hashirama/uvg266
1
0
Fork 0
mirror of https://github.com/ultravideo/uvg266.git synced 2024-11-27 19:24:06 +00:00
Code Issues Actions Projects Releases Wiki Activity

made to clear sum vector after it has been set to memory

Browse source
This commit is contained in:
Reima Hyvönen 2019-03-29 15:37:22 +02:00 committed by Pauli Oikkonen
parent f146de7acb
commit 0e0f2d3490
1 changed files with 5 additions and 1 deletions
Show stats Download patch file Download diff file Expand all files Collapse all files

6
src/strategies/avx2/sao-avx2.c
View file

@ -95,6 +95,7 @@ static int sao_edge_ddistortion_avx2(const kvz_pixel *orig_data,
__m256i tmp1_vec_epi32;
__m256i tmp2_vec_epi32;
int sum = 0;
for (y = 1; y < block_height - 1; ++y) {
for (x = 1; x < block_width - 8; x+=8) {
const kvz_pixel *c_data = &rec_data[y * block_width + x];
@ -155,8 +156,11 @@ static int sao_edge_ddistortion_avx2(const kvz_pixel *orig_data,
tmp_sum_epi32 = _mm256_hadd_epi32(tmp_sum_epi32, tmp_sum_epi32);
tmp_sum_epi32 = _mm256_add_epi32(tmp_sum_epi32, _mm256_shuffle_epi32(tmp_sum_epi32, _MM_SHUFFLE(0, 1, 0, 1)));
sum += _mm_cvtsi128_si32(_mm256_castsi256_si128(tmp_sum_epi32));
tmp_sum_epi32 = _mm256_setzero_si256();
}
return (_mm_cvtsi128_si32(_mm256_castsi256_si128(tmp_sum_epi32)));
return sum;
}
/**