_w16_sse4_1() local [all …]
|
/external/libaom/libaom/aom_dsp/arm/ |
D | blend_a64_mask_neon.c | 23 const int16x8_t v_maxval, int16x8_t *res) { in blend8x1() argument 25 const int16x8_t max_minus_mask = vsubq_s16(v_maxval, mask); in blend8x1() 43 int16x8_t mask3, const int16x8_t v_maxval, in blend_8x4() argument 55 blend8x1(mask0, src0_0, src1_0, v_maxval, &im_res_0); in blend_8x4() 56 blend8x1(mask1, src0_1, src1_1, v_maxval, &im_res_1); in blend_8x4() 57 blend8x1(mask2, src0_2, src1_2, v_maxval, &im_res_2); in blend_8x4() 58 blend8x1(mask3, src0_3, src1_3, v_maxval, &im_res_3); in blend_8x4() 84 int16x4_t mask3, const int16x8_t v_maxval, in blend_4x4() argument 106 blend8x1(mask0_1, src0_0, src1_0, v_maxval, &res0); in blend_4x4() 107 blend8x1(mask2_3, src0_1, src1_1, v_maxval, &res1); in blend_4x4() [all …]
|