Searched refs:res_8 (Results 1 – 8 of 8) sorted by relevance
| /external/libaom/libaom/av1/common/x86/ |
| D | jnt_convolve_avx2.c | 108 const __m256i res_8 = _mm256_packus_epi16(round_result, round_result); in av1_dist_wtd_convolve_x_avx2() local
109 const __m128i res_0 = _mm256_castsi256_si128(res_8); in av1_dist_wtd_convolve_x_avx2()
110 const __m128i res_1 = _mm256_extracti128_si256(res_8, 1); in av1_dist_wtd_convolve_x_avx2()
163 const __m256i res_8 = _mm256_packus_epi16(round_result, round_result); in av1_dist_wtd_convolve_x_avx2() local
164 const __m128i res_0 = _mm256_castsi256_si128(res_8); in av1_dist_wtd_convolve_x_avx2()
165 const __m128i res_1 = _mm256_extracti128_si256(res_8, 1); in av1_dist_wtd_convolve_x_avx2()
305 const __m256i res_8 = in av1_dist_wtd_convolve_y_avx2() local
307 const __m128i res_0 = _mm256_castsi256_si128(res_8); in av1_dist_wtd_convolve_y_avx2()
308 const __m128i res_1 = _mm256_extracti128_si256(res_8, 1); in av1_dist_wtd_convolve_y_avx2()
375 const __m256i res_8 = in av1_dist_wtd_convolve_y_avx2() local
[all …]
|
| D | jnt_convolve_sse2.c | 85 const __m128i res_8 = _mm_packus_epi16(round_result, round_result); in av1_dist_wtd_convolve_x_sse2() local
86 *(uint32_t *)(&dst0[0]) = _mm_cvtsi128_si32(res_8); in av1_dist_wtd_convolve_x_sse2()
142 const __m128i res_8 = _mm_packus_epi16(round_result, round_result); in av1_dist_wtd_convolve_x_sse2() local
143 _mm_storel_epi64((__m128i *)(&dst0[i * dst_stride0 + j]), res_8); in av1_dist_wtd_convolve_x_sse2()
233 const __m128i res_8 = _mm_packus_epi16(round_result, round_result); in av1_dist_wtd_convolve_y_sse2() local
234 *(uint32_t *)(&dst0[0]) = _mm_cvtsi128_si32(res_8); in av1_dist_wtd_convolve_y_sse2()
262 const __m128i res_8 = _mm_packus_epi16(round_result, round_result); in av1_dist_wtd_convolve_y_sse2() local
263 *(uint32_t *)(&dst0[0]) = _mm_cvtsi128_si32(res_8); in av1_dist_wtd_convolve_y_sse2()
339 const __m128i res_8 = _mm_packus_epi16(round_result, round_result); in av1_dist_wtd_convolve_y_sse2() local
340 _mm_storel_epi64((__m128i *)(&dst0[i * dst_stride0 + j]), res_8); in av1_dist_wtd_convolve_y_sse2()
[all …]
|
| D | reconinter_sse4.c | 141 const __m128i res_8 = _mm_packus_epi16(diff_const_16, diff_const_16); in av1_build_compound_diffwtd_mask_d16_sse4_1() local
147 _mm_storel_epi64(dst, res_8); in av1_build_compound_diffwtd_mask_d16_sse4_1()
149 *(uint32_t *)dst = _mm_cvtsi128_si32(res_8); in av1_build_compound_diffwtd_mask_d16_sse4_1()
|
| D | convolve_2d_sse2.c | 425 const __m128i res_8 = in av1_dist_wtd_convolve_2d_copy_sse2() local
428 _mm_store_si128((__m128i *)(&dst0[j]), res_8); in av1_dist_wtd_convolve_2d_copy_sse2()
456 const __m128i res_8 = _mm_packus_epi16(round_result, round_result); in av1_dist_wtd_convolve_2d_copy_sse2() local
459 _mm_storel_epi64((__m128i *)(&dst0[j]), res_8); in av1_dist_wtd_convolve_2d_copy_sse2()
461 *(uint32_t *)(&dst0[j]) = _mm_cvtsi128_si32(res_8); in av1_dist_wtd_convolve_2d_copy_sse2()
|
| D | jnt_convolve_ssse3.c | 218 const __m128i res_8 = _mm_packus_epi16(round_result, round_result); in av1_dist_wtd_convolve_2d_ssse3() local
221 _mm_storel_epi64((__m128i *)(&dst0[i * dst_stride0 + j]), res_8); in av1_dist_wtd_convolve_2d_ssse3()
224 _mm_cvtsi128_si32(res_8); in av1_dist_wtd_convolve_2d_ssse3()
|
| /external/libaom/libaom/aom_dsp/x86/ |
| D | convolve_avx2.h | 190 const __m256i res_8 = _mm256_packus_epi16(round_result, round_result); \
191 const __m128i res_0 = _mm256_castsi256_si128(res_8); \
192 const __m128i res_1 = _mm256_extracti128_si256(res_8, 1); \
219 const __m256i res_8 = _mm256_packus_epi16(round_result, round_result); \
220 const __m128i res_0 = _mm256_castsi256_si128(res_8); \
221 const __m128i res_1 = _mm256_extracti128_si256(res_8, 1); \
|
| /external/gemmlowp/internal/ |
| D | output_sse.h | 39 __m128i res_8 = _mm_packus_epi16(res_16, res_16);
40 output.reg[0] = _mm_cvtsi128_si32(res_8);
58 __m128i res_8 = _mm_packus_epi16(res_16, res_16);
59 output.reg[0] = _mm_extract_epi32(res_8, 0);
60 output.reg[1] = _mm_extract_epi32(res_8, 1);
|
| D | output_neon.h | 39 uint8x8_t res_8 = vqmovun_s16(vcombine_s16(res_16, res_16));
40 output.reg[0] = vget_lane_u32(vreinterpret_u32_u8(res_8), 0);
|