Searched refs:sum_high (Results 1 – 4 of 4) sorted by relevance
/external/libgav1/src/dsp/x86/ |
D | warp_sse4.cc | 108 __m128i sum_high = sum_low; in WriteVerticalFilter() local 124 sum_high = _mm_add_epi32(sum_high, product_high); in WriteVerticalFilter() 127 sum_high = RightShiftWithRounding_S32(sum_high, kRoundBitsVertical); in WriteVerticalFilter() 129 const __m128i sum = _mm_packs_epi32(sum_low, sum_high); in WriteVerticalFilter() 132 const __m128i sum = _mm_packus_epi32(sum_low, sum_high); in WriteVerticalFilter() 145 __m128i sum_high = _mm_setzero_si128(); in WriteVerticalFilter() local 156 sum_high = _mm_add_epi32(sum_high, product_high); in WriteVerticalFilter() 159 sum_high = RightShiftWithRounding_S32(sum_high, kRoundBitsVertical); in WriteVerticalFilter() 161 const __m128i sum = _mm_packs_epi32(sum_low, sum_high); in WriteVerticalFilter() 164 const __m128i sum = _mm_packus_epi32(sum_low, sum_high); in WriteVerticalFilter()
|
/external/libgav1/src/dsp/arm/ |
D | warp_neon.cc | 306 int32x4_t sum_high = sum_low; in Warp_NEON() local 311 sum_high = in Warp_NEON() 312 vmlal_n_s16(sum_high, vget_high_s16(filter[k]), intermediate); in Warp_NEON() 316 vrshrn_n_s32(sum_high, kRoundBitsVertical)); in Warp_NEON() 412 int32x4_t sum_high = sum_low; in Warp_NEON() local 417 sum_high = vmlal_s16(sum_high, vget_high_s16(filter[k]), in Warp_NEON() 422 vrshrn_n_s32(sum_high, kRoundBitsVertical)); in Warp_NEON() 739 int32x4_t sum_high = sum_low; in Warp_NEON() local 744 sum_high = in Warp_NEON() 745 vmlal_n_s16(sum_high, vget_high_s16(filter[k]), intermediate); in Warp_NEON() [all …]
|
/external/libvpx/vpx_dsp/arm/ |
D | deblock_neon.c | 289 static uint8x8_t combine_mask(const int16x4_t sum_low, const int16x4_t sum_high, in combine_mask() argument 295 calculate_mask(sum_high, sumsq_high, f, fifteen); in combine_mask() 321 int16x4_t sum_low, sum_high; in vpx_mbpost_proc_across_ip_neon() local 334 sum_high = vdup_n_s16(sum); in vpx_mbpost_proc_across_ip_neon() 359 sum_low = vdup_lane_s16(sum_high, 3); in vpx_mbpost_proc_across_ip_neon() 366 sum_high = vdup_lane_s16(sum_low, 3); in vpx_mbpost_proc_across_ip_neon() 369 accumulate_sum_sumsq(vget_high_s16(x), xy_high, &sum_high, &sumsq_high); in vpx_mbpost_proc_across_ip_neon() 371 mask = combine_mask(sum_low, sum_high, sumsq_low, sumsq_high, f); in vpx_mbpost_proc_across_ip_neon() 373 output = filter_pixels(vcombine_s16(sum_low, sum_high), s); in vpx_mbpost_proc_across_ip_neon()
|
/external/libaom/av1/encoder/x86/ |
D | av1_k_means_avx2.c | 20 const __m128i sum_high = _mm_unpackhi_epi64(sum, sum); in k_means_horizontal_sum_avx2() local 22 _mm_storel_epi64((__m128i *)&res, _mm_add_epi64(sum, sum_high)); in k_means_horizontal_sum_avx2()
|